From acallejas at appliedminds.com  Sat Sep  1 18:41:46 2007
From: acallejas at appliedminds.com (Augusto Callejas)
Date: Sat, 01 Sep 2007 11:41:46 -0700
Subject: [Developers] querying a property in reverse
In-Reply-To: <111CAC82-63E7-4CF8-BD87-2C03AB3E5180@metaweb.com>
Message-ID: <C2FF007A.FD6%acallejas@appliedminds.com>

hi-

some properties do not have an inverse, like "Country of origin" for the
type "/film/film":

  http://mjtemplate.org/examples/schemas/type.html?id=/film/film

so its easy to query the country for a given film in one direction:

{
  "query":[{
    "country":{
      "name":null,
      "type":"/location/country"
    },
    "name":"The Matrix",
    "type":"/film/film"
  }]
}

but not in the other direction:

{
  "query":[{
    "/type/reflect/any_reverse":[{
      "link":{
        "master_property":"/film/film/country"
      },
      "name":"The Matrix",
      "type":"/film/film"
    }],
    "name":null,
    "type":"/location/country"
  }]
}

this problem arises when you have a general tree as a query and you need to
traverse in a certain direction.  the simplest example is a query of the
form:

TypeA-->TypeB<--TypeC

where the directed arrow is a master property but does not have a reverse
property.  any choice of the start of the query results in having to specify
the reverse property of a master property that doesn't have one.

augusto.

On 8/28/07 9:22 PM, "John Giannandrea" <jg at metaweb.com> wrote:

> 
> Augusto Callejas wrote:
>> however, say i want to always want to start from the performance,
>> but i don't know the relationship from performance to film is
>> "film" but i
>> know the relationship from film to performance is "starring".  how
>> can i
>> query on the "starring" property but in reverse without having to
>> perform a
>> separate query for the reverse property?
> 
> You can do this in a roundabout way, with something like this:
> 
> {
>    "query":[{
>      "/type/reflect/any_master":[{
>        "link":{
>          "master_property":{
>            "reverse_property":"/film/film/starring"
>          }
>        },
>        "name":"The Star Wars Holiday Special",
>        "type":"/film/film"
>      }],
>      "character":"Han Solo",
>      "type":"/film/performance"
>    }]
> }
> 
> However this requires that you know which property is the master and
> reverse
> otherwise you would need two clauses mirroring each other.
> 
> Im curious about your use case that would know one property id but
> not be able to find the inverse from the schema.   Can you say a
> little more
> about how this situation would arise in your app?
> 
> -jg
> 



From mikelove1 at gmail.com  Sun Sep  2 04:07:15 2007
From: mikelove1 at gmail.com (Mike Love)
Date: Sat, 1 Sep 2007 21:07:15 -0700
Subject: [Developers] CVT query question
Message-ID: <f121137a0709012107r661def4cw7a5d5d41806e9739@mail.gmail.com>

Hi,

I have a question on how to query through the CVT in this peer relationship:
http://www.freebase.com/view/schema/user/mikelove/default_domain/peer_relationship

I wanted to make a symmetrical analog of the influenced / influenced_by
relationship, and had to make it a CVT to get this result.

Here's a query that I would like to use:

{
  "query":{
    "id":null,
    "influenced":[{}],
    "influenced_by":[{}],
    "name":"William Wordsworth",
    "peers":[{}],
    "type":"/user/mikelove/default_domain/influence_node"
  }
}

This query below works, although it gives both names in each relationship
and the whole query comes back empty if the person had no peers:

{
  "query":{
    "id":null,
    "influenced":[{}],
    "influenced_by":[{}],
    "name":"William Wordsworth",
    "peers":[{
      "peers":[]
    }],
    "type":"/user/mikelove/default_domain/influence_node"
  }
}

Thanks muchly!

Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070901/ea734cd6/attachment.htm 

From jg at metaweb.com  Sun Sep  2 05:20:33 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sat, 1 Sep 2007 22:20:33 -0700
Subject: [Developers] CVT query question
In-Reply-To: <f121137a0709012107r661def4cw7a5d5d41806e9739@mail.gmail.com>
References: <f121137a0709012107r661def4cw7a5d5d41806e9739@mail.gmail.com>
Message-ID: <2C93245D-68C7-4695-B795-41F1DEA2F0F8@metaweb.com>


Mike Love wrote:
> I wanted to make a symmetrical analog of the influenced /  
> influenced_by relationship, and had to make it a CVT to get this  
> result.

Shooting from the hip I would suggest that 'peer' could be a property  
on influence_node rather than a new type.

> This query below works, although it gives both names in each  
> relationship and the whole query comes back empty if the person had  
> no peers:

You need to expand it further with an optional clause.

     "peers":[{
       "peers":[{
         "name":null,
         "optional":true
       }]

-jg




From kurt at metaweb.com  Tue Sep  4 01:14:40 2007
From: kurt at metaweb.com (Kurt Bollacker)
Date: Tue, 4 Sep 2007 01:14:40 +0000
Subject: [Developers] Unlinking Objects of Type...
In-Reply-To: <cbf942d90708251535m7523267bx502f594fa97fdf6@mail.gmail.com>
References: <cbf942d90708222031w5d1f2c6etb72aa322aa72404d@mail.gmail.com>
	<20070823050509.GK14323@metaweb.com>
	<cbf942d90708251535m7523267bx502f594fa97fdf6@mail.gmail.com>
Message-ID: <20070904011440.GN31621@metaweb.com>


On Sat, Aug 25, 2007 at 03:35:34PM -0700, Michael Jones wrote:
> Thanks folks.
> 
> I finally just bit the bullet and converted the python unlinker.py
> over to Perl.  It works fine.  If anyone wants the code, let me know,
> but it is not at all polished.

Please post/email this code.  From the perspective of a MQL user,
there are multiple flavors of topic deletion:

#1 You want to "hide" a topic as a type by simply deleting the typing
   link.  The properties for that type still exist, and so "undo" of
   the de-typing is easy.

#2 topic A has type T0 and you want to delete the typing and
   properties for T0.

#3 You want to fully delete topic A.  Note that this is a subset of
   the union of #2 for all types Tx on A, because there are possibly
   incoming links that are not visible as properties on A.

The unlink.py python code in the MQL docs does not consider incoming
links, and may not fully delete a topic.

The Freebase UI client has a feature to allow users to flag a topic
with "mark for delete".  Metaweb has an administrative process to vet
such marks and execute deletions.  You could do this through the Web
client if the number of deletions is small enough.

								Kurt
								:-)



 
> MJ
> 
> On 8/22/07, Kurt Bollacker <kurt at metaweb.com> wrote:
> > On Wed, Aug 22, 2007 at 08:31:09PM -0700, Michael Jones wrote:
> > > Hi Folks,
> > >
> > > I'm looking for a way to unlink some instances of a specific type (I'd
> > > prefer to delete them entirely, but that another discussion.)  There
> > > are a lot of them, so manually doing it is out.  I could write a
> > > script, but it seems this ought to be doable with a single query.
> > >
> > > My attempts, which all fail due to uniqueness errors, look like this:
> > >
> > > {"query":[{
> > >     type:[{
> > >       {"id":"/user/szaijan/fantasy_football/statline",
> > >        "type":"/type/type"
> > >       },
> > >       {"connect":"delete",
> > >        "id":"/user/szaijan/fantasy_football/projection",
> > >        "type":"/type/type"
> > >       }]
> > >   }]
> > > }
> > >
> > > Does anyone know how to go about this correctly without writing a
> > > script?  The brute force method would be to query all the ids of
> > > instances with the types I want to remove, then remove those types in
> > > a seperate query.  Hopefully, there's a better way.
> >
> > I should point out that if you delete the /type/object/type link of a
> > topic, it does not automatically remove the properties associated with
> > that type.  They become harder to find (you must use the full path ID
> > of the property, and they don't show up in the Web UI), but are still
> > exist and are usable.
> >
> >                                                                Kurt :-)
> >
> >
> >
> > > ____________________________________________________________________
> > >
> > >  Michael Jones, Gamer at Large
> > > szaijan at ternea.com
> > > ____________________________________________________________________
> > > _______________________________________________
> > > Developers mailing list
> > > Developers at freebase.com
> > > http://lists.freebase.com/mailman/listinfo/developers
> >
> 
> 
> -- 
> ____________________________________________________________________
> 
>  Michael Jones, Gamer at Large
> szaijan at ternea.com
> ____________________________________________________________________
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers

From szaijan at gmail.com  Tue Sep  4 21:16:45 2007
From: szaijan at gmail.com (Michael Jones)
Date: Tue, 4 Sep 2007 14:16:45 -0700
Subject: [Developers] Unlinking Objects of Type...
In-Reply-To: <20070904011440.GN31621@metaweb.com>
References: <cbf942d90708222031w5d1f2c6etb72aa322aa72404d@mail.gmail.com>
	<20070823050509.GK14323@metaweb.com>
	<cbf942d90708251535m7523267bx502f594fa97fdf6@mail.gmail.com>
	<20070904011440.GN31621@metaweb.com>
Message-ID: <cbf942d90709041416v64252a7aua9a5b4b3d0b4220e@mail.gmail.com>

Hi Kurt,

> I finally just bit the bullet and converted the python unlinker.py
> > over to Perl.  It works fine.  If anyone wants the code, let me know,
> > but it is not at all polished.
>
> Please post/email this code.  From the perspective of a MQL user,
> there are multiple flavors of topic deletion:


To where shall I post the code?  As mentioned, it is neither polished nor
suitable for use as a module.

The unlink.py python code in the MQL docs does not consider incoming
> links, and may not fully delete a topic.


My preference would have been to truly delete a type I created and all
instances of that type I had created.  In this case, I failed to consider
"incoming" links, just as the unlinker.py code did.

The Freebase UI client has a feature to allow users to flag a topic
> with "mark for delete".  Metaweb has an administrative process to vet
> such marks and execute deletions.  You could do this through the Web
> client if the number of deletions is small enough.


I had ~250 instances to address, so manual deletion was out of the
question.  Flagging for deletion was also not useful as it was requested
that I remove these instances in a set amount of time, and no instances that
I have flagged manually have been addressed at all, to my knowledge.  The
flagging approach works for public types that require moderation to keep
under control, but not so well for a cases where changes need to be made
quickly.

In general, applications built on Freebase data are likely to have some
instances which are not meant to be viewed by users but which are used by
the application as enumerated types, complex links, etc.  These objects
could be built in a separate database, but it's far simpler and more elegant
to only have a single database interface for the application.  Given that,
the ability to change/delete such objects without having to wait for a
moderator to review the request is desirable.

MJ

> On 8/22/07, Kurt Bollacker <kurt at metaweb.com> wrote:
> > > On Wed, Aug 22, 2007 at 08:31:09PM -0700, Michael Jones wrote:
> > > > Hi Folks,
> > > >
> > > > I'm looking for a way to unlink some instances of a specific type
> (I'd
> > > > prefer to delete them entirely, but that another discussion.)  There
> > > > are a lot of them, so manually doing it is out.  I could write a
> > > > script, but it seems this ought to be doable with a single query.
> > > >
> > > > My attempts, which all fail due to uniqueness errors, look like
> this:
> > > >
> > > > {"query":[{
> > > >     type:[{
> > > >       {"id":"/user/szaijan/fantasy_football/statline",
> > > >        "type":"/type/type"
> > > >       },
> > > >       {"connect":"delete",
> > > >        "id":"/user/szaijan/fantasy_football/projection",
> > > >        "type":"/type/type"
> > > >       }]
> > > >   }]
> > > > }
> > > >
> > > > Does anyone know how to go about this correctly without writing a
> > > > script?  The brute force method would be to query all the ids of
> > > > instances with the types I want to remove, then remove those types
> in
> > > > a seperate query.  Hopefully, there's a better way.
> > >
> > > I should point out that if you delete the /type/object/type link of a
> > > topic, it does not automatically remove the properties associated with
> > > that type.  They become harder to find (you must use the full path ID
> > > of the property, and they don't show up in the Web UI), but are still
> > > exist and are usable.
> > >
> > >                                                                Kurt
> :-)
> > >
> > >
> > >
> > > > ____________________________________________________________________
> > > >
> > > >  Michael Jones, Gamer at Large
> > > > szaijan at ternea.com
> > > > ____________________________________________________________________
> > > > _______________________________________________
> > > > Developers mailing list
> > > > Developers at freebase.com
> > > > http://lists.freebase.com/mailman/listinfo/developers
> > >
> >
> >
> > --
> > ____________________________________________________________________
> >
> >  Michael Jones, Gamer at Large
> > szaijan at ternea.com
> > ____________________________________________________________________
> > _______________________________________________
> > Developers mailing list
> > Developers at freebase.com
> > http://lists.freebase.com/mailman/listinfo/developers
>



-- 
____________________________________________________________________

Michael Jones, Gamer at Large
szaijan at ternea.com
____________________________________________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070904/22228c1d/attachment.htm 

From mikelove1 at gmail.com  Tue Sep  4 22:06:56 2007
From: mikelove1 at gmail.com (Mike Love)
Date: Tue, 4 Sep 2007 15:06:56 -0700
Subject: [Developers] symmetric relationships and CVT
Message-ID: <f121137a0709041506n3540e189j2c8fe7fb1c1b72a0@mail.gmail.com>

I have only noticed symmetric relationships that are also their own compound
value types

Adjoin: http://www.freebase.com/view/schema/location/adjoining_relationship
Sibling: http://www.freebase.com/view/schema/people/sibling_relationship

and Dan recommends this for 'Similar Artist':

http://www.freebase.com/view/discuss/music/artist#%239202a8c04000641f8000000005b8a960

is there a way to make symmetric relationships that are easier to build and
query?

thanks!

On 9/1/07, John Giannandrea <jg at metaweb.com> wrote:
>
>
> Mike Love wrote:
> > I wanted to make a symmetrical analog of the influenced /
> > influenced_by relationship, and had to make it a CVT to get this
> > result.
>
> Shooting from the hip I would suggest that 'peer' could be a property
> on influence_node rather than a new type.
>
> > This query below works, although it gives both names in each
> > relationship and the whole query comes back empty if the person had
> > no peers:
>
> You need to expand it further with an optional clause.
>
>      "peers":[{
>        "peers":[{
>          "name":null,
>          "optional":true
>        }]
>
> -jg
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070904/978bc5c5/attachment.htm 

From sarah at metaweb.com  Tue Sep  4 22:40:36 2007
From: sarah at metaweb.com (Sarah Milstein)
Date: Tue, 04 Sep 2007 15:40:36 -0700
Subject: [Developers] Freebase User Group--Sept 19
Message-ID: <46DDDEE4.8010204@metaweb.com>

Exciting news: Our first Freebase user group meeting is set for 
Wednesday, September 19, 6 - 8p in our offices. The event is open to 
Freebase enthusiasts of all stripes and will include pizza and beer.

In addition to exchanging ideas with other Freebasers, the meeting will 
include a couple of juicy (but mercifully short), behind-the-scenes 
presentations.

Our address is below. If you have any questions, please don't hesitate 
to ask!

Metaweb Technologies Inc.
631 Howard Street (@ New Montgomery)
Suite 400
San Francisco, CA 94105
http://maps.google.com/maps?q=631+Howard+Street+san+francisco+california&spn=0.022941,0.040036&hl=en

-- 
Sarah Milstein
Community Manager
Metaweb Technologies, Inc.
sarah at metaweb.com


From niallo at metaweb.com  Tue Sep  4 23:29:50 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 4 Sep 2007 16:29:50 -0700
Subject: [Developers] Weekly sandbox update
Message-ID: <20070904232950.GQ7152@digdug.inet>

Hi,

Data refresh and software update of http://sandbox.freebase.com/
happening over the next few minutes.

-- 
Niall O'Higgins
Software Engineer
Metaweb Technologies, Inc.

From niallo at metaweb.com  Wed Sep  5 00:45:01 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 4 Sep 2007 17:45:01 -0700
Subject: [Developers] Weekly sandbox update
In-Reply-To: <20070904232950.GQ7152@digdug.inet>
References: <20070904232950.GQ7152@digdug.inet>
Message-ID: <20070905004501.GS7152@digdug.inet>

Update complete.

On Tue, Sep 04, 2007 at 04:29:50PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Data refresh and software update of http://sandbox.freebase.com/
> happening over the next few minutes.
> 
> -- 
> Niall O'Higgins
> Software Engineer
> Metaweb Technologies, Inc.
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers

-- 
Niall O'Higgins
Software Engineer
Metaweb Technologies, Inc.

From ninjascience at gmail.com  Wed Sep  5 20:00:32 2007
From: ninjascience at gmail.com (Ryan Miller)
Date: Wed,  5 Sep 2007 21:00:32 +0100 (BST)
Subject: [Developers] Invite from Ryan Miller (ninjascience@gmail.com)
Message-ID: <20070905200032.516052B26A8@mail2.quechup.com>

RyanMiller (ninjascience at gmail.com)
has invited you as a friend on Quechup... 
	...the social networking platform sweeping the globe

Go to: http://quechup.com/join.php/aT0wMDAwMDAwMDA5NDM5OTAzJmM9OTg2ODE%3D
to accept Ryan's invite

You can use Quechup to meet new people, catch up with old friends,
maintain a blog, share videos &amp; photos, chat with other members, play
games, and more.
It's no wonder Quechup is fast becoming 'The Social Networking site to be
on'

Join Ryan and his friends today:
http://quechup.com/join.php/aT0wMDAwMDAwMDA5NDM5OTAzJmM9OTg2ODE%3D
------------------------------------------------------------------
You received this because Ryan Miller (ninjascience at gmail.com) knows and
agreed to invite you. You will only receive one invitation from
ninjascience at gmail.com. Quechup will not spam or sell your email address,
see our privacy policy - http://quechup.com/privacy.php
Go to
http://quechup.com/emailunsubscribe.php/ZW09ZGV2ZWxvcGVyc0BmcmVlYmFzZS5jb20%3D
if you do not wish to receive any more emails from Quechup.
------------------------------------------------------------------
Copyright Quechup.com 2007.
------------------------------------
Go to
http://quechup.com/emailunsubscribe.php/ZW09ZGV2ZWxvcGVyc0BmcmVlYmFzZS5jb20%3D
if you do not wish to receive any more emails from Quechup
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070905/c6260d52/attachment-0001.htm 

From sarah at metaweb.com  Wed Sep  5 23:21:06 2007
From: sarah at metaweb.com (Sarah Milstein)
Date: Wed, 05 Sep 2007 16:21:06 -0700
Subject: [Developers] Invite from Ryan Miller (ninjascience@gmail.com)
In-Reply-To: <20070905200032.516052B26A8@mail2.quechup.com>
References: <20070905200032.516052B26A8@mail2.quechup.com>
Message-ID: <46DF39E2.2050206@metaweb.com>

An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070905/38675eed/attachment.htm 

From ninjascience at gmail.com  Thu Sep  6 00:05:26 2007
From: ninjascience at gmail.com (Ryan Miller)
Date: Wed, 5 Sep 2007 17:05:26 -0700
Subject: [Developers] Invite from Ryan Miller (ninjascience@gmail.com)
In-Reply-To: <46DF39E2.2050206@metaweb.com>
References: <20070905200032.516052B26A8@mail2.quechup.com>
	<46DF39E2.2050206@metaweb.com>
Message-ID: <8037c20b0709051705o5c8576d4sccc42541ebfe52a4@mail.gmail.com>

yes, very sorry all.  If you fell for it, there's some indication that if
you cancel your account before they spam your contact list, it won't
happen.  Not sure though.  It took them about 24 hours to start sending
emails out on my behalf.

On 9/5/07, Sarah Milstein <sarah at metaweb.com> wrote:
>
>  Hey, folks. Ryan isn't spamming us. He sent this note earlier:
>
> You have all probably received an email invite from me to join quechup.com
> .
>
> It's an evil scam!  Don't do it!  I can't believe I fell for it either.
> How many years have I been using the internet? Not enough apparently.
>
> I'm very sorry for the spam!
>
> Thanks,
> Ryan
>
> Ryan Miller wrote:
>
>  [image: Quechup.com] <http://quechup.com/>
>  Trouble viewing this e-mail - click here<http://quechup.com/Qmail/contacts.php/ZW09bmluamFzY2llbmNlQGdtYWlsLmNvbSZpPTAwMDAwMDAwMDk0Mzk5MDMmYz05ODY4MSZmbj1SeWFuJmxuPU1pbGxlciZhbD1uaW5qYXNjaWVuY2UmcGg9bmluamFrYW4zMjgyMTQwMzA5MDcxNjAwNDZjcm9wX3RodW1iLmpwZyZ0bT1kZXZlbG9wZXJzQGZyZWViYXNlLmNvbQ%3D%3D>
>    <http://quechup.com/join.php/aT0wMDAwMDAwMDA5NDM5OTAzJmM9OTg2ODE%3D> *Ryan
> Miller* (ninjascience at gmail.com)
> has invited you as a friend on Quechup...
> ...the social networking platform sweeping the globe
> Click here to accept Ryan's invite<http://quechup.com/join.php/aT0wMDAwMDAwMDA5NDM5OTAzJmM9OTg2ODE%3D>
> You can use Quechup to meet new people, catch up with old friends,
> maintain a blog, share videos & photos, chat with other members, play games,
> and more. It's no wonder Quechup is fast becoming 'The Social Networking
> site to be on'.
> Join Ryan and his friends today:
> http://quechup.com/join.php/aT0wMDAwMDAwMDA5NDM5OTAzJmM9OTg2ODE%3D
>  You received this because Ryan Miller (ninjascience at gmail.com) knows and
> agreed to invite you. You will only receive one invite from
> ninjascience at gmail.com. Quechup will not spam or sell your email address -
> privacy policy <http://quechup.com/privacy.php>. (c) Quechup 2007.
> Click here if you do not wish to receive any more emails from Quechup<http://quechup.com/emailunsubscribe.php/ZW09ZGV2ZWxvcGVyc0BmcmVlYmFzZS5jb20%3D>
>
> ------------------------------
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.comhttp://lists.freebase.com/mailman/listinfo/developers
>
>
> --
> Sarah Milstein
> Community Manager
> Metaweb Technologies, Inc.
> sarah at metaweb.com
>
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070905/4cc55169/attachment.htm 

From robert at metaweb.com  Thu Sep  6 00:19:06 2007
From: robert at metaweb.com (Robert Cook)
Date: Wed, 5 Sep 2007 17:19:06 -0700
Subject: [Developers] Invite from Ryan Miller (ninjascience@gmail.com)
In-Reply-To: <8037c20b0709051705o5c8576d4sccc42541ebfe52a4@mail.gmail.com>
References: <20070905200032.516052B26A8@mail2.quechup.com>
	<46DF39E2.2050206@metaweb.com>
	<8037c20b0709051705o5c8576d4sccc42541ebfe52a4@mail.gmail.com>
Message-ID: <5DDCC034-B070-477C-B1EC-661960AABBD8@metaweb.com>

Now adding the "evil scam" boolean to the "Website" type...

On Sep 5, 2007, at 5:05 PM, Ryan Miller wrote:

> yes, very sorry all.  If you fell for it, there's some indication  
> that if you cancel your account before they spam your contact list,  
> it won't happen.  Not sure though.  It took them about 24 hours to  
> start sending emails out on my behalf.
>
> On 9/5/07, Sarah Milstein <sarah at metaweb.com> wrote:
> Hey, folks. Ryan isn't spamming us. He sent this note earlier:
>
> You have all probably received an email invite from me to join  
> quechup.com.
>
> It's an evil scam!  Don't do it!  I can't believe I fell for it  
> either.  How many years have I been using the internet? Not enough  
> apparently.
>
> I'm very sorry for the spam!
>
> Thanks,
> Ryan
>
> Ryan Miller wrote:
>>
>> Trouble viewing this e-mail - click here
>>
>> Ryan Miller (ninjascience at gmail.com)
>> has invited you as a friend on Quechup...
>> ...the social networking platform sweeping the globe
>> Click here to accept Ryan's invite
>> You can use Quechup to meet new people, catch up with old friends,  
>> maintain a blog, share videos & photos, chat with other members,  
>> play games, and more. It's no wonder Quechup is fast becoming 'The  
>> Social Networking site to be on'.
>> Join Ryan and his friends today:
>> http://quechup.com/join.php/aT0wMDAwMDAwMDA5NDM5OTAzJmM9OTg2ODE%3D
>> You received this because Ryan Miller (ninjascience at gmail.com)  
>> knows and agreed to invite you. You will only receive one invite  
>> from ninjascience at gmail.com. Quechup will not spam or sell your  
>> email address - privacy policy. ? Quechup 2007.
>> Click here if you do not wish to receive any more emails from Quechup
>> _______________________________________________
>> Developers mailing list
>> Developers at freebase.com
>> http://lists.freebase.com/mailman/listinfo/developers
>>
>
> -- 
> Sarah Milstein
> Community Manager
> Metaweb Technologies, Inc.
> sarah at metaweb.com
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers
>
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070905/9804b16c/attachment.htm 

From william at scissor.com  Fri Sep  7 12:15:54 2007
From: william at scissor.com (William Pietri)
Date: Fri, 07 Sep 2007 05:15:54 -0700
Subject: [Developers] Invite from Ryan Miller (ninjascience@gmail.com)
In-Reply-To: <5DDCC034-B070-477C-B1EC-661960AABBD8@metaweb.com>
References: <20070905200032.516052B26A8@mail2.quechup.com>	<46DF39E2.2050206@metaweb.com>	<8037c20b0709051705o5c8576d4sccc42541ebfe52a4@mail.gmail.com>
	<5DDCC034-B070-477C-B1EC-661960AABBD8@metaweb.com>
Message-ID: <46E140FA.5070506@scissor.com>

Robert Cook wrote:
> Now adding the "evil scam" boolean to the "Website" type...

Evil scam? Heck, that's a relationship to a whole set of types. Let me 
know when you get there and I'll pull out my shelf of related books. 
That'd be fun.

Interestingly, the question of what's a scam and what's not is a source 
for a variety of small battles on Wikipedia. A lot of scams work through 
some sort of information asymmetry, and Wikipedia is a big informational 
leveler. For example, take Almeda University:

http://en.wikipedia.org/wiki/Almeda_University

They sell "life experience" degrees, which in my view is partly a scam 
(for the people who buy them naively) and partly a scam enabler (for 
people who buy them knowingly and use them to get jobs, promotions, and 
the like. So there's a big divergence between what Almeda University 
says about itself and what reputable third parties say about them.

Unfortunately for Almeda's operators, the Wikipedia article about them 
is the second Google hit, and that must be putting a crimp in their 
business. They've spent a fair bit of effort trying to pretty up the 
article, and even issued their very first press release. When they got 
chased off, they advertised on a freelancer site for somebody willing to 
go to the Wikipedia article daily and revert them.

Eventually, they apparently gave up, but I'm sure there's quiet editing 
going on somewhere else to preserve somebody's favorite way of fleecing 
the rubes.

William


-- 
William Pietri - william at scissor.com - +1-415-643-1024
Agile consulting, coaching, and development: http://www.scissor.com/
Use your geek-fu to fight poverty: http://www.mifos.org/

From szaijan at gmail.com  Fri Sep  7 18:26:03 2007
From: szaijan at gmail.com (Michael Jones)
Date: Fri, 7 Sep 2007 11:26:03 -0700
Subject: [Developers] Too Many Writes Error
Message-ID: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>

Hi,

I seem to have reached some sort of quota on writes to Freebase and now get
the error :

"result status not okay: /api/status/error; error: Too many writes"

each time I try a write.  Will this be reset eventually?  If so, when.  What
is the limit?

I am curious about the long term intent of the Freebase team for the
database.  Is it intended to be a Wikipedia alternate, with more formalized
and flexible relationships between its instances, or is the value going to
lie in external applications that use the data in Freebase?  If the latter
is part of the plan at all, some thought should be given to allowing private
instances and types.

External applications will often have need of grafting their own types onto
Freebase instances in order to flag them or add meta data that would not be
of interest to the general public browsing Freebase.  You can force
developers of Freebase applications to use private databases for this
information, which seems to be the intent now, but in my opinion that is a
mistake.  If you want to encourage widespread development of Freebase
powered apps, then make them as easy to develop as possible.  Allowing
private types and instances in Freebase will mean a developer can build a
Freebase powered app and only have to worry about an MQL data access model
and only one database for connectivity, greatly simplifying the development
effort.  Further, the object oriented nature of JSON/MQL is much more
intuitive than RDB/SQL and will make such development accessible to a much
wider audience of fledgling developers.
____________________________________________________________________

Michael Jones, Gamer at Large
szaijan at ternea.com
____________________________________________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070907/5d00d710/attachment.htm 

From sm at metaweb.com  Fri Sep  7 18:50:42 2007
From: sm at metaweb.com (Scott Meyer)
Date: Fri, 07 Sep 2007 11:50:42 -0700
Subject: [Developers] Too Many Writes Error
In-Reply-To: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>
References: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>
Message-ID: <46E19D82.1090909@metaweb.com>

Michael Jones wrote:
> Hi,
> 
> I seem to have reached some sort of quota on writes to Freebase and now 
> get the error :
> 
> "result status not okay: /api/status/error; error: Too many writes"
> 
> each time I try a write.  Will this be reset eventually?  If so, when.  
> What is the limit?

10,000 graph primitives per day, I believe.  If you have a large amount
of data to load, the usual procedure is to debug it in the sandbox and
then get an exemption from the 10K limit when performing the upload.

http://www.freebase.com/view/helptopic/%239202a8c04000641f80000000041e92b0

> I am curious about the long term intent of the Freebase team for the 
> database.  Is it intended to be a Wikipedia alternate, with more 
> formalized and flexible relationships between its instances, or is the 
> value going to lie in external applications that use the data in 
> Freebase?

Um, both.  Actually Wikipedia, being basically text, owns the flexible
end of the spectrum, we're more structured.  The advantage of structure
is that both humans and computer programs can find what they want.
Supporting external applications is definitely important to us.

> External applications will often have need of grafting their own types 
> onto Freebase instances in order to flag them or add meta data that 
> would not be of interest to the general public browsing Freebase.

Currently, any user can create a type and use it to type existing
objects.  In fact, we would much rather you do that, than create
duplicate objects.  Is that what you want?

-Scott




From szaijan at gmail.com  Fri Sep  7 20:18:29 2007
From: szaijan at gmail.com (Michael Jones)
Date: Fri, 7 Sep 2007 13:18:29 -0700
Subject: [Developers] Too Many Writes Error
In-Reply-To: <46E19D82.1090909@metaweb.com>
References: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>
	<46E19D82.1090909@metaweb.com>
Message-ID: <cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>

Hi Scott,

On 9/7/07, Scott Meyer <sm at metaweb.com> wrote:
>
> > I seem to have reached some sort of quota on writes to Freebase and now
> > get the error :
> >
> > "result status not okay: /api/status/error; error: Too many writes"
> 10,000 graph primitives per day, I believe.  If you have a large amount
> of data to load, the usual procedure is to debug it in the sandbox and
> then get an exemption from the 10K limit when performing the upload.


It's only about 3-400 records, but I am testing an upload script, and ran it
about 15-20 times yesterday, refreshing exisitng objects.  It doesn't seem
like this should have triggered the limit, but it's possible.  Is there a
reason the limit is on total writes and not on instance creations?

> External applications will often have need of grafting their own types
> > onto Freebase instances in order to flag them or add meta data that
> > would not be of interest to the general public browsing Freebase.
>
> Currently, any user can create a type and use it to type existing
> objects.  In fact, we would much rather you do that, than create
> duplicate objects.  Is that what you want?


That is exactly what I want, and I am doing now for a Fantasy Football Draft
simulator I'm building.

I have grafted the type '/user/szaijan/fantasy_football/player' to all of
the instances of type '/american_football/football_player' that I have data
on.  I've also created types and instances for the objects that will support
the application (League, Owner, Team, et. al.)  In addition, I have created
Projection objects, which hold my estimate of a player's performance and
average fantasy draft position over the course of a given season.  My app
will essentially predict each selection of a fantasy draft based on the
point projections and average draft position, then rank the teams.  A lot of
meta data is used in the sorting, and several enumerated types required near
duplicates of types already in the database.

Ideally, I'd like to make sure no one who isn't running my app can see,
change, or add to this data.  There are two reasons for this.  In the case
of enumerated types, I need to control what is available so I can script for
it.  Allowing arbitrary objects to be added could break my app.  In the case
of meta data, most of it does not make sense outside the context of my app,
and it is thus not suitble for public viewing.

I know that one viewpoint is that such data does not belong on Freebase.  My
contention is that allowing for and supporting such data will be of benefit
in encouraging people to develop Freebase powered apps, for the reasons
specified earlier.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070907/4a65e5bb/attachment.htm 

From tim at metaweb.com  Fri Sep  7 21:20:48 2007
From: tim at metaweb.com (Tim Kientzle)
Date: Fri, 07 Sep 2007 14:20:48 -0700
Subject: [Developers] Too Many Writes Error
In-Reply-To: <cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>
References: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>	<46E19D82.1090909@metaweb.com>
	<cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>
Message-ID: <46E1C0B0.60202@metaweb.com>

> I seem to have reached some sort of quota on writes to Freebase...

We do have such a quota.  It does not limit the number of write
requests that you make, however.  It limits the number of
"primitives" that you create.  If you encountered it, you
did create a lot of instances.  (Depending on exactly how you
did it, "refreshing" something usually does create a complete
new instance that supersedes the old one.)

> It's only about 3-400 records, but I am testing an upload script, and 
> ran it about 15-20 times yesterday, ...

We ask people to test such things on sandbox.freebase.com,
which has no write limits and was built specifically to handle
this kind of testing.

We've developed a pretty standard process for handling robots
and automated applications (we use essentially this same
process even for applications we develop internally):

  1) Test and debug the application against sandbox.freebase.com.

  2) Get a customized account (with an elevated write limit, if
     necessary) to run the application against www.freebase.com.

The second step here (a customized account for the application)
is mostly to help us track overall usage.  (We audit robot usage
a little differently than usage by "real people.")

> Ideally, I'd like to make sure no one who isn't running my app
 > can see, change, or add to this data.

We do not support private data.  Our system has limited
support for restricting writes to critical information, but
no support for restricting read access.  (Freebase itself stores
public account information in Freebase where it can be freely
read but private account information---including passwords and
email addresses, for instance, are stored outside of Freebase.)

Let us know how we can help,

Tim Kientzle

From sm at metaweb.com  Fri Sep  7 23:32:35 2007
From: sm at metaweb.com (Scott Meyer)
Date: Fri, 07 Sep 2007 16:32:35 -0700
Subject: [Developers] Too Many Writes Error
In-Reply-To: <cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>
References: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>	<46E19D82.1090909@metaweb.com>
	<cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>
Message-ID: <46E1DF93.5000607@metaweb.com>

Michael Jones wrote:

> It's only about 3-400 records, but I am testing an upload script, and 
> ran it about 15-20 times yesterday, refreshing exisitng objects.  It 
> doesn't seem like this should have triggered the limit, but it's 
> possible.  Is there a reason the limit is on total writes and not on 
> instance creations?

Because it is hard to make that distinction.  The concern is DOS,
either intentional or because of a broken script.

>      > External applications will often have need of grafting their own
>     types
>      > onto Freebase instances in order to flag them or add meta data that
>      > would not be of interest to the general public browsing Freebase.
> 
>     Currently, any user can create a type and use it to type existing
>     objects.  In fact, we would much rather you do that, than create
>     duplicate objects.  Is that what you want? 
> 
>  
> That is exactly what I want, and I am doing now for a Fantasy Football 
> Draft simulator I'm building.
>  
> I have grafted the type '/user/szaijan/fantasy_football/player' to all 
> of the instances of type '/american_football/football_player' that I 
> have data on.  I've also created types and instances for the objects 
> that will support the application (League, Owner, Team, et. al.)  In 
> addition, I have created Projection objects, which hold my estimate of a 
> player's performance and average fantasy draft position over the course 
> of a given season.  My app will essentially predict each selection of a 
> fantasy draft based on the point projections and average draft position, 
> then rank the teams.  A lot of meta data is used in the sorting, and 
> several enumerated types required near duplicates of types already in 
> the database.
>  
> Ideally, I'd like to make sure no one who isn't running my app can see, 
> change, or add to this data.

Unfortunately, pretty much everything in Freebase is always going to
be visible, that's sort of the point.  Data that isn't public or isn't
connected is of no use to a community.

The problem of publishing data that is, in essence, guaranteed by
the publisher, is one that we're aware of but you're the first person
to bring it up in this context.

Our permission system provides write protection on a per-instance basis
so in this case it might actually be better to create your own instances
of FantasyFootballPlayer and associate them with the actual football
player they represent.

As to the ultimate suitability of Freebase for your application, I think it
depends on the extent to which you can imagine other people eventually using
"your" types and instances.

Regards,

-Scott

From jack.alves at gmail.com  Sat Sep  8 00:00:23 2007
From: jack.alves at gmail.com (Jack Alves)
Date: Fri, 7 Sep 2007 17:00:23 -0700
Subject: [Developers] Need an alternative query for topics with date
	properties
Message-ID: <554723b0709071700q67e28456m170499ea750293b@mail.gmail.com>

Howdy,

I'm still tinkering with a MJT script I wrote awhile ago.

www.tigerlight.com/mjt/day_history.html

The script shows topics that have a property value matching a specific date.
I iterate through the specified years to to get results. I want to add a
feature to show topics with dates any day of a specified year. The query
times out if I use a date range. Is there another way to accomplish what I
want to do?

{ /type/reflect/any_value: [  { link: { master_property: { schema: {
name: *null
*
domain: { id: * null *
name: * null *
}
}
name: * null *
}
}
type: "/type/datetime"
value: * null *
value>=: "2007"
value<: "2008"
}
]
sort: [  "/type/reflect/any_value.link.master_property.name"
 "/type/reflect/any_value.link.master_property.schema.name"
]
id: * null *
name: * null *
/common/topic/image: [  { id: * null *
optional: * true *
limit: 1
}
]
type: [  { name: * null *
}
]
}


Also, I would love to have functionality to query a date like "*-09-07".
Interesting topics in freebase go back about 150 years. Iterating through
each year is a little slow.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070907/22b8f065/attachment-0001.htm 

From rdm at cfcl.com  Sat Sep  8 00:15:51 2007
From: rdm at cfcl.com (Rich Morin)
Date: Fri, 7 Sep 2007 17:15:51 -0700
Subject: [Developers] Metaweb Ruby module
Message-ID: <p0623092ec30797980ec8@[192.168.1.205]>

Shawn Simister said, on July 24:
> I've uploaded what code I've hacked together thus far here
> <http://www.mediafire.com/?fm7sz0n8hbg> but I'd love to get
> some feedback on how to go about building this code into a
> more robust API.  The current code supports basic ActiveRecord
> style queries and pluralization when dealing with the resulting
> objects.  There is no error handling, write support or lazy
> loading of related entities....yet.

I don't see any follow-up to this posting.  Worse, the link seems
to be broken (320 error).  Help?

More generally, I'd love to hear about any progress in building
Rails models for the API, etc.

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From tomgordon at mac.com  Sat Sep  8 00:14:36 2007
From: tomgordon at mac.com (Tom Gordon)
Date: Fri, 7 Sep 2007 20:14:36 -0400
Subject: [Developers] Lifecycle events? (was: Re:  Too Many Writes Error)
In-Reply-To: <46E1DF93.5000607@metaweb.com>
References: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>
	<46E19D82.1090909@metaweb.com>
	<cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>
	<46E1DF93.5000607@metaweb.com>
Message-ID: <A059F5F7-76FD-4EAF-84E2-1DAE4C2581B8@mac.com>


This is an interesting topic, because I also plan to keep additional  
metadata about objects.

My main question is this: is there any plan to expose "lifecycle"  
events?  I'm fairly sure at this point that I will be storing my own  
metadata and cached objects.  It would be nice if I could listen to  
create/delete/merge/split/update events for domains I'm interested  
in.  Otherwise anomalies will creep into my application.  I can work  
around it by tuning the cache parameters and trying various tricks,  
but consuming change notifications would be more interesting to me.   
If there are no plans to implement this then I'll need to start  
considering that in my design.

Michael brought up an interesting idea - if I *could* store private  
metadata in freebase itself, I might consider that.  I doubt it  
though since my metadata relies on heavy server-side processing and  
lots of updates.  My browsers need to talk to my server anyway so  
I'll probably proxy-and-cache for the actual items.

Thanks,
Tom Gordon


On Sep 7, 2007, at 7:32 PM, Scott Meyer wrote:

> Michael Jones wrote:
>
>> It's only about 3-400 records, but I am testing an upload script, and
>> ran it about 15-20 times yesterday, refreshing exisitng objects.  It
>> doesn't seem like this should have triggered the limit, but it's
>> possible.  Is there a reason the limit is on total writes and not on
>> instance creations?
>
> Because it is hard to make that distinction.  The concern is DOS,
> either intentional or because of a broken script.
>
>>> External applications will often have need of grafting their own
>>     types
>>> onto Freebase instances in order to flag them or add meta data that
>>> would not be of interest to the general public browsing Freebase.
>>
>>     Currently, any user can create a type and use it to type existing
>>     objects.  In fact, we would much rather you do that, than create
>>     duplicate objects.  Is that what you want?
>>
>>
>> That is exactly what I want, and I am doing now for a Fantasy  
>> Football
>> Draft simulator I'm building.
>>
>> I have grafted the type '/user/szaijan/fantasy_football/player' to  
>> all
>> of the instances of type '/american_football/football_player' that I
>> have data on.  I've also created types and instances for the objects
>> that will support the application (League, Owner, Team, et. al.)  In
>> addition, I have created Projection objects, which hold my  
>> estimate of a
>> player's performance and average fantasy draft position over the  
>> course
>> of a given season.  My app will essentially predict each selection  
>> of a
>> fantasy draft based on the point projections and average draft  
>> position,
>> then rank the teams.  A lot of meta data is used in the sorting, and
>> several enumerated types required near duplicates of types already in
>> the database.
>>
>> Ideally, I'd like to make sure no one who isn't running my app can  
>> see,
>> change, or add to this data.
>
> Unfortunately, pretty much everything in Freebase is always going to
> be visible, that's sort of the point.  Data that isn't public or isn't
> connected is of no use to a community.
>
> The problem of publishing data that is, in essence, guaranteed by
> the publisher, is one that we're aware of but you're the first person
> to bring it up in this context.
>
> Our permission system provides write protection on a per-instance  
> basis
> so in this case it might actually be better to create your own  
> instances
> of FantasyFootballPlayer and associate them with the actual football
> player they represent.
>
> As to the ultimate suitability of Freebase for your application, I  
> think it
> depends on the extent to which you can imagine other people  
> eventually using
> "your" types and instances.
>
> Regards,
>
> -Scott
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From tim at metaweb.com  Sat Sep  8 00:33:05 2007
From: tim at metaweb.com (Tim Kientzle)
Date: Fri, 07 Sep 2007 17:33:05 -0700
Subject: [Developers] Metaweb Ruby module
In-Reply-To: <p0623092ec30797980ec8@[192.168.1.205]>
References: <p0623092ec30797980ec8@[192.168.1.205]>
Message-ID: <46E1EDC1.7020603@metaweb.com>

Someone here did take a quick stab at a Ruby client
for Freebase a few months ago.   I found the code
and have put a copy at:
    http://dev.mqlx.com/~tim/metaweb-ruby.tgz

This is unofficial and unsupported.  To be honest,
I don't personally know if it works at all.  But maybe
some of the Ruby folks out there will find it a useful
starting point for something better.  We're releasing
it under a BSD license, so you should be able to do
just about anything you like with it.

Tim Kientzle

Rich Morin wrote:
> Shawn Simister said, on July 24:
>> I've uploaded what code I've hacked together thus far here
>> <http://www.mediafire.com/?fm7sz0n8hbg> but I'd love to get
>> some feedback on how to go about building this code into a
>> more robust API.  The current code supports basic ActiveRecord
>> style queries and pluralization when dealing with the resulting
>> objects.  There is no error handling, write support or lazy
>> loading of related entities....yet.
> 
> I don't see any follow-up to this posting.  Worse, the link seems
> to be broken (320 error).  Help?
> 
> More generally, I'd love to hear about any progress in building
> Rails models for the API, etc.
> 
> -r

From jason at metaweb.com  Sat Sep  8 05:25:07 2007
From: jason at metaweb.com (Jason Douglas)
Date: Fri, 7 Sep 2007 22:25:07 -0700
Subject: [Developers] Metaweb Ruby module
In-Reply-To: <p0623092ec30797980ec8@[192.168.1.205]>
References: <p0623092ec30797980ec8@[192.168.1.205]>
Message-ID: <5901A988-33FF-4C79-9BD7-C1DC0690EE58@metaweb.com>

I'm pretty sure I downloaded the files at the time, but unfortunately  
I can't find them either.

I've also heard about a developer in a start-up that built an  
ActiveRecord-style interface to Freebase.  I pinged them about  
posting it to the list at the time and they seemed open to it but I  
never saw anything... I'll try them again.

-jason

On Sep 7, 2007, at 5:15 PM, Rich Morin wrote:

> Shawn Simister said, on July 24:
>> I've uploaded what code I've hacked together thus far here
>> <http://www.mediafire.com/?fm7sz0n8hbg> but I'd love to get
>> some feedback on how to go about building this code into a
>> more robust API.  The current code supports basic ActiveRecord
>> style queries and pluralization when dealing with the resulting
>> objects.  There is no error handling, write support or lazy
>> loading of related entities....yet.
>
> I don't see any follow-up to this posting.  Worse, the link seems
> to be broken (320 error).  Help?
>
> More generally, I'd love to hear about any progress in building
> Rails models for the API, etc.
>
> -r
> -- 
> http://www.cfcl.com/rdm            Rich Morin
> http://www.cfcl.com/rdm/resume     rdm at cfcl.com
> http://www.cfcl.com/rdm/weblog     +1 650-873-7841
>
> Technical editing and writing, programming, and web development
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From jason at metaweb.com  Sat Sep  8 05:41:06 2007
From: jason at metaweb.com (Jason Douglas)
Date: Fri, 7 Sep 2007 22:41:06 -0700
Subject: [Developers] Lifecycle events? (was: Re: Too Many Writes Error)
In-Reply-To: <A059F5F7-76FD-4EAF-84E2-1DAE4C2581B8@mac.com>
References: <cbf942d90709071126s6bd68474yb9ad33905b8bbb3@mail.gmail.com>
	<46E19D82.1090909@metaweb.com>
	<cbf942d90709071318k59815084t567cd77b1d547d8c@mail.gmail.com>
	<46E1DF93.5000607@metaweb.com>
	<A059F5F7-76FD-4EAF-84E2-1DAE4C2581B8@mac.com>
Message-ID: <0AC7D6B7-7B3A-46D6-9A40-37748B117DE4@metaweb.com>

On Sep 7, 2007, at 5:14 PM, Tom Gordon wrote:
> It would be nice if I could listen to
> create/delete/merge/split/update events for domains I'm interested
> in.  Otherwise anomalies will creep into my application.  I can work
> around it by tuning the cache parameters and trying various tricks,
> but consuming change notifications would be more interesting to me.
> If there are no plans to implement this then I'll need to start
> considering that in my design.

We don't have concrete plans for that yet, but how best to  
"syndicate" Freebase data to applications is something we're thinking  
about and we're trying to get a better sense of what approach(es)  
would work most broadly.  Can you say any more about your use case...  
how many domains and types?  how data is accessed, cached, etc.?   
whether your app will write data too? etc...  Whatever you're willing  
to share would be helpful.

-jason

From rdm at cfcl.com  Sat Sep  8 09:42:26 2007
From: rdm at cfcl.com (Rich Morin)
Date: Sat, 8 Sep 2007 02:42:26 -0700
Subject: [Developers] mapping Freebase to KR systems?
Message-ID: <p06230933c3081b25e444@[192.168.1.205]>

Has anyone done anything like mapping Freebase to other
Knowledge Representation systems?  For example, Protege
might be an interesting tool for exploring the Freebase
ontology and PowerLoom might be able to reason about
(subsets of) Freebase's relationships.

Speaking of which, if I wanted to download a full copy
of the Freebase ontology, is there a better way than
crawling through the domains and types?

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From rdm at cfcl.com  Sat Sep  8 09:47:35 2007
From: rdm at cfcl.com (Rich Morin)
Date: Sat, 8 Sep 2007 02:47:35 -0700
Subject: [Developers] mapping Zip codes to other location info?
Message-ID: <p06230934c3081eaab73d@[192.168.1.205]>

Although Freebase recognizes 94066 as a postal code, I
don't see any relations that would link this to "San
Bruno", Latitude: 37.63 / Longitude: -122.43, etc.  Am
I simply looking in the wrong places?

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From philip-freebase at shadowmagic.org.uk  Sat Sep  8 15:04:50 2007
From: philip-freebase at shadowmagic.org.uk (Philip Kendall)
Date: Sat, 8 Sep 2007 16:04:50 +0100
Subject: [Developers] Importing data from World of Spectrum
Message-ID: <20070908150450.GA8772@sphinx.mythic-beasts.com>

One thing I'm considering with Freebase is to upload at least some of
the data from World of Spectrum <http://www.worldofspectrum.org/>. This
contains an essentially complete copy of everything the computer game
type has for every program ever released for the ZX Spectrum (currently
14,688 programs).

Some questions:

1) Is this something which it is a good idea to upload to Freebase? I'd
   obviously test everything in the sandbox first.

2) If so, the thing which is worrying me most is accidently creating
   a large number duplicate topics for things like software houses,
   where the Freebase entry happens to have a slightly different name
   from the name used by World of Spectrum. Two questions here, really:

   * How much should I worry about this? (ie should I apply a tactic
     of upload it all and worry about merges later, or should I spend
     a non-trivial amount of time trying to match things up first?).

   * Does anyone have any suggestions about how to match up names with
     already existing topics?

3) There are a couple of points relating to the preferred usage of the
   types within Freebase, but I'll raise these as discussions "inside"
   Freebase. (If anyone's interested, see my question about the
   designer/developer distinction for 8-bit games on the Computer
   Game type).

4) What other pitfalls am I going to need to avoid? :-)

Cheers,

Phil

-- 
  Philip Kendall <philip at shadowmagic.org.uk>
  http://www.shadowmagic.org.uk/

From narphorium at gmail.com  Sat Sep  8 19:58:19 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Sat, 08 Sep 2007 15:58:19 -0400
Subject: [Developers] Metaweb Ruby module
In-Reply-To: <46E1EDC1.7020603@metaweb.com>
References: <p0623092ec30797980ec8@[192.168.1.205]>
	<46E1EDC1.7020603@metaweb.com>
Message-ID: <46E2FEDB.1070508@gmail.com>

Sorry for not updating the thread in a while. That code which I posted a 
while back was very hacked together and not ready for any serious use. I 
am continuing to work on a Ruby API for Freebase that will allow you to 
access Freebase data in a Rails web app. The new code should be 
available sometime this week so that you can start playing around with it.

The Ruby API will let you make queries sort of like ActiveRecord and the 
results will be Ruby objects that model the Freebase data you queried. 
An example of the syntax might be:

bill_gates = Metaweb::ReadService.find(:first, :type => 
'/people/person', :name => 'Bill Gates')
puts bill_gates.gender.name

Shawn

Tim Kientzle wrote:
> Someone here did take a quick stab at a Ruby client
> for Freebase a few months ago.   I found the code
> and have put a copy at:
>     http://dev.mqlx.com/~tim/metaweb-ruby.tgz
>
> This is unofficial and unsupported.  To be honest,
> I don't personally know if it works at all.  But maybe
> some of the Ruby folks out there will find it a useful
> starting point for something better.  We're releasing
> it under a BSD license, so you should be able to do
> just about anything you like with it.
>
> Tim Kientzle
>
> Rich Morin wrote:
>   
>> Shawn Simister said, on July 24:
>>     
>>> I've uploaded what code I've hacked together thus far here
>>> <http://www.mediafire.com/?fm7sz0n8hbg> but I'd love to get
>>> some feedback on how to go about building this code into a
>>> more robust API.  The current code supports basic ActiveRecord
>>> style queries and pluralization when dealing with the resulting
>>> objects.  There is no error handling, write support or lazy
>>> loading of related entities....yet.
>>>       
>> I don't see any follow-up to this posting.  Worse, the link seems
>> to be broken (320 error).  Help?
>>
>> More generally, I'd love to hear about any progress in building
>> Rails models for the API, etc.
>>
>> -r
>>     
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers
>
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070908/a96a0dea/attachment-0001.htm 

From narphorium at gmail.com  Sat Sep  8 19:58:37 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Sat, 08 Sep 2007 15:58:37 -0400
Subject: [Developers] mapping Freebase to KR systems?
In-Reply-To: <p06230933c3081b25e444@[192.168.1.205]>
References: <p06230933c3081b25e444@[192.168.1.205]>
Message-ID: <46E2FEED.9090202@gmail.com>

I have been developing a tool in Ruby which translates Freebase types 
into OWL ontologies that are viewable in tools like Protege. The results 
have been encouraging and I would like to try to start mapping Freebase 
topics to RDF instances using these OWL ontologies. I have attached a 
sample OWL ontology for the /people/person type and I can make the code 
available to you as soon as I have the Freebase Ruby API cleaned up.

Shawn

Rich Morin wrote:
> Has anyone done anything like mapping Freebase to other
> Knowledge Representation systems?  For example, Protege
> might be an interesting tool for exploring the Freebase
> ontology and PowerLoom might be able to reason about
> (subsets of) Freebase's relationships.
>
> Speaking of which, if I wanted to download a full copy
> of the Freebase ontology, is there a better way than
> crawling through the domains and types?
>
> -r
>   

-------------- next part --------------
A non-text attachment was scrubbed...
Name: person.owl
Type: text/xml
Size: 5063 bytes
Desc: not available
Url : http://lists.freebase.com/pipermail/developers/attachments/20070908/1a4e9a08/attachment.bin 

From jg at metaweb.com  Sat Sep  8 20:09:29 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sat, 8 Sep 2007 13:09:29 -0700
Subject: [Developers] mapping Freebase to KR systems?
In-Reply-To: <p06230933c3081b25e444@[192.168.1.205]>
References: <p06230933c3081b25e444@[192.168.1.205]>
Message-ID: <7712A8DA-F32B-4159-8185-650AC5889C0D@metaweb.com>


Rich Morin wrote:
> Has anyone done anything like mapping Freebase to other
> Knowledge Representation systems?  For example, Protege
> might be an interesting tool for exploring the Freebase
> ontology and PowerLoom might be able to reason about
> (subsets of) Freebase's relationships.

Freebase is more of a database than an ontology.
There is no direct mapping from our types to, say, OWL representations.
This is not to say it cant be done, just that it is not there yet.

A different thing would be mapping between the topics and
ontology/databases like Cyc, SUMO or WordNet.
This is possible and desirable, it requires reconciliation though.

However we do plan to provide an RDF export of the freebase data
which should be compatible with some SW tools.

> Speaking of which, if I wanted to download a full copy
> of the Freebase ontology, is there a better way than
> crawling through the domains and types?

Currently you have to export all the types with MQL.
You can also explore with a tool like
	http://www.mjtemplate.org/examples/schemas/

We do plan to provide 'data dumps' which would include
the schemas in some flattened form.

-jg


From jg at metaweb.com  Sat Sep  8 20:19:49 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sat, 8 Sep 2007 13:19:49 -0700
Subject: [Developers] Importing data from World of Spectrum
In-Reply-To: <20070908150450.GA8772@sphinx.mythic-beasts.com>
References: <20070908150450.GA8772@sphinx.mythic-beasts.com>
Message-ID: <AAFDCFA8-28BC-4746-B11A-9B5528EC1E9B@metaweb.com>


Philip Kendall wrote:
> 1) Is this something which it is a good idea to upload to Freebase?  
> I'd
>    obviously test everything in the sandbox first.

yes, this would be a fine candidate for freebase, assuming that the
data you plan to upload is compatible with the Creative Commons  
Attribution license.

We have some general guidelines here:
http://www.freebase.com/view/?id=%239202a8c04000641f80000000010c2cf1

> 2) If so, the thing which is worrying me most is accidently creating
>    a large number duplicate topics for things like software houses,
>    where the Freebase entry happens to have a slightly different name
>    from the name used by World of Spectrum. Two questions here,  
> really:
>    * How much should I worry about this? (ie should I apply a tactic
>      of upload it all and worry about merges later, or should I spend
>      a non-trivial amount of time trying to match things up first?).

We have some tools in house (that we plan to release when they are
further along) that can help with this.  If you would like to contact  
me off
list we can probably help with some of this reconciliation.

-jg


From rictic at gmail.com  Sun Sep  9 01:24:45 2007
From: rictic at gmail.com (Peter Burns)
Date: Sat, 8 Sep 2007 18:24:45 -0700
Subject: [Developers] Any good sources of license-compatible data that need
	importing into Freebase?
Message-ID: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>

Hey all,

I'm a coder with some time on his hands, a penchant for screen-scraping and
data transforming, and a deep ideological agreement with putting data into a
convenient form for easy access.

Know databases that need to be imported into Freebase?  I'd like to
contribute.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070908/db0244d0/attachment.htm 

From ptomblin at xcski.com  Sun Sep  9 01:28:47 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Sat, 8 Sep 2007 21:28:47 -0400
Subject: [Developers] Any good sources of license-compatible data that
	need importing into Freebase?
In-Reply-To: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
References: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
Message-ID: <20070909012847.GA18685@allhats2.xcski.com>

Quoting Peter Burns (rictic at gmail.com):
> Hey all,
> 
> I'm a coder with some time on his hands, a penchant for screen-scraping and
> data transforming, and a deep ideological agreement with putting data into a
> convenient form for easy access.
> 
> Know databases that need to be imported into Freebase?  I'd like to
> contribute.

Do you know anything about scraping data from PDF files?  Many countries
publish their aeronautical data only in PDF formats.  I used to get this
data from the DAFIF (Digital Aeronautical Flight Information File), but
the US Government that published it withdrew it from public use.


-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
"Guns don't shoot people.  Vice Presidents shoot people."

From rictic at gmail.com  Sun Sep  9 01:38:07 2007
From: rictic at gmail.com (Peter Burns)
Date: Sat, 8 Sep 2007 18:38:07 -0700
Subject: [Developers] Any good sources of license-compatible data that
	need importing into Freebase?
In-Reply-To: <20070909012847.GA18685@allhats2.xcski.com>
References: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
	<20070909012847.GA18685@allhats2.xcski.com>
Message-ID: <2f9bef730709081838h5c792049n89ccaa14e44cd02b@mail.gmail.com>

Do you have a link to one of them?  I've never tried to parse a PDF
document, but it looks like there are a number of tools available in a few
languages I'm competent in.

On 9/8/07, Paul Tomblin <ptomblin at xcski.com> wrote:
>
> Quoting Peter Burns (rictic at gmail.com):
> > Hey all,
> >
> > I'm a coder with some time on his hands, a penchant for screen-scraping
> and
> > data transforming, and a deep ideological agreement with putting data
> into a
> > convenient form for easy access.
> >
> > Know databases that need to be imported into Freebase?  I'd like to
> > contribute.
>
> Do you know anything about scraping data from PDF files?  Many countries
> publish their aeronautical data only in PDF formats.  I used to get this
> data from the DAFIF (Digital Aeronautical Flight Information File), but
> the US Government that published it withdrew it from public use.
>
>
> --
> Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
> "Guns don't shoot people.  Vice Presidents shoot people."
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070908/de0fa6b3/attachment.htm 

From narphorium at gmail.com  Sun Sep  9 01:41:28 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Sat, 08 Sep 2007 21:41:28 -0400
Subject: [Developers] Any good sources of license-compatible data that
 need	importing into Freebase?
In-Reply-To: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
References: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
Message-ID: <46E34F48.6010409@gmail.com>

I'd really like to see a crawler for keeping track of the relationships 
between professional athletes and the teams that they play for. The 
information is readily available from a wide range of sources but 
changes periodically throughout the year depending on which sport you're 
following. I think that updating these relationships by hand would be 
rather tedious task for a human but can easily done by a screen-scraping 
crawler. Just having the NFL, NBA, MLB, NHL and PGA would provide a lot 
of opportunity for mashups.

Shawn

Peter Burns wrote:
> Hey all,
>
> I'm a coder with some time on his hands, a penchant for 
> screen-scraping and data transforming, and a deep ideological 
> agreement with putting data into a convenient form for easy access.
>
> Know databases that need to be imported into Freebase?  I'd like to 
> contribute.
> ------------------------------------------------------------------------
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070908/8b8ad70c/attachment.htm 

From rdm at cfcl.com  Sun Sep  9 01:48:52 2007
From: rdm at cfcl.com (Rich Morin)
Date: Sat, 8 Sep 2007 18:48:52 -0700
Subject: [Developers] keeping track of Freebase content/prospects
In-Reply-To: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
References: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
Message-ID: <p0623093ac30900619ef6@[192.168.1.205]>

I'd like there to be a wiki (or whatever) that would aid
in tracking existing and/or prospective Freebase content.
It should contain a description, license notes, location
(e.g., URL), etc.

Using this as a communal resource, we could eliminate (or
at least reduce) wasted effort in obtaining and keeping
Freebase content up to date.

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From jason at metaweb.com  Sun Sep  9 02:41:37 2007
From: jason at metaweb.com (Jason Douglas)
Date: Sat, 8 Sep 2007 19:41:37 -0700
Subject: [Developers] keeping track of Freebase content/prospects
In-Reply-To: <p0623093ac30900619ef6@[192.168.1.205]>
References: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
	<p0623093ac30900619ef6@[192.168.1.205]>
Message-ID: <DED966A7-9FC8-4CD7-B787-624267D5B22A@metaweb.com>

Great idea!  That should be easy enough to track in Freebase  
itself... something similar was done with the "Desired Application"  
type.

I'm cc'ing some folks from our data acquisition team who might have a  
good list for starters.

-jason


On Sep 8, 2007, at 6:48 PM, Rich Morin wrote:

> I'd like there to be a wiki (or whatever) that would aid
> in tracking existing and/or prospective Freebase content.
> It should contain a description, license notes, location
> (e.g., URL), etc.
>
> Using this as a communal resource, we could eliminate (or
> at least reduce) wasted effort in obtaining and keeping
> Freebase content up to date.
>
> -r
> -- 
> http://www.cfcl.com/rdm            Rich Morin
> http://www.cfcl.com/rdm/resume     rdm at cfcl.com
> http://www.cfcl.com/rdm/weblog     +1 650-873-7841
>
> Technical editing and writing, programming, and web development
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From rdm at cfcl.com  Sun Sep  9 05:15:26 2007
From: rdm at cfcl.com (Rich Morin)
Date: Sat, 8 Sep 2007 22:15:26 -0700
Subject: [Developers] keeping track of Freebase content/prospects
In-Reply-To: <DED966A7-9FC8-4CD7-B787-624267D5B22A@metaweb.com>
References: <2f9bef730709081824h4b462519hb3a9eacc32267c15@mail.gmail.com>
	<p0623093ac30900619ef6@[192.168.1.205]>
	<DED966A7-9FC8-4CD7-B787-624267D5B22A@metaweb.com>
Message-ID: <p0623093ec309301dcf28@[192.168.1.205]>

At 19:41 -0700 9/8/07, Jason Douglas wrote:
> Great idea!  That should be easy enough to track in Freebase
> itself... something similar was done with the "Desired Application"
> type.
>
> I'm cc'ing some folks from our data acquisition team who might have
> a good list for starters.

I wasn't sure whether it made sense to use Freebase to track the info,
in part because some of it is going to be pretty fuzzy (e.g., comments
on whether and why a database might be useful).

However, if you're taking that path, be sure to brainstorm a bit on what
kinds of information could be saved in a way that would allow Freebase
to make useful discriminations.  For example, the date/time of the last
revision and the most recent upload might prove handy...

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From skud at infotrope.net  Sun Sep  9 07:25:38 2007
From: skud at infotrope.net (Kirrily Robert)
Date: Sun, 9 Sep 2007 17:25:38 +1000
Subject: [Developers] Ontological forum?
Message-ID: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>

This is probably one for Jeff.  Where's the best place to discuss
matters of ontology/types/data structures?  I presume it's not this
mailing list, but it seems to me as if there should be a sibling list,
or at least some equivalent forum, for such discussions.

K.

-- 
Kirrily Robert
skud at infotrope.net
http://infotrope.net

From narphorium at gmail.com  Sun Sep  9 07:45:12 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Sun, 09 Sep 2007 03:45:12 -0400
Subject: [Developers] Ontological forum?
In-Reply-To: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>
Message-ID: <46E3A488.2000301@gmail.com>

 From what I've seen, the discussion areas in the type descriptions are 
often where people discuss those sorts of things.

Kirrily Robert wrote:
> This is probably one for Jeff.  Where's the best place to discuss
> matters of ontology/types/data structures?  I presume it's not this
> mailing list, but it seems to me as if there should be a sibling list,
> or at least some equivalent forum, for such discussions.
>
> K.
>
>   


From skud at infotrope.net  Sun Sep  9 08:08:05 2007
From: skud at infotrope.net (Kirrily Robert)
Date: Sun, 9 Sep 2007 18:08:05 +1000
Subject: [Developers] Ontological forum?
In-Reply-To: <46E3A488.2000301@gmail.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>
	<46E3A488.2000301@gmail.com>
Message-ID: <b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>

On 9/9/07, Shawn Simister <narphorium at gmail.com> wrote:
>  From what I've seen, the discussion areas in the type descriptions are
> often where people discuss those sorts of things.

That's all very well for existing, clearly defined types, but what if
one wishes to have a broad-ranging philosophical discussion that
encompasses more than one type, or a theoretical one covering a type
that doesn't yet exist?

K.

-- 
Kirrily Robert
skud at infotrope.net
http://infotrope.net

From narphorium at gmail.com  Sun Sep  9 08:19:52 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Sun, 09 Sep 2007 04:19:52 -0400
Subject: [Developers] Ontological forum?
In-Reply-To: <b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>	
	<46E3A488.2000301@gmail.com>
	<b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>
Message-ID: <46E3ACA8.6030801@gmail.com>

Kirrily Robert wrote:
> On 9/9/07, Shawn Simister <narphorium at gmail.com> wrote:
>   
>>  From what I've seen, the discussion areas in the type descriptions are
>> often where people discuss those sorts of things.
>>     
>
> That's all very well for existing, clearly defined types, but what if
> one wishes to have a broad-ranging philosophical discussion that
> encompasses more than one type, or a theoretical one covering a type
> that doesn't yet exist?
>
> K.
>   
I agree that the discussion pages really limit the range and the 
audience of such discussions but I haven't found anything else yet that 
covers those sorts of topics.
I suppose a mailing list for information architects and domain 
specialist would be useful.

Shawn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070909/a0a51028/attachment.htm 

From skud at infotrope.net  Sun Sep  9 09:52:08 2007
From: skud at infotrope.net (Kirrily Robert)
Date: Sun, 9 Sep 2007 19:52:08 +1000
Subject: [Developers] Ontological forum?
In-Reply-To: <b08e1490709090251m3a82562bm8e355b2851ed2e8d@mail.gmail.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>
	<46E3A488.2000301@gmail.com>
	<b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>
	<46E3ACA8.6030801@gmail.com>
	<b08e1490709090251m3a82562bm8e355b2851ed2e8d@mail.gmail.com>
Message-ID: <b08e1490709090252j534c4e0co822533af2ab9dca8@mail.gmail.com>

Dammit, I meant to send this to the list.

On 9/9/07, Kirrily Robert <skud at infotrope.net> wrote:
> On 9/9/07, Shawn Simister <narphorium at gmail.com> wrote:
> >  I agree that the discussion pages really limit the range and the audience
> > of such discussions but I haven't found anything else yet that covers those
> > sorts of topics.
> >  I suppose a mailing list for information architects and domain specialist
> > would be useful.
>
> Yes please!
>
> Meanwhile, I'll ask my question here.  I found myself wondering about
> abbreviations this evening.  Australian states (for which I have
> created a user type,
> http://www.freebase.com/view/filter/user/skud/default_domain/australian_state)
> have an official abbreviated form of their name.  For instance,
> Victoria is VIC and New South Wales is NSW.  Currently these are just
> listed in the "Also known as" field, but it seems to me that it would
> also be useful to have a type "Abbreviation" in its own right.  Where
> one abbreviation can mean multiple things, it could link to all those
> terms.  But I'm sure there are complications to this that I haven't
> foreseen, so I wanted to open up discussion on it.
>
> K.
>
> --
> Kirrily Robert
> skud at infotrope.net
> http://infotrope.net
>


-- 
Kirrily Robert
skud at infotrope.net
http://infotrope.net

From ksrinivs at gmail.com  Sun Sep  9 14:00:56 2007
From: ksrinivs at gmail.com (Kavitha Srinivas)
Date: Sun, 9 Sep 2007 10:00:56 -0400
Subject: [Developers] mapping Freebase to KR systems?
In-Reply-To: <7712A8DA-F32B-4159-8185-650AC5889C0D@metaweb.com>
References: <p06230933c3081b25e444@[192.168.1.205]>
	<7712A8DA-F32B-4159-8185-650AC5889C0D@metaweb.com>
Message-ID: <C2EF639E-E67B-439A-8702-2FF7FAB1DDF7@gmail.com>

John,
    Any idea about when we can get data dumps from Freebase?  I'm  
interested in RDF dumps like Shawn, but will settle for any other  
format, if its easier.
-Kavitha

On Sep 8, 2007, at 4:09 PM, John Giannandrea wrote:

>
> Rich Morin wrote:
>> Has anyone done anything like mapping Freebase to other
>> Knowledge Representation systems?  For example, Protege
>> might be an interesting tool for exploring the Freebase
>> ontology and PowerLoom might be able to reason about
>> (subsets of) Freebase's relationships.
>
> Freebase is more of a database than an ontology.
> There is no direct mapping from our types to, say, OWL  
> representations.
> This is not to say it cant be done, just that it is not there yet.
>
> A different thing would be mapping between the topics and
> ontology/databases like Cyc, SUMO or WordNet.
> This is possible and desirable, it requires reconciliation though.
>
> However we do plan to provide an RDF export of the freebase data
> which should be compatible with some SW tools.
>
>> Speaking of which, if I wanted to download a full copy
>> of the Freebase ontology, is there a better way than
>> crawling through the domains and types?
>
> Currently you have to export all the types with MQL.
> You can also explore with a tool like
> 	http://www.mjtemplate.org/examples/schemas/
>
> We do plan to provide 'data dumps' which would include
> the schemas in some flattened form.
>
> -jg
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From esessions at gmail.com  Sun Sep  9 14:38:03 2007
From: esessions at gmail.com (Eric Sessions)
Date: Sun, 9 Sep 2007 10:38:03 -0400
Subject: [Developers] "Unique query may have at most one result" Error
Message-ID: <efecaf730709090738hcef4515tfc9098360a92ba1e@mail.gmail.com>

Hi,

I am sorry for asking such a basic question.  I am having a lot of trouble
with a query (trying to run it in the query editor.) It runs great for the
first three tries then fails at the fourth try. I am using cursors and
taking the cursor from the previous result and using it at the next query. I
am really lost as to what I am doing wrong.

The query I am trying to run is
{
"cursor":true,
"query":[{
"*":[],
"/business/retail_location/address":{
"citytown":[],
"postal_code":[],
"street_address":[],
"street_address_2":[]
},
"name":[],
"type":"/dining/restaurant"
}]
}

The error message I get is
"Unique query may have at most one result. Got 56". It seems like on the
fourth attempt, it is losing the outside data like restaurant id etc and
only returning the address information only. What am I doing wrong?

Thanks,

-Eric
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070909/699f27d1/attachment.htm 

From skud at infotrope.net  Sun Sep  9 15:33:53 2007
From: skud at infotrope.net (Kirrily Robert)
Date: Mon, 10 Sep 2007 01:33:53 +1000
Subject: [Developers] Metadata about types
Message-ID: <b08e1490709090833w3cc404f3w80cab53fa12ca565@mail.gmail.com>

How can I find out the following information about types?

* who created it
* when they created it
* public vs private

I hoped that this might tell me the answer:

{
  "query":[{
    "id":"/type/type",
    "properties":[],
    "type":"/type/type"
  }]
}

... but it turns out it doesn't.

K.

-- 
Kirrily Robert
skud at infotrope.net
http://infotrope.net

From robert at metaweb.com  Sun Sep  9 17:06:27 2007
From: robert at metaweb.com (Robert Cook)
Date: Sun, 9 Sep 2007 10:06:27 -0700
Subject: [Developers] Ontological forum?
In-Reply-To: <b08e1490709090252j534c4e0co822533af2ab9dca8@mail.gmail.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>
	<46E3A488.2000301@gmail.com>
	<b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>
	<46E3ACA8.6030801@gmail.com>
	<b08e1490709090251m3a82562bm8e355b2851ed2e8d@mail.gmail.com>
	<b08e1490709090252j534c4e0co822533af2ab9dca8@mail.gmail.com>
Message-ID: <7D05FAD7-E54F-40D2-9E31-92E95B2A3168@metaweb.com>

Kirrily and Shawn --

Right now the best place to discuss new types that don't correspond  
to existing domains is in the discussion attached to the "Data" page  
in the top navbar.  There are already some lively debates here:

   http://www.freebase.com/view/discuss/% 
239202a8c04000641f8000000003c1b16b

Specifically for your question about state abbreviations, that would  
best be discussed in the Location domain:

   http://www.freebase.com/view/domain/location

or even the "Administrative division" type, which would be an  
included type of your Australian State type, and already includes the  
ISO and FIPS codes (which may satisfy your abbreviation requirement):

   http://www.freebase.com/view/schema/location/administrative_division

I actually think a "data modeler" mailing list is a good idea, but  
I'll have to check with Sarah and others to see if that fits longer  
term plans.  Generally, I think that there's not enough documentation  
on data modeling and having a more general question forum would be a  
good idea, either as a list or as a discussion area on the site.  One  
reasonable place would be the discussion attached to the "Data  
Modeling Guide" (if we made it more prominent):

   http://www.freebase.com/view/?id=%239202a8c04000641f80000000045637f3

Robert

On Sep 9, 2007, at 2:52 AM, Kirrily Robert wrote:

> Dammit, I meant to send this to the list.
>
> On 9/9/07, Kirrily Robert <skud at infotrope.net> wrote:
>> On 9/9/07, Shawn Simister <narphorium at gmail.com> wrote:
>>>  I agree that the discussion pages really limit the range and the  
>>> audience
>>> of such discussions but I haven't found anything else yet that  
>>> covers those
>>> sorts of topics.
>>>  I suppose a mailing list for information architects and domain  
>>> specialist
>>> would be useful.
>>
>> Yes please!
>>
>> Meanwhile, I'll ask my question here.  I found myself wondering about
>> abbreviations this evening.  Australian states (for which I have
>> created a user type,
>> http://www.freebase.com/view/filter/user/skud/default_domain/ 
>> australian_state)
>> have an official abbreviated form of their name.  For instance,
>> Victoria is VIC and New South Wales is NSW.  Currently these are just
>> listed in the "Also known as" field, but it seems to me that it would
>> also be useful to have a type "Abbreviation" in its own right.  Where
>> one abbreviation can mean multiple things, it could link to all those
>> terms.  But I'm sure there are complications to this that I haven't
>> foreseen, so I wanted to open up discussion on it.
>>
>> K.
>>
>> --
>> Kirrily Robert
>> skud at infotrope.net
>> http://infotrope.net
>>
>
>
> -- 
> Kirrily Robert
> skud at infotrope.net
> http://infotrope.net
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From narphorium at gmail.com  Sun Sep  9 17:28:14 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Sun, 09 Sep 2007 13:28:14 -0400
Subject: [Developers] Metadata about types
In-Reply-To: <b08e1490709090833w3cc404f3w80cab53fa12ca565@mail.gmail.com>
References: <b08e1490709090833w3cc404f3w80cab53fa12ca565@mail.gmail.com>
Message-ID: <46E42D2E.4030706@gmail.com>

Kirrily Robert wrote:
> How can I find out the following information about types?
>
> * who created it
> * when they created it
> * public vs private
>
> I hoped that this might tell me the answer:
>
> {
>   "query":[{
>     "id":"/type/type",
>     "properties":[],
>     "type":"/type/type"
>   }]
> }
>
> ... but it turns out it doesn't.
>
> K.
>   
That one confused me as well. You need to use the type name as the id. 
For example:

{
  "query":[{
    "*":[{}],
    "id":"/people/person",
    "properties":[{}],
    "type":"/type/type"
  }]
}

This will give you the first 2 requirements. Not too sure about public 
vs. private. As far as I can tell thats not exposed in the API.

From jg at metaweb.com  Sun Sep  9 21:05:16 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sun, 9 Sep 2007 14:05:16 -0700
Subject: [Developers] Metadata about types
In-Reply-To: <b08e1490709090833w3cc404f3w80cab53fa12ca565@mail.gmail.com>
References: <b08e1490709090833w3cc404f3w80cab53fa12ca565@mail.gmail.com>
Message-ID: <52A6B990-7E9F-4A0E-B784-6C42777887C8@metaweb.com>


Kirrily Robert wrote:
> How can I find out the following information about types?
> * who created it
> * when they created it
> * public vs private

There are a lot of types, so this query limits it by domain (in this  
case /location)

{
   "query":[{
     "/freebase/type_hints/included_types":[],
     "/freebase/type_profile/published":null,
     "creator":null,
     "domain":"/location",
     "id":null,
     "limit":0,
     "sort":"-timestamp",
     "timestamp":null,
     "type":"/type/type"
   }]
}

-jg




From jg at metaweb.com  Sun Sep  9 21:13:50 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sun, 9 Sep 2007 14:13:50 -0700
Subject: [Developers] "Unique query may have at most one result" Error
In-Reply-To: <efecaf730709090738hcef4515tfc9098360a92ba1e@mail.gmail.com>
References: <efecaf730709090738hcef4515tfc9098360a92ba1e@mail.gmail.com>
Message-ID: <AF0D1884-82C2-4FBA-938F-96FD6066A106@metaweb.com>


Eric Sessions wrote:
> I am sorry for asking such a basic question.

There are no basic questions, this is exactly the right place to ask  
for help.

>   I am having a lot of trouble with a query (trying to run it in  
> the query editor.) It runs great for the first three tries then  
> fails at the fourth try. I am using cursors and taking the cursor  
> from the previous result and using it at the next query. I am  
> really lost as to what I am doing wrong.

You came across a restaurant with more than one address,
specifically http://www.freebase.com/view/village_inn

Put a [] clause around address like this:

{
   "cursor":true,
   "query":[{
     "*":null,
     "/business/retail_location/address":[{
       "citytown":[],
       "postal_code":[],
       "street_address":[],
       "street_address_2":[]
     }],
     "name":[],
     "type":"/dining/restaurant"
   }]
}


Uniqueness errors are discussed in more depth here:

http://www.freebase.com/view/helptopic?id=% 
239202a8c04000641f800000000544e13e#id2612623

-jg


From jg at metaweb.com  Sun Sep  9 21:23:08 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sun, 9 Sep 2007 14:23:08 -0700
Subject: [Developers] Metadata about types
In-Reply-To: <46E42D2E.4030706@gmail.com>
References: <b08e1490709090833w3cc404f3w80cab53fa12ca565@mail.gmail.com>
	<46E42D2E.4030706@gmail.com>
Message-ID: <9E255592-8E8F-4386-8D05-79CBA47B107F@metaweb.com>


Shawn Simister wrote:
> This will give you the first 2 requirements. Not too sure about public
> vs. private. As far as I can tell thats not exposed in the API.

It is often a little hard to figure out what properties are in use for a
specific object.  One practical tool that shows you whats going on
under the hood is our explore view.  For example:

     http://www.freebase.com/view/explore/location/continent

this shows you the programmer property ids being used
at the lowest layer.  If you click on Africa for example you can
see how languages are modeled and all the other assertions
for that instance.

(you can get to the explore view from any topic page by typing F8
and clicking on 'Explore'.

Hope that is helpful in understanding the data better.

-jg


From jg at metaweb.com  Sun Sep  9 21:26:46 2007
From: jg at metaweb.com (John Giannandrea)
Date: Sun, 9 Sep 2007 14:26:46 -0700
Subject: [Developers] mapping Freebase to KR systems?
In-Reply-To: <C2EF639E-E67B-439A-8702-2FF7FAB1DDF7@gmail.com>
References: <p06230933c3081b25e444@[192.168.1.205]>
	<7712A8DA-F32B-4159-8185-650AC5889C0D@metaweb.com>
	<C2EF639E-E67B-439A-8702-2FF7FAB1DDF7@gmail.com>
Message-ID: <6376A86D-A0A2-4B4A-82DF-FEDB2422729E@metaweb.com>


Kavitha Srinivas wrote:
>    Any idea about when we can get data dumps from Freebase?  I'm  
> interested in RDF dumps like Shawn, but will settle for any other  
> format, if its easier.

We are committed to doing it, but are still trying to figure out the  
formats
and timing.  We are still in alpha and have not finalized the details.

You can dump pretty much everything with MQL, and we have some folks
who are experimenting with that against sandbox.freebase.com.

We will make an announcement later when we have figured out the
details of what we are supporting long term.

-jg


From skud at infotrope.net  Mon Sep 10 05:03:07 2007
From: skud at infotrope.net (Kirrily Robert)
Date: Mon, 10 Sep 2007 15:03:07 +1000
Subject: [Developers] Descriptions
Message-ID: <b08e1490709092203t5d7bdd8ar5c629d2fbb2e7300@mail.gmail.com>

How does one access the description of an object via MQL?

Eg. for something like:

{
  "query":{
    "*":null,
    "name":"Influence Walker",
    "type":"/freebase/metaweb_application"
  }
}

I'd expect to find somewhere in there, the description "A la Digg Labs
BigSpy, Influence Walker displays weighted lines..." as seen at
http://www.freebase.com/view/%239202a8c04000641f8000000005c724e8

K.

-- 
Kirrily Robert
skud at infotrope.net
http://infotrope.net

From crism at metaweb.com  Mon Sep 10 05:15:14 2007
From: crism at metaweb.com (Christopher R. Maden)
Date: Mon, 10 Sep 2007 01:15:14 -0400 (EDT)
Subject: [Developers] Descriptions
In-Reply-To: <b08e1490709092203t5d7bdd8ar5c629d2fbb2e7300@mail.gmail.com>
Message-ID: <10231638.116661189401314503.JavaMail.root@h00215>

Kirrily Robert <skud at infotrope.net> wrote:
> How does one access the description of an object via MQL?

Hey, Kirrily,

The description is attached via the article property:

{
  "query":{
    "/common/topic/article":[{}],
    "name":"Influence Walker",
    "type":"/freebase/metaweb_application"
  }
}

> {
>   "query":{
>     "*":null,
>     "name":"Influence Walker",
>     "type":"/freebase/metaweb_application"
>   }
> }

When you specify the type in a query, the "*" wildcard only picks up properties of /type/object and the specified type; since the article is a property of /common/topic, it doesn?t show up there.

You won?t actually see the text of the description in that query response, since articles are accessed via the document service, but the query response points to the relevant document.

~Chris
-- 
Christopher R. Maden
Data Architect
Metaweb Technologies, Inc.
<URL: http://www.metaweb.com/ >

From jamie at metaweb.com  Mon Sep 10 17:11:33 2007
From: jamie at metaweb.com (Jamie Taylor)
Date: Mon, 10 Sep 2007 10:11:33 -0700
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <p06230934c3081eaab73d@[192.168.1.205]>
References: <p06230934c3081eaab73d@[192.168.1.205]>
Message-ID: <E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>

Rich -

Postal Code now has an included co-type of /location/location (a  
recent change!)

This gives you the contains/contained_by properties so you can make  
the association between Postal Codes and the cities/neighborhoods  
they are associated with.  The /location/location co-type also gives  
you the geolocation property so you can include the centroid for the  
Postal Code area (or ZCTA bounding region since US Postal Codes are  
carrier line routes.)

We have an open gardening task to co-type existing Postal Code  
instances and provide as many of the /location/location properties as  
we can get our hands on.  That said, please fill in the details you  
know, as I don't think our data will get us to the neighborhood  
level, which is where it gets really useful.

J

On Sep 8, 2007, at 2:47 AM, Rich Morin wrote:

> Although Freebase recognizes 94066 as a postal code, I
> don't see any relations that would link this to "San
> Bruno", Latitude: 37.63 / Longitude: -122.43, etc.  Am
> I simply looking in the wrong places?
>
> -r
> -- 
> http://www.cfcl.com/rdm            Rich Morin
> http://www.cfcl.com/rdm/resume     rdm at cfcl.com
> http://www.cfcl.com/rdm/weblog     +1 650-873-7841
>
> Technical editing and writing, programming, and web development
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From ptomblin at xcski.com  Mon Sep 10 17:21:59 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Mon, 10 Sep 2007 13:21:59 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
Message-ID: <20070910172159.GA27250@allhats2.xcski.com>

Quoting Jamie Taylor (jamie at metaweb.com):
> Rich -
> 
> Postal Code now has an included co-type of /location/location (a  
> recent change!)

One big honking problem leaps out at me with that.  If I have, say, a
house that has a location and a zip code, I want the location of the
house, not the centroid of the zip code area.  How would that work?

-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
There's going to be no serious problem after this.  --Ken Thompson

From karell.stemarie at gmail.com  Mon Sep 10 17:38:14 2007
From: karell.stemarie at gmail.com (Karell Ste-Marie)
Date: Mon, 10 Sep 2007 13:38:14 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
Message-ID: <46e5810b.395c220a.1816.650b@mx.google.com>

What about adding the long/lat to the zip code?

I have all of this information for most of the US/CAN postal codes, nothing
would make me happier than to batch upload this information to the system.

There are many applications to this information.


-----------------------------------------------
Karell Ste-Marie
MCP, MCP+SB, MCP+I, MCAD, MCSE
http://karell.blogspot.com/
http://bodyspace.bodybuilding.com/n5d/
http://blog.bodybuilding.com/n5d

-----Original Message-----
From: developers-bounces at freebase.com
[mailto:developers-bounces at freebase.com] On Behalf Of Jamie Taylor
Sent: Monday, September 10, 2007 1:12 PM
To: Rich Morin
Cc: developers at freebase.com
Subject: Re: [Developers] mapping Zip codes to other location info?

Rich -

Postal Code now has an included co-type of /location/location (a  
recent change!)

This gives you the contains/contained_by properties so you can make  
the association between Postal Codes and the cities/neighborhoods  
they are associated with.  The /location/location co-type also gives  
you the geolocation property so you can include the centroid for the  
Postal Code area (or ZCTA bounding region since US Postal Codes are  
carrier line routes.)

We have an open gardening task to co-type existing Postal Code  
instances and provide as many of the /location/location properties as  
we can get our hands on.  That said, please fill in the details you  
know, as I don't think our data will get us to the neighborhood  
level, which is where it gets really useful.

J

On Sep 8, 2007, at 2:47 AM, Rich Morin wrote:

> Although Freebase recognizes 94066 as a postal code, I
> don't see any relations that would link this to "San
> Bruno", Latitude: 37.63 / Longitude: -122.43, etc.  Am
> I simply looking in the wrong places?
>
> -r
> -- 
> http://www.cfcl.com/rdm            Rich Morin
> http://www.cfcl.com/rdm/resume     rdm at cfcl.com
> http://www.cfcl.com/rdm/weblog     +1 650-873-7841
>
> Technical editing and writing, programming, and web development
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers

_______________________________________________
Developers mailing list
Developers at freebase.com
http://lists.freebase.com/mailman/listinfo/developers


From ptomblin at xcski.com  Mon Sep 10 17:46:10 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Mon, 10 Sep 2007 13:46:10 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <46e5810b.395c220a.1816.650b@mx.google.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
	<46e5810b.395c220a.1816.650b@mx.google.com>
Message-ID: <20070910174610.GA30644@allhats2.xcski.com>

Quoting Karell Ste-Marie (karell.stemarie at gmail.com):
> What about adding the long/lat to the zip code?

It's unfortunate that there isn't any way to model a polygon
in Freebase, because a geographic boundary for the zip code region makes a
lot more sense than a single lat/long[1].  Especially since all the
physical objects in the zip code have their own lat/long which probably
doesn't correspond to whatever lat/long you assign to the zip code?


[1] And can I just say for the record that whoever decided that
Geolocation should have the longitude first and that the latitude and
longitude should be separated by a dash should be taken out and shot?
First of all, I know of no single application in the geographic world that
puts the longitude first.  Second of all, since there is a whole
hemisphere where the second number is negative, so separating the numbers
by a dash is visually confusing.

-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
You mean [Exchange] was deliberately written?  I thought someone had
transcribed the writing on a football-stadium restroom wall, found
that it compiled, so shipped it.
              -- Tanuki

From sarah at metaweb.com  Mon Sep 10 17:47:43 2007
From: sarah at metaweb.com (Sarah Milstein)
Date: Mon, 10 Sep 2007 10:47:43 -0700
Subject: [Developers] Ontological forum?
In-Reply-To: <7D05FAD7-E54F-40D2-9E31-92E95B2A3168@metaweb.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>	<46E3A488.2000301@gmail.com>	<b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>	<46E3ACA8.6030801@gmail.com>	<b08e1490709090251m3a82562bm8e355b2851ed2e8d@mail.gmail.com>	<b08e1490709090252j534c4e0co822533af2ab9dca8@mail.gmail.com>
	<7D05FAD7-E54F-40D2-9E31-92E95B2A3168@metaweb.com>
Message-ID: <46E5833F.1060204@metaweb.com>

An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070910/685ecd0f/attachment.htm 

From rdm at cfcl.com  Mon Sep 10 18:03:19 2007
From: rdm at cfcl.com (Rich Morin)
Date: Mon, 10 Sep 2007 11:03:19 -0700
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <p06230934c3081eaab73d@[192.168.1.205]>
References: <p06230934c3081eaab73d@[192.168.1.205]>
Message-ID: <p06230950c30b3657462d@[192.168.1.205]>

Some of the need for polygonal information might be reduced if
Zip+4 data were available, as it is much higher resolution.  Of
course, this assumes that the query itself uses Zip+4...

Also, I'd like to suggest that each location be linked to its
"neighbors".  For example, 94006 (San Bruno, CA) is just over
the ridge from 94044 (Pacifica, CA).  I suspect that this can
be largely automated, based on Lat/Lon information, but that
occasional oddities may have to be handled by hand.

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From jamie at metaweb.com  Mon Sep 10 18:06:31 2007
From: jamie at metaweb.com (Jamie Taylor)
Date: Mon, 10 Sep 2007 11:06:31 -0700
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <20070910172159.GA27250@allhats2.xcski.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
	<20070910172159.GA27250@allhats2.xcski.com>
Message-ID: <BD199E87-6D38-4BCD-893A-B0AF77F8E8CE@metaweb.com>

Paul -

The house would be an /architecture/structure which has a mailing  
address property (which is co-typed as a location, so it would have  
its own geocode.)  The Postal Code referenced in the address, like a  
city, would also have a centroid lat/lon.

J

On Sep 10, 2007, at 10:21 AM, Paul Tomblin wrote:

> Quoting Jamie Taylor (jamie at metaweb.com):
>> Rich -
>>
>> Postal Code now has an included co-type of /location/location (a
>> recent change!)
>
> One big honking problem leaps out at me with that.  If I have, say, a
> house that has a location and a zip code, I want the location of the
> house, not the centroid of the zip code area.  How would that work?
>
> -- 
> Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
> There's going to be no serious problem after this.  --Ken Thompson


From karell.stemarie at gmail.com  Mon Sep 10 18:14:11 2007
From: karell.stemarie at gmail.com (Karell Ste-Marie)
Date: Mon, 10 Sep 2007 14:14:11 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <20070910174610.GA30644@allhats2.xcski.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
	<46e5810b.395c220a.1816.650b@mx.google.com>
	<20070910174610.GA30644@allhats2.xcski.com>
Message-ID: <46e58976.3e61220a.44c5.58f5@mx.google.com>

Hi Paul,

Well, it's not that bad, a zip code can have a "reference" long/lat for the
purpose of seeing the distance of one zip code versus another. Of course, it
is much more precise to measure the distance between my driveway and the
front of your house but for most applications I still think that the
long/lat addition to the zip code may be precise enough. In most web
applications that I've seen I will provide my zip code and it will use
long/lat from my zip code and compare that to the destination zip code and
calculate the distance between both. Of course, over a long distance this is
not very accurate since the earth is curved on the surface but some also
take that into account. Elevation is considered to be the holy grail in such
applications as well.

I do agree that the US's decision to go backwards on the international
standard is quite backwards, about as much as driving on the left side of
the road. But then we can always debate the imperial system versus the
metric system.

As for the separator being a dash, I can only completely agree.


-----------------------------------------------
Karell Ste-Marie
MCP, MCP+SB, MCP+I, MCAD, MCSE
http://karell.blogspot.com/
http://bodyspace.bodybuilding.com/n5d/
http://blog.bodybuilding.com/n5d


-----Original Message-----
From: Paul Tomblin [mailto:ptomblin at xcski.com] 
Sent: Monday, September 10, 2007 1:46 PM
To: Karell Ste-Marie
Cc: developers at freebase.com
Subject: Re: [Developers] mapping Zip codes to other location info?

Quoting Karell Ste-Marie (karell.stemarie at gmail.com):
> What about adding the long/lat to the zip code?

It's unfortunate that there isn't any way to model a polygon
in Freebase, because a geographic boundary for the zip code region makes a
lot more sense than a single lat/long[1].  Especially since all the
physical objects in the zip code have their own lat/long which probably
doesn't correspond to whatever lat/long you assign to the zip code?


[1] And can I just say for the record that whoever decided that
Geolocation should have the longitude first and that the latitude and
longitude should be separated by a dash should be taken out and shot?
First of all, I know of no single application in the geographic world that
puts the longitude first.  Second of all, since there is a whole
hemisphere where the second number is negative, so separating the numbers
by a dash is visually confusing.

-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
You mean [Exchange] was deliberately written?  I thought someone had
transcribed the writing on a football-stadium restroom wall, found
that it compiled, so shipped it.
              -- Tanuki


From ptomblin at xcski.com  Mon Sep 10 18:26:00 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Mon, 10 Sep 2007 14:26:00 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <46e58976.3e61220a.44c5.58f5@mx.google.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
	<46e5810b.395c220a.1816.650b@mx.google.com>
	<20070910174610.GA30644@allhats2.xcski.com>
	<46e58976.3e61220a.44c5.58f5@mx.google.com>
Message-ID: <20070910182600.GA4207@allhats2.xcski.com>

Quoting Karell Ste-Marie (karell.stemarie at gmail.com):
> Well, it's not that bad, a zip code can have a "reference" long/lat for the
> purpose of seeing the distance of one zip code versus another. Of course, it
> is much more precise to measure the distance between my driveway and the
> front of your house but for most applications I still think that the
> long/lat addition to the zip code may be precise enough. In most web
> applications that I've seen I will provide my zip code and it will use

"Most" applications.  An airport has a Location, but it also has an
address.  That address has a Zip Code.  If you include the type "Zip Code"
and the type "Location", which lat/long are you going to get, the one for
the airport, or the one for the zip code centroid?  That might not make
any difference for calculating UPS shipping rates, but it sure as hell
makes a difference if you're flying there.

-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
There are two ways to write error-free programs.  Only the third one works.

From ptomblin at xcski.com  Mon Sep 10 18:51:54 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Mon, 10 Sep 2007 14:51:54 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <BD199E87-6D38-4BCD-893A-B0AF77F8E8CE@metaweb.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
	<20070910172159.GA27250@allhats2.xcski.com>
	<BD199E87-6D38-4BCD-893A-B0AF77F8E8CE@metaweb.com>
Message-ID: <20070910185154.GA9390@allhats2.xcski.com>

Quoting Jamie Taylor (jamie at metaweb.com):
> Paul -
> 
> The house would be an /architecture/structure which has a mailing  
> address property (which is co-typed as a location, so it would have  
> its own geocode.)  The Postal Code referenced in the address, like a  
> city, would also have a centroid lat/lon.

Never mind house, let's look at Airport, something which needs to be
co-typed with location a lot more than a house does.  And it also has an
address.

Ok, so the address property would be of type "address" (say) and that
would have a property that was of type "postal code".  In that case the
location in the postal code wouldn't interfere with the location in the
Airport?

I guess I was getting confused between "IS A" and "HAS A" relationships.
Airport "IS A" location, but it "HAS A" zip code.  So it doesn't matter
that zip code "IS A" location as well.

Sorry about the confusion, and I'll just shut up now.

-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
I forsee one of those "open your wallet and repeat after me,
_help yourself_" moments in your local friendly workshop.
              -- Tanuki

From karell.stemarie at gmail.com  Mon Sep 10 18:58:59 2007
From: karell.stemarie at gmail.com (Karell Ste-Marie)
Date: Mon, 10 Sep 2007 14:58:59 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <20070910182600.GA4207@allhats2.xcski.com>
References: <p06230934c3081eaab73d@[192.168.1.205]>
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com>
	<46e5810b.395c220a.1816.650b@mx.google.com>
	<20070910174610.GA30644@allhats2.xcski.com>
	<46e58976.3e61220a.44c5.58f5@mx.google.com>
	<20070910182600.GA4207@allhats2.xcski.com>
Message-ID: <46e593f1.30e1220a.0ce1.67d4@mx.google.com>

Well, that depends on what we are trying to accomplish here. If we want we
can put a structure that will allow every house to have it's own geolocation
- fine. But what do we use? The front door? The center of the house? The
center of the property? What if the property is not a square or rectangle?
This opens up a huge can of worms.

While I understand that you are not suggesting that we implement geolocation
for every rock on the planet I question the need to have such granular data.
What you are suggesting is that every house (or property) should have it's
own long/lat and I personally think that this is overkill. While I can
understand the argument of the required precision of having a proper
long/lat for an airport again I must ask: the geolocation of what? The
communication tower? The parking lot? The landing runway? Which one? An
airport is a pretty bit place and if you compare it to a zip code I think
it's safe to assume (at least in Canada where the zip codes are usually
quite precise) that an airport can span multiple zip codes in the sense of
square feet. So if that is the case, then the zip code is actually more
precise for this exercise than the airport's geolocation itself.

http://srmwww.gov.bc.ca/gis/storageprecision.html
On a 9 significant-digit geolocation long/lat you have meter (4 four feet)
precision, using 11 significant-digit precision you will have centimeter
precision (1/3 inch). Then with single-precision you will have 100-meter
precision (most GPSes), with double precision you will have 1.9cm precision.
So taking this into account, you would have to include an array or a
3-dimension address (or a from/to/elevation to create a rectangle).

Personally, I still think that dead-center is the best thing to do and if
you need more precision then you can put it in on a per-case basis. But
expecting someone to figure out their long/lat for their own home will lead
to many mistakes. The UPS man may think that it's 50 miles away to send a
package to my next-door neighbor.

Can we meet in the middle and say that you can have geolocation for a zip
code and for an address but that the address may default to using the zip
code's?

-----------------------------------------------
Karell Ste-Marie
MCP, MCP+SB, MCP+I, MCAD, MCSE
http://karell.blogspot.com/
http://bodyspace.bodybuilding.com/n5d/
http://blog.bodybuilding.com/n5d

-----Original Message-----
From: Paul Tomblin [mailto:ptomblin at xcski.com] 
Sent: Monday, September 10, 2007 2:26 PM
To: Karell Ste-Marie
Cc: developers at freebase.com
Subject: Re: [Developers] mapping Zip codes to other location info?

"Most" applications.  An airport has a Location, but it also has an
address.  That address has a Zip Code.  If you include the type "Zip Code"
and the type "Location", which lat/long are you going to get, the one for
the airport, or the one for the zip code centroid?  That might not make
any difference for calculating UPS shipping rates, but it sure as hell
makes a difference if you're flying there.


From ptomblin at xcski.com  Mon Sep 10 20:13:24 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Mon, 10 Sep 2007 16:13:24 -0400
Subject: [Developers] mapping Zip codes to other location info?
In-Reply-To: <46e593f1.30e1220a.0ce1.67d4@mx.google.com>
References: <p06230934c3081eaab73d@[192.168.1.205]> 
	<E173AB56-D196-4B66-A67C-3A63AA47D178@metaweb.com> 
	<46e5810b.395c220a.1816.650b@mx.google.com> 
	<20070910174610.GA30644@allhats2.xcski.com> 
	<46e58976.3e61220a.44c5.58f5@mx.google.com> 
	<20070910182600.GA4207@allhats2.xcski.com> 
	<46e593f1.30e1220a.0ce1.67d4@mx.google.com>
Message-ID: <8074-SnapperMsgFB251063C30B58AB@[10.78.92.179]>


On Mon, 10 Sep 2007 14:58:59 -0400 "Karell Ste-Marie" 
<karell.stemarie at gmail.com> wrote:
>What you are suggesting is that every house (or property) should have it's
>own long/lat and I personally think that this is overkill. While I can
>understand the argument of the required precision of having a proper
>long/lat for an airport again I must ask: the geolocation of what? The
>communication tower? The parking lot? The landing runway? Which one?

Every airport in the world (at least the ones listed in its country's AIP) have a designated 
location point.  There are ICAO rules on how to place it, but there will be a survey monument 
at that location.


___
Sent with SnapperMail
www.snappermail.com

From tim at metaweb.com  Mon Sep 10 20:30:18 2007
From: tim at metaweb.com (Tim Kientzle)
Date: Mon, 10 Sep 2007 13:30:18 -0700
Subject: [Developers] Importing data from World of Spectrum
In-Reply-To: <20070908150450.GA8772@sphinx.mythic-beasts.com>
References: <20070908150450.GA8772@sphinx.mythic-beasts.com>
Message-ID: <46E5A95A.7060903@metaweb.com>

Philip Kendall wrote:
> 
> 2) If so, the thing which is worrying me most is accidently creating
>    a large number duplicate topics for things like software houses,
>    where the Freebase entry happens to have a slightly different name
>    from the name used by World of Spectrum. Two questions here, really:

We call this the "reconciliation problem."  It is probably
our single biggest headache importing new data.  (There
are entire academic conferences on this subject; it's not a
simple problem to solve in any generic fashion.)

I suggest you start by doing some experiments against our
sandbox system to see how many things you can match.
Start with manual experiments and searches, then try
automating them.  From there, a lot depends on the nature
of the data you have and how well it happens to line up.

What you'll often find is that some percentage of your
data matches up very easily.  By using some creativity
("companies with the same first four letters in the same
city/state", perhaps?) you might find a number of other
matches and/or determine that your objects don't already
exist.

There will always be a few errors and there will always
be room to make manual corrections afterwards, so try not
to get too hung up getting everything exactly correct.

It's mostly just a tricky judgment call when your work
so far is "good enough." Finding someone else who can
collaborate would certainly help.  It might also be useful
to import subsets of the data at a time (first, the
data that matches exactly, then additional subsets as
you try different matching strategies).  That would also
make it easier for others to look at the result and give
you feedback.

Good luck, and let us know how it goes.

Tim Kientzle
Metaweb Technologies

From ptomblin at xcski.com  Mon Sep 10 21:59:43 2007
From: ptomblin at xcski.com (Paul Tomblin)
Date: Mon, 10 Sep 2007 17:59:43 -0400
Subject: [Developers] mapping Zip codes to other location info?
Message-ID: <20070910215943.GB8475@allhats2.xcski.com>

Quoting Paul Tomblin (ptomblin at xcski.com):
> Every airport in the world (at least the ones listed in its country's
> AIP) have a designated location point.  There are ICAO rules on how to
> place it, but there will be a survey monument at that location.

Sorry, I meant to expand those abbreviations, but I was composing the
message while in the air and it was nearly my turn to fly.

AIP == Aeronautical Information Publication.  These document all the
airports and navaids in a country's airspace.  There are strict
requirements for the format of these documents.  It's amusing to look at
the AIP for a country like Albania which only has one airport, but still
includes all the front matter and back matter required by the standards.

ICAO == International Civil Aviation Authority.  They are responsible for
making the standards that allow a pilot from Albania to make sense of the
airspace and airports in Zambia.

-- 
Paul Tomblin <ptomblin at xcski.com> http://blog.xcski.com/
"Power corrupts.  Powerpoint corrupts absolutely." - Vint Cerf

From skud at infotrope.net  Tue Sep 11 00:10:58 2007
From: skud at infotrope.net (Kirrily Robert)
Date: Tue, 11 Sep 2007 10:10:58 +1000
Subject: [Developers] Ontological forum?
In-Reply-To: <46E5833F.1060204@metaweb.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>
	<46E3A488.2000301@gmail.com>
	<b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>
	<46E3ACA8.6030801@gmail.com>
	<b08e1490709090251m3a82562bm8e355b2851ed2e8d@mail.gmail.com>
	<b08e1490709090252j534c4e0co822533af2ab9dca8@mail.gmail.com>
	<7D05FAD7-E54F-40D2-9E31-92E95B2A3168@metaweb.com>
	<46E5833F.1060204@metaweb.com>
Message-ID: <b08e1490709101710q27a6d841l1d5b13ae660f38b5@mail.gmail.com>

On 9/11/07, Sarah Milstein <sarah at metaweb.com> wrote:
>
>  So as Robert noted, one possibility--and it's available right now--is to
> start threads on the existing data modeling guide. In addition to immediacy,
> this spot has the advantage of collecting info in an area where other
> Freebasers will look in the future. The disadvantage is that we don't yet
> have RSS for individual discussions, so you'd have to check back for
> updates.

I'd say that RSS feeds and/or email notification for discussions were
pretty much at the top of my wishlist for FB right now.  It's one of
the things that makes it hardest to use, having to manually go and
check for messages somewhere where I'm not already looking.  And I'm
betting that heaps of people never know about replies they get, or
discussions that might be of interest to them, and hence don't stay
involved as much as they might.

>  If there's a preference for listserving these conversations, we could keep
> them on this list for a little while (a week? a month?) until there's
> critical mass to spawn a new list. Or, if there's the desire for such a list
> right now, we can set one up. We do, of course, keep archives of the list/s,
> but at the moment, they're separate from Freebase.com.

Sounds good to me.  I think that quick little off-the-cuff comments
are well suited to FB itself, but in-depth discussions are better in
email.

K.

-- 
Kirrily Robert
skud at infotrope.net
http://infotrope.net

From niallo at metaweb.com  Tue Sep 11 00:26:58 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 10 Sep 2007 17:26:58 -0700
Subject: [Developers] Weekly sandbox data refresh
Message-ID: <20070911002658.GL7152@digdug.inet>

Hi,

Weekly data refresh of http://sandbox.freebase.com happening over the
next few minutes.

-- 
Niall O'Higgins
Software Engineer
Metaweb Technologies, Inc.

From narphorium at gmail.com  Tue Sep 11 00:34:51 2007
From: narphorium at gmail.com (Shawn Simister)
Date: Mon, 10 Sep 2007 20:34:51 -0400
Subject: [Developers] Ontological forum?
In-Reply-To: <b08e1490709101710q27a6d841l1d5b13ae660f38b5@mail.gmail.com>
References: <b08e1490709090025s567803e7h9eda75e5dc37fd1d@mail.gmail.com>	<46E3A488.2000301@gmail.com>	<b08e1490709090108o1b8f4258gfa3939d5e224bc6f@mail.gmail.com>	<46E3ACA8.6030801@gmail.com>	<b08e1490709090251m3a82562bm8e355b2851ed2e8d@mail.gmail.com>	<b08e1490709090252j534c4e0co822533af2ab9dca8@mail.gmail.com>	<7D05FAD7-E54F-40D2-9E31-92E95B2A3168@metaweb.com>	<46E5833F.1060204@metaweb.com>
	<b08e1490709101710q27a6d841l1d5b13ae660f38b5@mail.gmail.com>
Message-ID: <46E5E2AB.1050007@gmail.com>

Kirrily Robert wrote:
> On 9/11/07, Sarah Milstein <sarah at metaweb.com> wrote:
>   
>>  So as Robert noted, one possibility--and it's available right now--is to
>> start threads on the existing data modeling guide. In addition to immediacy,
>> this spot has the advantage of collecting info in an area where other
>> Freebasers will look in the future. The disadvantage is that we don't yet
>> have RSS for individual discussions, so you'd have to check back for
>> updates.
>>     
> I'd say that RSS feeds and/or email notification for discussions were
> pretty much at the top of my wishlist for FB right now.  It's one of
> the things that makes it hardest to use, having to manually go and
> check for messages somewhere where I'm not already looking.  And I'm
> betting that heaps of people never know about replies they get, or
> discussions that might be of interest to them, and hence don't stay
> involved as much as they might
>>  If there's a preference for listserving these conversations, we could keep
>> them on this list for a little while (a week? a month?) until there's
>> critical mass to spawn a new list. Or, if there's the desire for such a list
>> right now, we can set one up. We do, of course, keep archives of the list/s,
>> but at the moment, they're separate from Freebase.com.
>>     
> Sounds good to me.  I think that quick little off-the-cuff comments
> are well suited to FB itself, but in-depth discussions are better in
> email.
>
> K
I'll second both those points. Email alerts would be a big help.

Shawn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070910/7f523236/attachment.htm 

From niallo at metaweb.com  Tue Sep 11 00:38:17 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 10 Sep 2007 17:38:17 -0700
Subject: [Developers] Weekly sandbox data refresh
In-Reply-To: <20070911002658.GL7152@digdug.inet>
References: <20070911002658.GL7152@digdug.inet>
Message-ID: <20070911003817.GN7152@digdug.inet>

Refresh complete.

On Mon, Sep 10, 2007 at 05:26:58PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Weekly data refresh of http://sandbox.freebase.com happening over the
> next few minutes.
> 
> -- 
> Niall O'Higgins
> Software Engineer
> Metaweb Technologies, Inc.
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers

-- 
Niall O'Higgins
Software Engineer
Metaweb Technologies, Inc.

From ksrinivs at gmail.com  Tue Sep 11 03:25:35 2007
From: ksrinivs at gmail.com (Kavitha Srinivas)
Date: Mon, 10 Sep 2007 23:25:35 -0400
Subject: [Developers] MQL newbie question
In-Reply-To: <6376A86D-A0A2-4B4A-82DF-FEDB2422729E@metaweb.com>
References: <p06230933c3081b25e444@[192.168.1.205]>
	<7712A8DA-F32B-4159-8185-650AC5889C0D@metaweb.com>
	<C2EF639E-E67B-439A-8702-2FF7FAB1DDF7@gmail.com>
	<6376A86D-A0A2-4B4A-82DF-FEDB2422729E@metaweb.com>
Message-ID: <A1790050-9B47-48E5-8105-AAF632DFFFF0@gmail.com>

Hi,
    I'm experimenting with MQL, and decided to try the query:
{
   "qname": {
     "query": [{
       "type":"/film/film",
       "name~=":"The Usual Suspects",
        "*":null
     }]
   }
}

in the query editor.  I do get a response, but its not one I expected.
Here's what I get -- notice the lack of information about awards, or  
about performances, both of which have entries when I pull up this  
object in the web page.  What am I doing wrong?  Thanks
Kavitha
------------------------------------------------------------------------ 
----------------------------------
({
   "status": "200 OK",
   "code": "/api/status/ok",
   "qname": {
     "code": "/api/status/ok",
     "result": [
       {
         "personal_appearances": [],
         "rating": [],
         "key": [
           "167857",
           "The_Usual_Suspects",
           "Usual_Suspects",
           "Verbal_Kint",
           "Roger_Kint",
           "the_usual_suspects",
           "The_Usual_Suspects_$0028film$0029"
         ],
         "creator": "/user/metaweb",
         "locations": [],
         "guid": "#9202a8c04000641f80000000001337ab",
         "id": "/topic/en/the_usual_suspects",
         "film_format": [],
         "produced_by": [
           "Michael McDonnell",
           "Bryan Singer"
         ],
         "name": "The Usual Suspects",
         "cinematography": [],
         "subjects": [],
         "music": [
           "John Ottman"
         ],
         "poster": [
           null
         ],
         "type": "/film/film",
         "edited_by": [],
         "initial_release_date": [
           "1995-08-16"
         ],
         "permission": "/boot/all_permission",
         "timestamp": "2006-10-22T12:35:48.0011Z",
         "sequel": [],
         "imdb_id": [
           "http://www.imdb.com/title/tt0114814"
         ],
         "soundtrack": [],
         "estimated_budget": [
           null
         ],
         "awards": [
           null
         ],
         "written_by": [
           "Christopher McQuarrie"
         ],
         "starring": [
           null,
           null,
           null,
           null,
           null,
           null,
           null,
           null,
           null,
           null
         ],
         "genre": [
           "Crime",
           "Mystery",
           "Thriller"
         ],
         "story_by": [],
         "featured_film_locations": [],
         "film_series": [],
         "language": [
           "English"
         ],
         "country": [],
         "dubbing_performances": [],
         "directed_by": [
           "Bryan Singer"
         ],
         "film_collections": [],
         "nominations": [],
         "runtime": [
           null
         ],
         "prequel": []
       },
       {
         "personal_appearances": [],
         "rating": [],
         "key": [],
         "creator": "/user/cg",
         "locations": [],
         "guid": "#9202a8c04000641f80000000042b08bd",
         "id": "#9202a8c04000641f80000000042b08bd",
         "film_format": [],
         "produced_by": [],
         "name": "the usual suspects",
         "cinematography": [],
         "subjects": [],
         "music": [],
         "poster": [],
         "type": "/film/film",
         "edited_by": [],
         "initial_release_date": [],
         "permission": "/boot/all_permission",
         "timestamp": "2007-03-14T22:48:08.0000Z",
         "sequel": [],
         "imdb_id": [],
         "soundtrack": [],
         "estimated_budget": [],
         "awards": [],
         "written_by": [],
         "starring": [],
         "genre": [],
         "story_by": [],
         "featured_film_locations": [],
         "film_series": [],
         "language": [],
         "country": [],
         "dubbing_performances": [],
         "directed_by": [],
         "film_collections": [],
         "nominations": [],
         "runtime": [],
         "prequel": []
       }
     ]
   }
})

From jg at metaweb.com  Tue Sep 11 05:41:59 2007
From: jg at metaweb.com (John Giannandrea)
Date: Mon, 10 Sep 2007 22:41:59 -0700
Subject: [Developers] MQL newbie question
In-Reply-To: <A1790050-9B47-48E5-8105-AAF632DFFFF0@gmail.com>
References: <p06230933c3081b25e444@[192.168.1.205]>
	<7712A8DA-F32B-4159-8185-650AC5889C0D@metaweb.com>
	<C2EF639E-E67B-439A-8702-2FF7FAB1DDF7@gmail.com>
	<6376A86D-A0A2-4B4A-82DF-FEDB2422729E@metaweb.com>
	<A1790050-9B47-48E5-8105-AAF632DFFFF0@gmail.com>
Message-ID: <F2163661-C160-4B52-83A2-B4F701255550@metaweb.com>


Kavitha Srinivas wrote:
> in the query editor.  I do get a response, but its not one I expected.
> Here's what I get -- notice the lack of information about awards, or
> about performances, both of which have entries when I pull up this
> object in the web page.  What am I doing wrong?  Thanks

Hi
There were two problems here, we had a second film which was
not the right one, which I have removed.

To se the performances and awards, you have to look one ply
out into the 'compound value types' that model these things.

For example:

{
   "query":[{
     "*":[{}],
     "name":"The Usual Suspects",
     "type":"/film/film"
   }]
}

The schema for a film performance is here:
	http://www.freebase.com/view/schema/film/performance

So to get a list of those, something like this would work:

{
   "query":[{
     "name":"The Usual Suspects",
     "starring":[{
       "actor":null,
       "part":null
     }],
     "type":"/film/film"
   }]
}

-jg


From niallo at metaweb.com  Tue Sep 11 23:06:38 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 11 Sep 2007 16:06:38 -0700
Subject: [Developers] Sandbox software update
Message-ID: <20070911230638.GS7152@digdug.inet>

Hi,

Weekly sandbox software update happening over the next few minutes.
You may notice short blips in service during our maintainence window.

-- 
Niall O'Higgins
Software Engineer
Metaweb Technologies, Inc.

From niallo at metaweb.com  Wed Sep 12 00:00:54 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 11 Sep 2007 17:00:54