From daepark at metaweb.com  Mon Jul  2 19:43:14 2007
From: daepark at metaweb.com (Dae Park)
Date: Mon, 2 Jul 2007 12:43:14 -0700
Subject: [Developers] mjt html generation
Message-ID: <B36DCA18-3679-447E-942E-FAAAA4B6B5C2@metaweb.com>

Is the mjt template generation synchronous? For example if I call a  
mjt.def, will I be able to access the element created in the mjt.def  
in a mjt.script:

<div mjt.def="foo()">
   <p id="bar">Hello world</p>
</div>
<div>
   ${foo()}
   <pre mjt.script="">
     alert(document.getElementById("bar"));
   </pre>
</div>

From lindner at inuus.com  Mon Jul  2 23:20:19 2007
From: lindner at inuus.com (Paul Lindner)
Date: Mon, 2 Jul 2007 16:20:19 -0700
Subject: [Developers] Schools, Colleges, Data loading and more...
Message-ID: <20070702232019.GP19982@inuus.com>

Hi there,

I was recently invited to the Freebase alpha.  I like what I see!  I
work for Hi5 and we can see a ton of potential here.

Couple of questions though.

* What about internationalization?  Are there facilities for storing
  different names based on language?  Or would that be a
  specialization of an existing type?

    For example the city New York City is Nueva York in Spanish

  This looks like a big problem and I hope it gets solved in a
  comprehensive way.  It would be nice to see the ability to mark
  certain text as the canonical form and allow other translations to
  be made.  Then if the canonical form is changed the other languages
  can be flagged for re-verification.

* Is there any planned integration with geonames.org?

* Hi5 has a large database of schools we could contribute.  Most of
  these are imported from official sources, we could just submit the
  data but I'm more interested in contributing tools that can be used to
  refresh data from offical sources.  What's the best way to do that?

* Would it be possible to integrate editing / submitting functionality
  on external sites?  Something better than just a link to the
  freebase edit page might be interesting.

* How about OpenID authentication for contributors?  This would
  obviously be something for post-alpha.


In any case, glad to see this effort and am excited about the
possibilities that it offers.

-- 
Paul Lindner        ||||| | | | |  |  |  |   |   |
lindner at inuus.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.freebase.com/pipermail/developers/attachments/20070702/93380c57/attachment.pgp 

From daepark at metaweb.com  Mon Jul  2 23:54:48 2007
From: daepark at metaweb.com (Dae Park)
Date: Mon, 2 Jul 2007 16:54:48 -0700
Subject: [Developers] mjt.def
Message-ID: <19E07FD9-D6CB-42E2-A70F-59FC6D5127C5@metaweb.com>

Is there anyway to share "mjt.def" declarations across multiple  
templates?

From crism at metaweb.com  Tue Jul  3 00:02:00 2007
From: crism at metaweb.com (Christopher R. Maden)
Date: Mon, 02 Jul 2007 17:02:00 -0700
Subject: [Developers] Schools, Colleges, Data loading and more...
In-Reply-To: <20070702232019.GP19982@inuus.com>
References: <20070702232019.GP19982@inuus.com>
Message-ID: <468991F8.1060107@metaweb.com>

Glad you like it, Paul!  I can address one of these questions for you.

Paul Lindner wrote:
> * What about internationalization?  Are there facilities for storing
>   different names based on language?  Or would that be a
>   specialization of an existing type?
> 
>     For example the city New York City is Nueva York in Spanish

The underlying data model is already there.  If you go look with the 
query editor, you will see that there are already polylingual names for 
things:

{
   "query": {
     "id": "/wikipedia/en/New_York_City",
     "name": [{}]
   }
}

{
   "result": {
     "id": "/wikipedia/en/New_York_City",
     "name": [
       {
         "lang": "/lang/fr",
         "type": "/type/text",
         "value": "New York"
       },
       {
         "lang": "/lang/de",
         "type": "/type/text",
         "value": "New York City"
       },
       {
         "lang": "/lang/it",
         "type": "/type/text",
         "value": "New York"
       },
       {
         "lang": "/lang/ja",
         "type": "/type/text",
         "value": "??????"
       },
       {
         "lang": "/lang/es",
         "type": "/type/text",
         "value": "Nueva York"
       },
       {
         "lang": "/lang/en",
         "type": "/type/text",
         "value": "New York"
       }
     ]
   }
}

This isn?t yet exposed in the UI, but the data is there and can be 
manipulated in queries.

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

From niallo at metaweb.com  Tue Jul  3 00:07:37 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 2 Jul 2007 17:07:37 -0700
Subject: [Developers] Sandbox update
Message-ID: <20070703000737.GO5935@digdug.inet>

Hi,

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

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

From robert at metaweb.com  Tue Jul  3 00:23:23 2007
From: robert at metaweb.com (Robert Cook)
Date: Mon, 2 Jul 2007 17:23:23 -0700
Subject: [Developers] Schools, Colleges, Data loading and more...
In-Reply-To: <20070702232019.GP19982@inuus.com>
References: <20070702232019.GP19982@inuus.com>
Message-ID: <DFB5F2EF-E9A5-45AE-8C02-E73BDF9B3912@metaweb.com>

Paul -- Your questions are very good.  You'll notice that we're still  
getting the basics going, but everything you've mentioned we've  
seriously considered.  Feedback like yours helps us prioritize which  
features we do first.

On Jul 2, 2007, at 4:20 PM, Paul Lindner wrote:

> Hi there,
>
> I was recently invited to the Freebase alpha.  I like what I see!  I
> work for Hi5 and we can see a ton of potential here.
>
> Couple of questions though.
>
> * What about internationalization?  Are there facilities for storing
>   different names based on language?  Or would that be a
>   specialization of an existing type?
>
>     For example the city New York City is Nueva York in Spanish
>
>   This looks like a big problem and I hope it gets solved in a
>   comprehensive way.  It would be nice to see the ability to mark
>   certain text as the canonical form and allow other translations to
>   be made.

Internationalization is part of our core infrastructure, although  
we've done little so far to surface that in the UI.  Hundreds of  
thousands of topics already have labels in multiple languages.  We'll  
be adding more UI support for internationalization soon.

>   Then if the canonical form is changed the other languages
>   can be flagged for re-verification.

This is a clever idea -- we view topic name changes as an important  
event (mostly for vandalism control), but this could help guide  
translators.

>
> * Is there any planned integration with geonames.org?

We've considered it, and I believe our Minister of Information (Jamie  
Taylor) has spoken with them.  Geo is also important to us, so we  
will be trying to work with existing systems.

> * Hi5 has a large database of schools we could contribute.  Most of
>   these are imported from official sources, we could just submit the
>   data but I'm more interested in contributing tools that can be  
> used to
>   refresh data from offical sources.  What's the best way to do that?

We have an internal tool that allows you to upload spreadsheets and  
(most importantly) reconcile the data with existing topics.  We're  
working hard to make this available to contributors.  In the  
meantime, we could work closely with you to get it loaded.

We should chat more (off this email list, so not to bore people)  
about the structure of your data.

> * Would it be possible to integrate editing / submitting functionality
>   on external sites?  Something better than just a link to the
>   freebase edit page might be interesting.
	
Surprisingly, Freebase is actually an "external site" in that it uses  
the same APIs that anybody else should use.  Our current  
functionality is a bit complex to package up for others to use, so we  
will be putting together some functions (autocomplete in particular)  
to make it easier for developers to create write-enabled applications.

> * How about OpenID authentication for contributors?  This would
>   obviously be something for post-alpha.

We've have a couple of questions about OpenID support.  We are fans  
of open standards and we'll try to support them to the extent we can.

> In any case, glad to see this effort and am excited about the
> possibilities that it offers.

Thank you.  Please keep the questions coming.

Robert


> -- 
> Paul Lindner        ||||| | | | |  |  |  |   |   |
> lindner at inuus.com
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From tim at metaweb.com  Tue Jul  3 00:36:24 2007
From: tim at metaweb.com (Tim Kientzle)
Date: Mon, 2 Jul 2007 17:36:24 -0700
Subject: [Developers] Schools, Colleges, Data loading and more...
In-Reply-To: <20070702232019.GP19982@inuus.com>
References: <20070702232019.GP19982@inuus.com>
Message-ID: <6D3FCB1F-6142-4947-9B02-B85D071EC11F@metaweb.com>

On Jul 2, 2007, at 4:20 PM, Paul Lindner wrote:
> I was recently invited to the Freebase alpha.  I like what I see!  I
> work for Hi5 and we can see a ton of potential here.
>
> Couple of questions though.
>
> * What about internationalization?  Are there facilities for storing
>   different names based on language?

As Chris Madsen mentioned, our underlying data model already
supports full internationalization (a "name" is the value of
a link between an object and a language; any object can have
names in any number of languages).  We do plan to fully
internationalize our own applications, but applications
using our public APIs can be fully internationalized today.

> * Is there any planned integration with geonames.org?

I'll let someone from our data group answer this.

> * Hi5 has a large database of schools we could contribute.  ...
> What's the best way to do that?

I'll let someone from our data group address that question
in detail, but I can say that other developers are building
tools against our public API to push data into our system.

We also have a number of internal tools and development libraries
that we plan to eventually expose for broad use.  The first such
is the MJT templating system http://mjtemplate.org/ that makes it
very easy to build browser-based applications that run against our  
system.

> * Would it be possible to integrate editing / submitting functionality
>   on external sites?  Something better than just a link to the
>   freebase edit page might be interesting.

Our API includes write facilities that developers can use
to integrate editing capabilities into externally-hosted
applications.  To limit abuse, we do have some default
restrictions on bulk writes, but we're eager to work with
outside developers.  Please let us know what you need.

> * How about OpenID authentication for contributors?  This would
>   obviously be something for post-alpha.

We're looking into it.

Tim Kientzle


From niallo at metaweb.com  Tue Jul  3 00:55:55 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 2 Jul 2007 17:55:55 -0700
Subject: [Developers] Sandbox update
In-Reply-To: <20070703000737.GO5935@digdug.inet>
References: <20070703000737.GO5935@digdug.inet>
Message-ID: <20070703005555.GR5935@digdug.inet>

Update complete.

On Mon, Jul 02, 2007 at 05:07:37PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Data refresh and software update on 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 nix at metaweb.com  Tue Jul  3 02:28:17 2007
From: nix at metaweb.com (Nick Thompson)
Date: Mon, 02 Jul 2007 19:28:17 -0700
Subject: [Developers] mjt.def
In-Reply-To: <19E07FD9-D6CB-42E2-A70F-59FC6D5127C5@metaweb.com>
References: <19E07FD9-D6CB-42E2-A70F-59FC6D5127C5@metaweb.com>
Message-ID: <4689B441.4090602@metaweb.com>

Dae Park wrote:
 > Is there anyway to share "mjt.def" declarations across multiple templates?

in the most recent version of mjt, mjt.run() returns a javascript object
that contains all the toplevel mjt.def= declarations from the template.
so you can do:

<THIS IS UNTESTED>
<div id="mytemplate">
   <div mjt.dev="mydef(arg1)">
     <b>$arg1</b>
   </div>
</div>
</THIS IS UNTESTED>

then later, in js:
   window.mytemplate = mjt.run('mydiv');
then you can use something like ${mytemplate.mydef(...)} in other
templates.

     nick

From daepark at metaweb.com  Tue Jul  3 02:50:13 2007
From: daepark at metaweb.com (Dae Park)
Date: Mon, 2 Jul 2007 19:50:13 -0700
Subject: [Developers] mjt.def
In-Reply-To: <4689B441.4090602@metaweb.com>
References: <19E07FD9-D6CB-42E2-A70F-59FC6D5127C5@metaweb.com>
	<4689B441.4090602@metaweb.com>
Message-ID: <EE9638BD-0CCF-46EF-8046-96DC1B4DAB53@metaweb.com>

I was thinking more along the lines of including a javasript library  
and maybe something similar could be applied to a library of mjt.def  
declarations.

Would you use a hidden frame to load your mjt.def declarations and  
have your visible (template) frame wait for it to load, set the  
mytemplate window variable and then call mjt.run on itself?





On Jul 2, 2007, at 7:28 PM, Nick Thompson wrote:

> Dae Park wrote:
> > Is there anyway to share "mjt.def" declarations across multiple  
> templates?
>
> in the most recent version of mjt, mjt.run() returns a javascript  
> object
> that contains all the toplevel mjt.def= declarations from the  
> template.
> so you can do:
>
> <THIS IS UNTESTED>
> <div id="mytemplate">
>   <div mjt.dev="mydef(arg1)">
>     <b>$arg1</b>
>   </div>
> </div>
> </THIS IS UNTESTED>
>
> then later, in js:
>   window.mytemplate = mjt.run('mydiv');
> then you can use something like ${mytemplate.mydef(...)} in other
> templates.
>
>     nick


From nix at metaweb.com  Tue Jul  3 02:50:26 2007
From: nix at metaweb.com (Nick Thompson)
Date: Mon, 02 Jul 2007 19:50:26 -0700
Subject: [Developers] mjt html generation
In-Reply-To: <B36DCA18-3679-447E-942E-FAAAA4B6B5C2@metaweb.com>
References: <B36DCA18-3679-447E-942E-FAAAA4B6B5C2@metaweb.com>
Message-ID: <4689B972.9030509@metaweb.com>

nope - all the templates are expanded into a single string, which is
then pasted in using innerHTML.  mjt.script="" is evaluated as the
string is being created, well before it gets turned into DOM elements.
this is usually what you want, but i've run into the situation you're
describing too.

my way of dealing with it has to set up a short timeout - in most cases
'bar' should be ready immediately after the current mjt.run() completes,
and the javascript interpreter won't run timeouts until then.  this
doesn't feel clean or robust but it's worked for me so far.

in the future i'm planning to add something like mjt.script="ondomready"
to declare that the script should execute after the dom is ready, but
before mjt.run() returns.  this should be simpler to use, more reliable,
and more efficient.  but i'm only going to be on email sporadically for
the next few weeks, so it will probably show up in mjt 0.6 sometime in
early to mid-august.

     nick

Dae Park wrote:
> Is the mjt template generation synchronous? For example if I call a 
> mjt.def, will I be able to access the element created in the mjt.def in 
> a mjt.script:
> 
> <div mjt.def="foo()">
>   <p id="bar">Hello world</p>
> </div>
> <div>
>   ${foo()}
>   <pre mjt.script="">
>     alert(document.getElementById("bar"));
>   </pre>
> </div>

From nix at metaweb.com  Tue Jul  3 03:21:31 2007
From: nix at metaweb.com (Nick Thompson)
Date: Mon, 02 Jul 2007 20:21:31 -0700
Subject: [Developers] mjt.def
In-Reply-To: <EE9638BD-0CCF-46EF-8046-96DC1B4DAB53@metaweb.com>
References: <19E07FD9-D6CB-42E2-A70F-59FC6D5127C5@metaweb.com>
	<4689B441.4090602@metaweb.com>
	<EE9638BD-0CCF-46EF-8046-96DC1B4DAB53@metaweb.com>
Message-ID: <4689C0BB.3020106@metaweb.com>

 > I was thinking more along the lines of including a javasript library and
 > maybe something similar could be applied to a library of mjt.def
 > declarations.

there is an undocumented library mechanism that might be you want in this
case.  be warned: it's undocumented because it is confusing and has only
been used in very restricted circumstances so far.  fetching a library
from a url is an asynchronous operation, so evaluation order can be
surprising.

to use library inclusion, you must be using mjt.run() with no arguments
to expand the entire <body> as a template.  if you do this, mjt.run will
also check the <head> for link tags.  if mjt.run() finds a tag of the form:

   <link rel="x-mjt-import" title="mylib" href="./mylib.html"></link>

then it will automatically fetch mylib.html using a dynamically generated
iframe.  when it detects that mylib.html has been loaded, it compiles
the <body> of mylib.html first, and then completes the mjt.run() on the
<body> of the current page.  the link title= attribute declares a variable
that will be bound to a namespace containing the toplevel mjt.def=""
calls from mylib.html.  so you can use ${mylib.mydef('xyz')} to refer
to the template function mydef inside mylib.html.

it's best not to include any <script> tags in mylib.html, since there is
no protection against including the same script twice.  keep the <script>
tags in the main .html file.

note that because of the asynchronous fetch, mjt.run() will return control
to the caller *before it has actually expanded the current page*.  there
is no way around this: it is an inevitable consequence of fetching a url
using javascript.  but you shouldn't notice it unless you do something after
mjt.run(); in your onload handler.

according to what i've found on the net, it should work across browsers but
i haven't tested.

use at your own risk :-)  the details of this will almost certainly change
in a future release.

     nick

Dae Park wrote:
> I was thinking more along the lines of including a javasript library and 
> maybe something similar could be applied to a library of mjt.def 
> declarations.
> 
> Would you use a hidden frame to load your mjt.def declarations and have 
> your visible (template) frame wait for it to load, set the mytemplate 
> window variable and then call mjt.run on itself?
> 
> 
> 
> 
> 
> On Jul 2, 2007, at 7:28 PM, Nick Thompson wrote:
> 
>> Dae Park wrote:
>> > Is there anyway to share "mjt.def" declarations across multiple 
>> templates?
>>
>> in the most recent version of mjt, mjt.run() returns a javascript object
>> that contains all the toplevel mjt.def= declarations from the template.
>> so you can do:
>>
>> <THIS IS UNTESTED>
>> <div id="mytemplate">
>>   <div mjt.dev="mydef(arg1)">
>>     <b>$arg1</b>
>>   </div>
>> </div>
>> </THIS IS UNTESTED>
>>
>> then later, in js:
>>   window.mytemplate = mjt.run('mydiv');
>> then you can use something like ${mytemplate.mydef(...)} in other
>> templates.
>>
>>     nick
> 

From niallo at metaweb.com  Tue Jul  3 23:41:59 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 3 Jul 2007 16:41:59 -0700
Subject: [Developers] www.freebase.com software update
Message-ID: <20070703234159.GE8944@digdug.inet>

Hi,

Weekly software update of www.freebase.com happening over the next few
minutes.

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

From niallo at metaweb.com  Wed Jul  4 00:21:05 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 3 Jul 2007 17:21:05 -0700
Subject: [Developers] www.freebase.com software update
In-Reply-To: <20070703234159.GE8944@digdug.inet>
References: <20070703234159.GE8944@digdug.inet>
Message-ID: <20070704002105.GG8944@digdug.inet>

Update complete.

On Tue, Jul 03, 2007 at 04:41:59PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Weekly software update of www.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 will at hamstersoup.com  Thu Jul  5 11:13:14 2007
From: will at hamstersoup.com (Will Moffat)
Date: Thu, 5 Jul 2007 13:13:14 +0200
Subject: [Developers] Documentation urgently needs a wiki + Getting started
	with PHP
Message-ID: <2f665fe20707050413x41ba3fean1c8808a024ecc5c9@mail.gmail.com>

I gave a demo of Freebase to some great Belgium PHP developers. They
totally got it and loved the concept. One of them was so fired up that
he started working through the tutorial at 4am.

It was not a positive experience.

Some of the problems:
* There is no visible feedback mechanism for the documentation pages.
At the very least a wiki or discussion system needs to be added so
that people share their problems. Even if all they do is to add a note
"This doesn't work" then that will save lots of pain and bad feeling.
* Some of the tutorials no longer work. It's very disheartening to a
new developer when the tutorials are buggy. Hello world apps must just
work out of the box. Developers new to Freebase just don't have the
experience to fix simple bugs and will just give up.
* It's easy to mess up cookie authentication, we need to emphasize
that developers should test this first.
* "Developing Metaweb-Enabled Web Applications" needs to be updated
with the API changes (results codes have changed which means examples
like albumlist.php are totally foobar)
* The tutorial apps are simple (which is good) but too fragile.
albumlist.php fails with PHP errors for queries like 'R.E.M' (missing
last period) or 'The Beatles' (more than one matching band). Somebody
new to Freebase will have no idea what went wrong.
* The help doesn't seem to be searchable. (It would be great if it was
exposed as static html without authentication and crawled by  Google.)
* If you receive an invitation, register but then lose the
confirmation email then you cannot access your account. The "I forgot
my password" does not seem to work in that situation.

Other minor issues:
* It would be great if the Query Editor generated queries as a mqlread
URL, a PHP query and a Perl query as well as JSON.
* Need a better error message if you call
http://freebase.com/api/service/mqlread without arguments. The error
should be 'missing query' not  the confusing "authentication required"

On a positive note:

I made a start on trying to improve the PHP documentation. See
http://hamstersoup.wordpress.com/
* Created simple shell and PHP scripts to test your Freebase cookie.
* Fixed albumlist.php, made it more robust and added a 'show query' option.
* Wrote step by step instructions for getting albumlist.php running

Your feedback would be very welcome.
regards,
--Will

PS It was the great documentation that got me into Freebase in the
first place. It's worth updating it!

From jeff at metaweb.com  Thu Jul  5 23:11:04 2007
From: jeff at metaweb.com (Jeff Prucher)
Date: Thu, 5 Jul 2007 16:11:04 -0700
Subject: [Developers] Moving types out of the /common domain
Message-ID: <001e01c7bf59$d4fa8190$bc01a8c0@p4>

The /common domain currently contains types that serve two different
purposes, which is starting to become a problem. The two kinds of types are
those, such as /common/topic, that are important to the workings of Freebase
itself, and those, such as /common/phone_number, that are used by a variety
of types in several domains but which are not necessary to the functioning
of Freebase.  We would like to create a new domain (/shared) for the second
kind of types (the ones that are merely used across domains) to make it
easier to manage the types that are crucial to the system.

Moving types around like this is something that we are still doing in Alpha
(though as little as possible), and we are working on ways to make addresses
stable.  We will make the following changes on Monday, July 9, if we don't
hear any objections.

The types currently in /common that will be moved to /shared are:
award
award_winning_work
award-winner
award_presenting_organization
award_nominee
award_nominated_work
award_nomination
award_honor
citation
collection
legal_entity
phone_number
review
weekday

The types that will remain in /common are:
topic
document
image
attribution_info
licensed_object
license
webpage

Additionally, /common/interpersonal_relationship will be deleted entirely.

Jeff Prucher
Type Librarian & Ontologist
Metaweb, Inc.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070705/f0b96162/attachment.htm 

From william at scissor.com  Fri Jul  6 00:49:57 2007
From: william at scissor.com (William Pietri)
Date: Thu, 05 Jul 2007 17:49:57 -0700
Subject: [Developers] Documentation urgently needs a wiki + Getting
 started with PHP
In-Reply-To: <2f665fe20707050413x41ba3fean1c8808a024ecc5c9@mail.gmail.com>
References: <2f665fe20707050413x41ba3fean1c8808a024ecc5c9@mail.gmail.com>
Message-ID: <468D91B5.1000905@scissor.com>

Will Moffat wrote:
> * Some of the tutorials no longer work. It's very disheartening to a
> new developer when the tutorials are buggy. Hello world apps must just
> work out of the box. Developers new to Freebase just don't have the
> experience to fix simple bugs and will just give up.
> [...]
> * "Developing Metaweb-Enabled Web Applications" needs to be updated
> with the API changes (results codes have changed which means examples
> like albumlist.php are totally foobar)


Sounds like these would be good to have as automated test cases, so you 
they don't have to be manually rechecked after each new release.

William

From dmilbrath at metaweb.com  Fri Jul  6 06:30:49 2007
From: dmilbrath at metaweb.com (Dan Milbrath)
Date: Fri, 6 Jul 2007 02:30:49 -0400 (EDT)
Subject: [Developers] Documentation urgently needs a wiki + Getting
 started with PHP
In-Reply-To: <468D31D7.6040902@metaweb.com>
Message-ID: <31866932.2421183703449085.JavaMail.root@h00215>

Will,
thanks for the useful feedback. Comments below.


>> From: "Will Moffat" <will at hamstersoup.com>
>> Date: July 5, 2007 4:13:14 AM PDT
>> To: developers at freebase.com
>> Subject: [Developers] Documentation urgently needs a wiki + Getting 
>> started with PHP
>>
>>
>> Some of the problems:
>> * There is no visible feedback mechanism for the documentation pages.
>> At the very least a wiki or discussion system needs to be added so
>> that people share their problems. Even if all they do is to add a note
>> "This doesn't work" then that will save lots of pain and bad feeling.

We're in the process of making the API into a series of Freebase topics, complete with discuss functionality. 

You can see this here: http://www.freebase.com/view/helptopic?id=%239202a8c04000641f800000000544e14d


>> * Some of the tutorials no longer work. It's very disheartening to a
>> new developer when the tutorials are buggy. Hello world apps must just
>> work out of the box. Developers new to Freebase just don't have the
>> experience to fix simple bugs and will just give up.

We'll try get these fixed shortly. 


>> * The help doesn't seem to be searchable. (It would be great if it was
>> exposed as static html without authentication and crawled by Google.)

We're working on addressing that in an upcoming release.






From jphekman at arborius.net  Fri Jul  6 21:58:20 2007
From: jphekman at arborius.net (Jessica Perry Hekman)
Date: Fri, 06 Jul 2007 17:58:20 -0400
Subject: [Developers] Possible mjt bug
Message-ID: <20070706215820.GB16488@arborius.net>

I may have discovered a bug in mjt. See

 http://pendaran.arborius.net:8080/~jphekman/freebase/ifelse.html

displaying the results of

 <div mjt.if="true">
   if=true
 </div>
 <div mjt.else="">
   else
 </div>

which I expect to display

 if=true

as per

 http://www.mjtemplate.org/tour/intro.html#mjt-else

but which instead displays

 if=true
 else

...If it is not a bug, I would love feedback on how I misinterpreted the 
doc.

I'm enjoying mjt, except for how hard it is to tell if the divs are 
properly balanced; I seem to be completely dependent on emacs telling me 
which curly brace I'm closing, and of course when it is closing a div it 
doesn't share that piece of information.
 
Jessica


From crism at metaweb.com  Fri Jul  6 22:19:03 2007
From: crism at metaweb.com (Christopher R. Maden)
Date: Fri, 06 Jul 2007 15:19:03 -0700
Subject: [Developers] Possible mjt bug
In-Reply-To: <20070706215820.GB16488@arborius.net>
References: <20070706215820.GB16488@arborius.net>
Message-ID: <468EBFD7.8030009@metaweb.com>

Jessica Perry Hekman wrote:
> I may have discovered a bug in mjt. See
> 
>  http://pendaran.arborius.net:8080/~jphekman/freebase/ifelse.html

I think the problem is that you downloaded a local copy of mjt.js from 
mjt 0.4.1, and mjt.else was added in mjt 0.5.  Try including mjt.js 
directly from mjtemplate.org or getting a fresh copy.

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

From jphekman at arborius.net  Fri Jul  6 22:41:50 2007
From: jphekman at arborius.net (Jessica Perry Hekman)
Date: Fri, 06 Jul 2007 18:41:50 -0400
Subject: [Developers] Possible mjt bug
In-Reply-To: <468EBFD7.8030009@metaweb.com>
References: <20070706215820.GB16488@arborius.net>
	<468EBFD7.8030009@metaweb.com>
Message-ID: <20070706224149.GC16488@arborius.net>

Up to date documentation? Who would have thought.

That worked; thanks.

Jessica

On Fri, Jul 06, 2007 at 03:19:03PM -0700, Christopher R. Maden wrote:
> Jessica Perry Hekman wrote:
> >I may have discovered a bug in mjt. See
> >
> > http://pendaran.arborius.net:8080/~jphekman/freebase/ifelse.html
> 
> I think the problem is that you downloaded a local copy of mjt.js from 
> mjt 0.4.1, and mjt.else was added in mjt 0.5.  Try including mjt.js 
> directly from mjtemplate.org or getting a fresh copy.
> 
> ~Chris
> -- 
> Christopher R. Maden
> Data Architect
> Metaweb Technologies, Inc.
> <URL: http://www.metaweb.com/ >

From dmilbrath at metaweb.com  Sat Jul  7 02:25:57 2007
From: dmilbrath at metaweb.com (Dan Milbrath)
Date: Fri, 6 Jul 2007 22:25:57 -0400 (EDT)
Subject: [Developers] new feature for schema designers: property hints
Message-ID: <8999526.4471183775157098.JavaMail.root@h00215>

Thought I'd give a shout out about a newly introduced feature. Property hints allow schema designers to provide field-level help. You can define these help descriptions from within the property definition interface in type editor - and the text then appears as a 'tool tip' when users are inputting data in the UI.

For more on this feature, you can check out the help topic here:
http://www.freebase.com/view/helptopic?id=%239202a8c04000641f8000000005453789

While I'll continue to update this list with significant features, release notes are also posted with each release (typically weekly) here:
http://www.freebase.com/view/helptopic?id=%239202a8c04000641f8000000003c196f0

Cheers,
Dan (danm)

From will at hamstersoup.com  Sun Jul  8 09:27:39 2007
From: will at hamstersoup.com (Will Moffat)
Date: Sun, 8 Jul 2007 11:27:39 +0200
Subject: [Developers] Possible mjt bug
In-Reply-To: <20070706215820.GB16488@arborius.net>
References: <20070706215820.GB16488@arborius.net>
Message-ID: <2f665fe20707080227jb494262p97c9b204aebcc7b0@mail.gmail.com>

> I'm enjoying mjt, except for how hard it is to tell if the divs are
> properly balanced; I seem to be completely dependent on emacs telling me
> which curly brace I'm closing, and of course when it is closing a div it
> doesn't share that piece of information.

Hi Jessica (and other Emacs fans),

I've had the same pain of unclosed div's and your email prompted me to
try Nick's suggestion of using Emacs nxml-mode. It's amazing, I wish I
had started using it earlier.

Here are my step by step instructions on how to get going with Emacs
nXhtml-mode (which includes nxml-mode)
http://hamstersoup.wordpress.com/

Please let me know if you find it helpful.
regards,
--Will

From will at hamstersoup.com  Mon Jul  9 13:56:09 2007
From: will at hamstersoup.com (Will Moffat)
Date: Mon, 9 Jul 2007 15:56:09 +0200
Subject: [Developers] cursors, limits, MJT and chocolates
Message-ID: <2f665fe20707090656k53583244y47e6ecb94c7b5bfd@mail.gmail.com>

Hi,

My map of freebase users is starting to give timeouts. Before I dive
in and start hacking I wonder if anybody could save me some time:
		
If I use a cursor can I still get a timeout?
Or does the use of a cursor guarantee a (possibly empty) set on each call?

Can I use 'limit' to fix the set size with a cursor?
Or does that refer to the total result limit?

Extra Belgian chocolates if somebody has an example of using the
cursor feature with MJT.

regards,
--Will

From kurt at metaweb.com  Mon Jul  9 16:19:46 2007
From: kurt at metaweb.com (Kurt Bollacker)
Date: Mon, 9 Jul 2007 09:19:46 -0700
Subject: [Developers] cursors, limits, MJT and chocolates
In-Reply-To: <2f665fe20707090656k53583244y47e6ecb94c7b5bfd@mail.gmail.com>
References: <2f665fe20707090656k53583244y47e6ecb94c7b5bfd@mail.gmail.com>
Message-ID: <20070709161946.GC6818@metaweb.com>

On Mon, Jul 09, 2007 at 03:56:09PM +0200, Will Moffat wrote:
> Hi,
> 
> My map of freebase users is starting to give timeouts. Before I dive
> in and start hacking I wonder if anybody could save me some time:
> 		
> If I use a cursor can I still get a timeout?

Yes, if the query is complex enough or is returning too many
sub-results.

> Or does the use of a cursor guarantee a (possibly empty) set on each call?

Currently, there's no guarantee of avoiding a timeout.  However, it's
not likely if you break your results into small enough pieces.

> Can I use 'limit' to fix the set size with a cursor?

When using cursors, "limit" at the top level of a MQL query acts like
you think "pagesize" should.  You can reduce the "limit" value to the
point where timeouts no longer occur.

{
 "query":
     [{"id":null,
       "name":null,
       "type":"/music/artist",
       "limit":10}],
 "cursor":true
}

Will iterate through music artists, 10 at a time.  When you get a
cursor, place it back into the "cursor" clause to get the next page.

It timeouts still occur when you set "limit":1, then you probably have
a very complex query or one that is trying to return a large number of
sub-results (e.g. iterating through publishers and giving all of their
books published as a sub-result).  If your query is simple, returns a
small amount of data, and "limit":1 still gives you a timeout, then we
have a bug somewhere and this should be reported.

> Or does that refer to the total result limit?

It's only the total limit when there are no cursors or "limit" appears
somewhere in the MQL query other than the top level.
 
> Extra Belgian chocolates if somebody has an example of using the
> cursor feature with MJT.

This should be possible, but I haven't done it. Nix?

							Kurt :-)

From darin at metaweb.com  Mon Jul  9 19:23:12 2007
From: darin at metaweb.com (Darin Wilson)
Date: Mon, 9 Jul 2007 12:23:12 -0700
Subject: [Developers] cursors, limits, MJT and chocolates
In-Reply-To: <2f665fe20707090656k53583244y47e6ecb94c7b5bfd@mail.gmail.com>
References: <2f665fe20707090656k53583244y47e6ecb94c7b5bfd@mail.gmail.com>
Message-ID: <444718DB-70B7-47D2-BA74-7A4BC1D1FB22@metaweb.com>

> Extra Belgian chocolates if somebody has an example of using the
> cursor feature with MJT.

That's some good incentive. ;-)

Below is a crude example. Cursors aren't in the official MJT docs, so  
I discovered this method by looking at the mjt source. Some of the  
implementation details could change (and Nix please correct me if  
this is not the right way to do this).

Note the "mjt.if='q.result.length > 0'" in the "ready" block.  
Normally, "next_cursor" will be null if there are no more results to  
read, but there is currently a known bug that prevents this from  
working correctly. Getting an empty result is the only way to be sure  
you've reached the end.

I hope this helps,
Darin

-----------------------------------------------------------
(headers omitted for brevity)


<body onload="mjt.run()" style="display:none">

<div mjt.script="" mjt.strip="1">
   var query = {
     'id':null,
     'name':null,
     'type':'/music/artist',
     'limit':20
   }

   var query_args = {}
   if (mjt.urlquery.cursor != null) {
     // use the cursor returned in the previous result
     query_args.cursor = mjt.urlquery.cursor
   }
</div>

<div mjt.task="q">
   mjt.mqlread([query], query_args)
</div>

<div mjt.choose="q.state">
   <div mjt.when="ready">
     <div mjt.if="q.result.length > 0">
         <h4>music artists in freebase:</h4>
         <ul>
            <li mjt.for="artist in q.result">
              $artist.name
            </li>
         </ul>
         <a href="?cursor=$q.next_cursor">next --></a>
     </div>
     <div mjt.else="">
         end of result set
     </div>
   </div>
   <div mjt.when="wait">
       loading...
   </div>
   <div mjt.when="error">
     error:
     <div mjt.for="msg in q.messages">
       $msg.message
     </div>
   </div>
</div>
</body>

From mikelove1 at gmail.com  Mon Jul  9 23:53:21 2007
From: mikelove1 at gmail.com (Mike Love)
Date: Mon, 9 Jul 2007 16:53:21 -0700
Subject: [Developers] uploaded my data and broke mjt inline example
Message-ID: <f121137a0707091653n40a29c03t9628a1937c150485@mail.gmail.com>

I finally got around to formatting and uploading a dataset I'd been working
on, about creative influences through Western history.  The documentation +
Query Editor is really great, so thanks to those who put that together.  I
was able to use the Wikipedia URL keys to identify people, which saved a lot
of time.

Nick: since I changed the type name from "Historic Figure" to "Influence
Node", I broke the inline example for mjt that you put together.  Should be
able to change historic_figure to influence_node and it will work.

Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freebase.com/pipermail/developers/attachments/20070709/03528124/attachment.htm 

From niallo at metaweb.com  Tue Jul 10 00:10:17 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 9 Jul 2007 17:10:17 -0700
Subject: [Developers] sandbox.freebase.com weekly update
Message-ID: <20070710001017.GC7284@digdug.inet>

Hi,

Data refresh and software update happening over the next few minutes.

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

From niallo at metaweb.com  Tue Jul 10 00:44:57 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 9 Jul 2007 17:44:57 -0700
Subject: [Developers] sandbox.freebase.com weekly update
In-Reply-To: <20070710001017.GC7284@digdug.inet>
References: <20070710001017.GC7284@digdug.inet>
Message-ID: <20070710004457.GE7284@digdug.inet>

Update complete.

On Mon, Jul 09, 2007 at 05:10:17PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Data refresh and software update 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 crism at metaweb.com  Tue Jul 10 03:15:34 2007
From: crism at metaweb.com (Christopher R. Maden)
Date: Mon, 09 Jul 2007 20:15:34 -0700
Subject: [Developers] Proposed changes in the music domain
In-Reply-To: <4684823A.2070302@metaweb.com>
References: <4684823A.2070302@metaweb.com>
Message-ID: <4692F9D6.8070607@metaweb.com>

The changes described below have been made on the main Freebase site.  I 
couldn?t find any Freebase applications that used these properties, but 
please let us know if anything breaks.

One addition to these changes is that a Composition may be recorded as a 
Musical Album (e.g., a symphony that shows up as multiple tracks on the 
album).

Christopher R. Maden wrote:
> If you care about data in the music domain, please take a moment to look 
> at it on sandbox: <URL: 
> http://sandbox.freebase.com/view/domain?id=/music >.
> 
> If you don?t care about the music domain, please skip the rest of this 
> e-mail.
> 
> The main change here is the new ?Composition? type, which has 
> ?Composer,? ?Lyricist,? and ?Recorded Versions? properties.  A ?Song? is 
> now just a kind of ?Composition? (by way of an included type).
> 
> A composition can include another composition ? so a symphony might 
> include its movements, a rock opera can include its songs, an operatic 
> cycle can include its operas.  These different semantically-meaningful 
> types can be added as the data is fleshed out.
> 
> This would affect a couple of other domains, too; a musical would be 
> both a ?Play? and a ?Composition,? with the composer and lyricist using 
> the ?Composition? properties and with its songs being compositions 
> themselves included in the larger work; similarly for operas.  (I do 
> realize that there is a connotative difference between a librettist and 
> a lyricist, but that is more a matter of era than anything else; which 
> does _Tommy_ have?)
> 
> Another new type is ?Arrangement,? which is, like ?Song,? itself a 
> ?Composition,? but which is also an arrangement of some other 
> composition.  This allows us to distinguish between covers of Jimi 
> Hendrix?s ?All Along the Watchtower? and Bob Dylan?s.
> 
> I am posting this same message to the discussion in the Music domain on 
> the main Freebase site, <URL: 
> http://www.freebase.com/discuss/view?id=/music#%239202a8c04000641f80000000051fcbc7 
>  >; please reply with comments there.

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

From synedra at gmail.com  Tue Jul 10 16:31:25 2007
From: synedra at gmail.com (Kirsten Jones)
Date: Tue, 10 Jul 2007 09:31:25 -0700
Subject: [Developers] sandbox.freebase.com weekly update
In-Reply-To: <20070710004457.GE7284@digdug.inet>
References: <20070710001017.GC7284@digdug.inet>
	<20070710004457.GE7284@digdug.inet>
Message-ID: <964D9795-0E4C-499D-B4D9-55EF72D3386B@gmail.com>

Hi all,

I'm trying to create an MJT implementation similar to the FMDb, but  
I'm running into an issue where it's not behaving correctly for me.

I have this, which probably looks suspiciously like the same code  
from FMDb:

<div mjt.task="image">
	mjt.mqlread({
		"id":mjt.urlquery.game,
		"type":[{}],
		"name":null,
		"/common/topic/image":({
			"id":null,
			"timestamp":null,
			"optional":true,
			"limit":1
		})
	})
</div>

<img alt="image" src="#" mjt.src="${mjt.imgurl(image.result['/common/ 
topic/image'].id, 265, 1000)}" />

For games with an image, this causes the image to appear, but I get  
an error:

error applying tfunc main() [object Object]
     TypeError: image.result has no properties
             										"limit":1
            									})
            								})
            							));
            __m[__n++]=__ts[52];
---174-->  if (image.result['/common/topic/image'].length) {
            __m[__n++]=__ts[53];
            __m[__n++]=(mjt.imgurl(image.result['/common/topic/ 
image'].id, 265, 1000));
            __m[__n++]=__ts[54];
            }

For games with no image, I just get the error, no image, and it  
doesn't display the rest of the stuff.

If I add the logic of <div mjt.if="image.result['/common/topic/ 
image'].length"> I don't get the picture ever, but I still get the  
error.

I assume I'm missing something obvious.  Help?

TIA,

Kirsten


From tsturge at metaweb.com  Tue Jul 10 19:00:44 2007
From: tsturge at metaweb.com (Tim Sturge)
Date: Tue, 10 Jul 2007 12:00:44 -0700
Subject: [Developers] sandbox.freebase.com weekly update
In-Reply-To: <964D9795-0E4C-499D-B4D9-55EF72D3386B@gmail.com>
References: <20070710001017.GC7284@digdug.inet>	<20070710004457.GE7284@digdug.inet>
	<964D9795-0E4C-499D-B4D9-55EF72D3386B@gmail.com>
Message-ID: <4693D75C.6080309@metaweb.com>

Hi Kirsten,

There's actually three separate cases here:

1) games with an image. The result should look something like

image.result=
{
"id":"#92...."
"type":[{ "name": "Game", 
           "id": "/game/game" 
       }],
"name":"SimCity",
"/common/topic/image":{
	"id":"#92...",
	"timestamp":"2007-06-01T..."
}
}

2) games without an image:

image.result=
{
"id":"#92...."
"type":[{ "name": "Game", 
           "id": "/game/game" 
       }],
"name":"SimCity",
"/common/topic/image":null
}

3) things that aren't there at all (mjt.urlquery.game points to 
something that doesn't exist)

image.result=null

(this is what seems to be causing the error you are seeing)

So you do need to check if image.result['/common/topic/image'] is null 
or not (to catch case 2) but you might also want to check if
image.result itself is null or not (to catch case 3). Of course if 
image.result is null, there's nothing to display at all.

If you're seeing case 3 a lot, it might be interesting to examine the 
"id" field you are getting from mjt.urlquery.game; it should either be 
"#92..." (that is 32 hex digits with a leading # sign) or 
"/wikipedia/en/SimCity" (that is a / separated list of words like an 
absolute UNIX pathname)

I hope this helps,

Tim


Kirsten Jones wrote:
> Hi all,
>
> I'm trying to create an MJT implementation similar to the FMDb, but  
> I'm running into an issue where it's not behaving correctly for me.
>
> I have this, which probably looks suspiciously like the same code  
> from FMDb:
>
> <div mjt.task="image">
> 	mjt.mqlread({
> 		"id":mjt.urlquery.game,
> 		"type":[{}],
> 		"name":null,
> 		"/common/topic/image":({
> 			"id":null,
> 			"timestamp":null,
> 			"optional":true,
> 			"limit":1
> 		})
> 	})
> </div>
>
> <img alt="image" src="#" mjt.src="${mjt.imgurl(image.result['/common/ 
> topic/image'].id, 265, 1000)}" />
>
> For games with an image, this causes the image to appear, but I get  
> an error:
>
> error applying tfunc main() [object Object]
>      TypeError: image.result has no properties
>              										"limit":1
>             									})
>             								})
>             							));
>             __m[__n++]=__ts[52];
> ---174-->  if (image.result['/common/topic/image'].length) {
>             __m[__n++]=__ts[53];
>             __m[__n++]=(mjt.imgurl(image.result['/common/topic/ 
> image'].id, 265, 1000));
>             __m[__n++]=__ts[54];
>             }
>
> For games with no image, I just get the error, no image, and it  
> doesn't display the rest of the stuff.
>
> If I add the logic of <div mjt.if="image.result['/common/topic/ 
> image'].length"> I don't get the picture ever, but I still get the  
> error.
>
> I assume I'm missing something obvious.  Help?
>
> TIA,
>
> Kirsten
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers
>   


From niallo at metaweb.com  Tue Jul 10 22:51:21 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 10 Jul 2007 15:51:21 -0700
Subject: [Developers] freebase.com software update
Message-ID: <20070710225121.GO7284@digdug.inet>

Hi,

Weekly software update of www.freebase.com will be happening over the
next few minutes.

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

From niallo at metaweb.com  Tue Jul 10 23:25:46 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 10 Jul 2007 16:25:46 -0700
Subject: [Developers] freebase.com software update
In-Reply-To: <20070710225121.GO7284@digdug.inet>
References: <20070710225121.GO7284@digdug.inet>
Message-ID: <20070710232546.GR7284@digdug.inet>

Update complete.

On Tue, Jul 10, 2007 at 03:51:21PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Weekly software update of www.freebase.com will be 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 crism at metaweb.com  Wed Jul 11 00:07:46 2007
From: crism at metaweb.com (Christopher R. Maden)
Date: Tue, 10 Jul 2007 17:07:46 -0700
Subject: [Developers] More proposed music changes
Message-ID: <46941F52.6090000@metaweb.com>

No, we won?t be sending mail like this every time we change any schema, 
but there?s so much music data and so much interest in it, we want to 
make sure we don?t break anything.

There are two proposed changes implemented on sandbox.freebase.com.

There?s one new type, ?Musical Group,? which now holds the group members 
property, instead of that being a property of ?Musical Artist.?

There?s another new type, ?Multipart Album,? which is used to model 
albums with more than one disc.  The actual track listings are part of 
the member discs, but the multipart album is what shows up in an 
artist?s discography.

Take a look at Pink Floyd (<URL: 
http://sandbox.freebase.com/view/?id=%239202a8c04000641f8000000003bd2794 
 >) for an example.  They?re now a ?Musical Group? as well as a ?Musical 
Artist,? and their albums _Ummagumma_, _The Wall_, _Delicate Sound of 
Thunder_, and the box set _Shine On_ are now modeled as ?Multipart Albums.?

What do you think?  Will this break any of your applications?  Please 
comment on the Freebase Music domain discussion page.

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

From will at hamstersoup.com  Thu Jul 12 17:26:31 2007
From: will at hamstersoup.com (Will Moffat)
Date: Thu, 12 Jul 2007 19:26:31 +0200
Subject: [Developers] Some notes on basic MQL and reflection
Message-ID: <2f665fe20707121026t35d09e77j1c8f87a32d797545@mail.gmail.com>

Hello,

Developers learning MQL might be interested in some notes I put together:
Step by step MQL with Jimi Hendrix - http://tinyurl.com/2zpd9n

When I show developers Freebase and the query language MQL, the first
thing they often try to do is extract all the info they can out of the
database for a generic topic.

This is actually quite a good learning experience, as you have to
learn lots of MQL concepts:

    * How "*" expands (default values v objects, type specified or not)
    * Multiple queries v nested queries
    * Reflection

I had been keeping my own notes, but I thought they might be worth
sharing even although they are very incomplete.

I'd be interested in your feedback.
Is this something worth working on?

regards,
--Will

From tim at metaweb.com  Thu Jul 12 18:48:14 2007
From: tim at metaweb.com (Tim Kientzle)
Date: Thu, 12 Jul 2007 11:48:14 -0700
Subject: [Developers] Sandbox.freebase.com outage today.
Message-ID: <A0C39C26-9A53-4F6D-AA76-CC5120D024B0@metaweb.com>

Between 4pm and 6pm Pacific Time (2300-0100 UTC), we'll be doing some  
maintenance work on sandbox.freebase.com.

The system will be read-only for some portion of that time and will  
be completely down for a couple of minutes.

Thanks for your patience,

Tim Kientzle


From tim at metaweb.com  Fri Jul 13 23:37:44 2007
From: tim at metaweb.com (Tim Kientzle)
Date: Fri, 13 Jul 2007 16:37:44 -0700
Subject: [Developers] Sandbox.freebase.com
Message-ID: <E95AA145-054F-4C0C-AE2B-977C817860B9@metaweb.com>

We apologize for the recent problems with sandbox.freebase.com.

Our recent maintenance uncovered some issues which should
now be resolved.

Tim Kientzle


From jeff at metaweb.com  Mon Jul 16 16:40:38 2007
From: jeff at metaweb.com (Jeff Prucher)
Date: Mon, 16 Jul 2007 09:40:38 -0700
Subject: [Developers] Moving types out of the /common domain
In-Reply-To: <001e01c7bf59$d4fa8190$bc01a8c0@p4>
Message-ID: <000c01c7c7c8$1c4bba20$bc01a8c0@p4>

We finally migrated the types out of the /common domain, but differently
(and later!) than in the original plan.  The final resting places for the
types formerly in /common are as follows:

Moved to a new domain, /award:
award 
award_winning_work 
award-winner 
award_presenting_organization 
award_nominee 
award_nominated_work 
award_nomination 
award_honor 

Moved to /media_common:
citation

Moved to /time
weekday

Deleted outright:
collection
review

Remaining in /common:
Topic
Document
Image
interpersonal relationship*
Required attribution/authorship info for licenses
Licensed Object
License
Webpage
Legal Entity
Phone Number

*Note that "interpersonal relationship" is still slated for deletion.

Jeff Prucher
Type Librarian & Ontologist
Metaweb, Inc.

________________________________

	From: developers-bounces at freebase.com
[mailto:developers-bounces at freebase.com] On Behalf Of Jeff Prucher
	Sent: Thursday, July 05, 2007 4:11 PM
	To: developers at freebase.com
	Subject: [Developers] Moving types out of the /common domain
	
	

	The /common domain currently contains types that serve two different
purposes, which is starting to become a problem. The two kinds of types are
those, such as /common/topic, that are important to the workings of Freebase
itself, and those, such as /common/phone_number, that are used by a variety
of types in several domains but which are not necessary to the functioning
of Freebase.  We would like to create a new domain (/shared) for the second
kind of types (the ones that are merely used across domains) to make it
easier to manage the types that are crucial to the system.

	Moving types around like this is something that we are still doing
in Alpha (though as little as possible), and we are working on ways to make
addresses stable.  We will make the following changes on Monday, July 9, if
we don't hear any objections.

	The types currently in /common that will be moved to /shared are: 
	award 
	award_winning_work 
	award-winner 
	award_presenting_organization 
	award_nominee 
	award_nominated_work 
	award_nomination 
	award_honor 
	citation 
	collection 
	legal_entity 
	phone_number 
	review 
	weekday 

	The types that will remain in /common are: 
	topic 
	document 
	image 
	attribution_info 
	licensed_object 
	license 
	webpage 

	Additionally, /common/interpersonal_relationship will be deleted
entirely. 

	Jeff Prucher 
	Type Librarian & Ontologist 
	Metaweb, Inc. 



From niallo at metaweb.com  Tue Jul 17 00:15:49 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 16 Jul 2007 17:15:49 -0700
Subject: [Developers] sandbox.freebase.com weekly update
Message-ID: <20070717001549.GG659@digdug.inet>

Hi,

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


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

From niallo at metaweb.com  Tue Jul 17 00:58:08 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Mon, 16 Jul 2007 17:58:08 -0700
Subject: [Developers] sandbox.freebase.com weekly update
In-Reply-To: <20070717001549.GG659@digdug.inet>
References: <20070717001549.GG659@digdug.inet>
Message-ID: <20070717005808.GJ659@digdug.inet>

Update complete.

On Mon, Jul 16, 2007 at 05:15:49PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Data refresh and software update for 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 niallo at metaweb.com  Tue Jul 17 23:20:03 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 17 Jul 2007 16:20:03 -0700
Subject: [Developers] www.freebase.com weekly update
Message-ID: <20070717232003.GQ659@digdug.inet>

Hi,

Weekly software update of www.freebase.com happening over the next few
minutes.

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

From niallo at metaweb.com  Tue Jul 17 23:54:05 2007
From: niallo at metaweb.com (Niall O'Higgins)
Date: Tue, 17 Jul 2007 16:54:05 -0700
Subject: [Developers] www.freebase.com weekly update
In-Reply-To: <20070717232003.GQ659@digdug.inet>
References: <20070717232003.GQ659@digdug.inet>
Message-ID: <20070717235405.GS659@digdug.inet>

Update complete.

On Tue, Jul 17, 2007 at 04:20:03PM -0700, Niall O'Higgins wrote:
> Hi,
> 
> Weekly software update of www.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 acallejas at appliedminds.com  Wed Jul 18 02:59:43 2007
From: acallejas at appliedminds.com (Augusto Callejas)
Date: Tue, 17 Jul 2007 19:59:43 -0700
Subject: [Developers] 888 types.
Message-ID: <C2C2D02F.E78%acallejas@appliedminds.com>

hi-

when i log into freebase, the current total number of types is 888.
with respect to the category of types:
  
http://freebase.com/view/helptopic?id=%239202a8c04000641f800000000544e143#id
2610292

does this count include both freebase types and user-defined types?
if so, is the only way to query for these types is to to exclude
types in the "/common" and "/type" domains?

thanks,
augusto.



From jg at metaweb.com  Wed Jul 18 06:46:54 2007
From: jg at metaweb.com (John Giannandrea)
Date: Tue, 17 Jul 2007 23:46:54 -0700
Subject: [Developers] 888 types.
In-Reply-To: <C2C2D02F.E78%acallejas@appliedminds.com>
References: <C2C2D02F.E78%acallejas@appliedminds.com>
Message-ID: <20070718064654.GA12227@metaweb.com>


Augusto Callejas wrote:
> when i log into freebase, the current total number of types is 888.
> with respect to the category of types:
> does this count include both freebase types and user-defined types?

There are 888 'published' types which means types that are
co-typed with a /freebase/type_profile.

There are actually around 2396 types currently in the system.

> if so, is the only way to query for these types is to to exclude
> types in the "/common" and "/type" domains?

The main distinctions between user's private types and
public types is that 1. the public types are in a namespace
which itself is in namespace '/' and 2. public types are 
marked as such with /freebase/type_profile/published = 'Published'

Here is a small MJT app that displays all the public and
then private types in the system:

	http://64.13.140.106/mw/mjt/count_types.html

-jg


From acallejas at appliedminds.com  Thu Jul 19 17:05:37 2007
From: acallejas at appliedminds.com (Augusto Callejas)
Date: Thu, 19 Jul 2007 10:05:37 -0700
Subject: [Developers] mediated properties
In-Reply-To: <835ED216-C2BA-4A4D-827D-BBA1CD7D1DB2@metaweb.com>
Message-ID: <C2C4E7F1.E83%acallejas@appliedminds.com>

hi-

regarding mediating types, how do i determine what two other types
it is mediating for?  for instance, there is the complex type:

  http://freebase.com/view/filter?id=/music/group_membership

reading the type description reveals that it relates these two types:

  /music/group_member (via property /music/group_membership/member)
  /music/musical_group (via property /music/group_membership/group)

how do i determine this programmatically?  at first, i was going to
use that fact that these two properties have "unique":true, but
so is the case for two other properties:

  /music/group_membership/start
  /music/group_membership/end

how can i make the distinction?

thanks,
augusto.


On 6/28/07 3:46 PM, "Dae Park" <daepark at metaweb.com> wrote:

> We perform a "schema query" for each type a topic is an instance
> which includes a query for each property and their expected_type. If
> the expected_type is marked as a mediator then we go another ply out
> and ask for the property's expected_type's disambiguating properties.
> 
> {
>    id: #topic_id
>    type: [{
>      id: null,
>      properties: [{
>        id: null,
>        master_property: null,
>        reverse_property: null,
>        unique: null,
>        expected_type: {
>          id: null,
>          "/freebase/type_hints/mediator": null,
>          properties: [{
>            master_property: null,
>            reverse_property: null,
>            id: null,
>            unique: null,
>            "/freebase/type_hints/disambiguator": true
>          }}
>        }
>      }]
>    }]
> }
> 
> When displaying these mediator properties (sub_properties) from a
> topic's property:
> 
> if the topic_property and sub_property are reciprocal of one another:
> if sub_property is unique:
> skip
> else:
> foreach value in sub-property:
> if value == topic we're viewing:
> continue:
> display value
> 
> 
> Two properties (prop1, prop2) are reciprocal's of one another if:
> 
> prop1.reverse_property == prop2
> or
> prop2.master_property == prop1
> 
> 
> 
> 
> 
> 
> On Jun 28, 2007, at 1:15 PM, Steve Sak wrote:
> 
>> 
>> How does the freebase.com "person" web pages "know" to display the
>> name
>> of the spouse/sibling or the name of the organization a person worked
>> for under "employment history"? Hard-coded?
>> 
>> Nick Thompson wrote:
>>> in general, i don't think your question has a well-defined answer -
>>> there's no reason for the system to believe that
>>> /business/employment_tenure/company is more relevant than
>>> /business/employment_tenure/title, for example.  you could be looking
>>> for a list of people at a particular company, in which case the job
>>> title would more interesting.
>>> 
>>> so, you might rule out /business/employment_tenure/person from
>>> context
>>> as chris described, but that doesn't determine a unique "other side".
>>> that's the nature of compound value types - if a relationship only
>>> had two sides, it would probably be modeled as a single link instead
>>> of being mediated by a compound value object.
>>> 
>>>     nick
>>> 
>>> Steve Sak wrote:
>>>> So, if I'm looking at /people/person/employment_history how would I
>>>> determine that the relevant property is
>>>> /business/employment_tenure/company?
>>>> 
>>>> 
>>>> Alec Flett wrote:
>>>>> I want to add that checking by is only valid if you know you're
>>>>> coming
>>>>> in and out on the same property (as is the case with marriage) - if
>>>>> the properties are different, you need to look at master/reverse
>>>>> properties because you can imagine a director who also produced a
>>>>> movie - if you were just going by ids, looking at the
>>>>> relationship of
>>>>> 
>>>>> "Who is the /director/ of this movie /produced by/ Martin
>>>>> Scorsese?" -
>>>>> you might get Martin Scorsese back, but that doesn't mean that
>>>>> "director" is the reverse of "produced by".
>>>>> 
>>>>> Alec
>>>>> 
>>>>> Christopher R. Maden wrote:
>>>>>> Steve Sak wrote:
>>>>>> 
>>>>>>> To restate the question: given a mediator object how do I
>>>>>>> determine
>>>>>>> which property is the one it's mediating?  For example, type A
>>>>>>> has a
>>>>>>> mediator M that has a bunch of properties mediates between A and
>>>>>>> another
>>>>>>> object.  How do I determine which property of M is the one that
>>>>>>> points
>>>>>>> to the other object being mediated?
>>>>>>> 
>>>>>> Mediators, or compound value types, are not especially different
>>>>>> from any other type; the UI treats them a little differently, but
>>>>>> that?s it.
>>>>>> 
>>>>>> But I think you are asking about instances rather than types,
>>>>>> correct? In other words, Arnold Schwarzenegger connects to a
>>>>>> marriage, and that marriage connects to two people, both Arnold
>>>>>> Schwarzenegger and Maria Shriver.  How can you tell which of the
>>>>>> spouses is the one you started from?
>>>>>> 
>>>>>> The only way to do that, really, is by looking at the IDs.
>>>>>> 
>>>>>> {
>>>>>>    "query": {
>>>>>>      "id": null,
>>>>>>      "name": "Arnold Schwarzenegger",
>>>>>>      "spouse_s": [
>>>>>>        {
>>>>>>          "spouse": [
>>>>>>            {
>>>>>>              "id":null,
>>>>>>              "name":null
>>>>>>            }
>>>>>>          ],
>>>>>>          "type":"/people/marriage"
>>>>>>        }
>>>>>>      ],
>>>>>>      "type":"/people/person"
>>>>>>    }
>>>>>> }
>>>>>> 
>>>>>> {
>>>>>>    "result": {
>>>>>>      "spouse_s": [
>>>>>>        {
>>>>>>          "spouse": [
>>>>>>            {
>>>>>>              "id": "#9202a8c04000641f80000000001da07e",
>>>>>>              "name": "Maria Shriver"
>>>>>>            },
>>>>>>            {
>>>>>>              "id": "#9202a8c04000641f8000000000006567",
>>>>>>              "name": "Arnold Schwarzenegger"
>>>>>>            }
>>>>>>          ],
>>>>>>          "type": "/people/marriage"
>>>>>>        }
>>>>>>      ],
>>>>>>      "type": "/people/person",
>>>>>>      "id": "#9202a8c04000641f8000000000006567",
>>>>>>      "name": "Arnold Schwarzenegger"
>>>>>>    }
>>>>>> }
>>>>>> 
>>>>>> You can see that one of the spouses in the marriage has the
>>>>>> same ID
>>>>>> as the person with whom you started, and that the other is
>>>>>> different.
>>>>>> 
>>>>>> Unfortunately, there?s not really an automatic way to say ?get all
>>>>>> of the values of these properties except the ones that point back
>>>>>> to the parent within the scope of this query.?
>>>>>> 
>>>>>> HTH,
>>>>>> Chris
>>>>>> 
>>>> 
>> 
>> -- 
>> Steven G Sak
>> Applied Minds, Inc.
>> 11718 Bowman Green Drive
>> Reston, Virginia 20190
>> o: (703) 483-2207
>> c: (703) 626-2557
>> 
>> _______________________________________________
>> 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 darin at metaweb.com  Thu Jul 19 17:52:22 2007
From: darin at metaweb.com (Darin Wilson)
Date: Thu, 19 Jul 2007 10:52:22 -0700
Subject: [Developers] mediated properties
In-Reply-To: <C2C4E7F1.E83%acallejas@appliedminds.com>
References: <C2C4E7F1.E83%acallejas@appliedminds.com>
Message-ID: <9E5AD426-0D3E-4748-A010-144118BC332F@metaweb.com>

Hi Augusto-

A mediating type, or complex value type (CVT), actually mediates for  
all of the properties that the type contains; however, the type  
designer can choose which of those properties should be displayed  
when an instance of the type linked from another topic. If you look  
at the type definition of /music/group_membership, you'll see that  
each of the properties has the "Display as disambiguator" box  
checked. This means that the property will be displayed, if the data  
is available.

If you look at the instance list for /music/group_membership, you'll  
see that some of the instances have years and other don't. That's  
because the "Period (start)" and/or "Period (end)" properties haven't  
been filled out yet. If the data were available, it would be displayed.

http://www.freebase.com/view/filter?id=%2Fmusic%2Fgroup_membership

To see this programmatically, you can use the following query. The / 
freebase/type_hints/mediator property on the type will tell you if  
it's a CVT or not. The /freebase/property_hints/disambiguator on the  
properties will tell you if the value of the property should be  
displayed.

{
   "query":[{
     "id":"/music/group_membership",
     "type":"/type/type",
     "/freebase/type_hints/mediator":null,
     "properties":[{
       "id":null,
       "/freebase/property_hints/disambiguator":null
     }]
   }]
}

HTH,
Darin


On Jul 19, 2007, at 10:05 AM, Augusto Callejas wrote:

> hi-
>
> regarding mediating types, how do i determine what two other types
> it is mediating for?  for instance, there is the complex type:
>
>   http://freebase.com/view/filter?id=/music/group_membership
>
> reading the type description reveals that it relates these two types:
>
>   /music/group_member (via property /music/group_membership/member)
>   /music/musical_group (via property /music/group_membership/group)
>
> how do i determine this programmatically?  at first, i was going to
> use that fact that these two properties have "unique":true, but
> so is the case for two other properties:
>
>   /music/group_membership/start
>   /music/group_membership/end
>
> how can i make the distinction?
>
> thanks,
> augusto.
>
>
> On 6/28/07 3:46 PM, "Dae Park" <daepark at metaweb.com> wrote:
>
>> We perform a "schema query" for each type a topic is an instance
>> which includes a query for each property and their expected_type. If
>> the expected_type is marked as a mediator then we go another ply out
>> and ask for the property's expected_type's disambiguating properties.
>>
>> {
>>    id: #topic_id
>>    type: [{
>>      id: null,
>>      properties: [{
>>        id: null,
>>        master_property: null,
>>        reverse_property: null,
>>        unique: null,
>>        expected_type: {
>>          id: null,
>>          "/freebase/type_hints/mediator": null,
>>          properties: [{
>>            master_property: null,
>>            reverse_property: null,
>>            id: null,
>>            unique: null,
>>            "/freebase/type_hints/disambiguator": true
>>          }}
>>        }
>>      }]
>>    }]
>> }
>>
>> When displaying these mediator properties (sub_properties) from a
>> topic's property:
>>
>> if the topic_property and sub_property are reciprocal of one another:
>> if sub_property is unique:
>> skip
>> else:
>> foreach value in sub-property:
>> if value == topic we're viewing:
>> continue:
>> display value
>>
>>
>> Two properties (prop1, prop2) are reciprocal's of one another if:
>>
>> prop1.reverse_property == prop2
>> or
>> prop2.master_property == prop1
>>
>>
>>
>>
>>
>>
>> On Jun 28, 2007, at 1:15 PM, Steve Sak wrote:
>>
>>>
>>> How does the freebase.com "person" web pages "know" to display the
>>> name
>>> of the spouse/sibling or the name of the organization a person  
>>> worked
>>> for under "employment history"? Hard-coded?
>>>
>>> Nick Thompson wrote:
>>>> in general, i don't think your question has a well-defined answer -
>>>> there's no reason for the system to believe that
>>>> /business/employment_tenure/company is more relevant than
>>>> /business/employment_tenure/title, for example.  you could be  
>>>> looking
>>>> for a list of people at a particular company, in which case the job
>>>> title would more interesting.
>>>>
>>>> so, you might rule out /business/employment_tenure/person from
>>>> context
>>>> as chris described, but that doesn't determine a unique "other  
>>>> side".
>>>> that's the nature of compound value types - if a relationship only
>>>> had two sides, it would probably be modeled as a single link  
>>>> instead
>>>> of being mediated by a compound value object.
>>>>
>>>>     nick
>>>>
>>>> Steve Sak wrote:
>>>>> So, if I'm looking at /people/person/employment_history how  
>>>>> would I
>>>>> determine that the relevant property is
>>>>> /business/employment_tenure/company?
>>>>>
>>>>>
>>>>> Alec Flett wrote:
>>>>>> I want to add that checking by is only valid if you know you're
>>>>>> coming
>>>>>> in and out on the same property (as is the case with marriage)  
>>>>>> - if
>>>>>> the properties are different, you need to look at master/reverse
>>>>>> properties because you can imagine a director who also produced a
>>>>>> movie - if you were just going by ids, looking at the
>>>>>> relationship of
>>>>>>
>>>>>> "Who is the /director/ of this movie /produced by/ Martin
>>>>>> Scorsese?" -
>>>>>> you might get Martin Scorsese back, but that doesn't mean that
>>>>>> "director" is the reverse of "produced by".
>>>>>>
>>>>>> Alec
>>>>>>
>>>>>> Christopher R. Maden wrote:
>>>>>>> Steve Sak wrote:
>>>>>>>
>>>>>>>> To restate the question: given a mediator object how do I
>>>>>>>> determine
>>>>>>>> which property is the one it's mediating?  For example, type A
>>>>>>>> has a
>>>>>>>> mediator M that has a bunch of properties mediates between A  
>>>>>>>> and
>>>>>>>> another
>>>>>>>> object.  How do I determine which property of M is the one that
>>>>>>>> points
>>>>>>>> to the other object being mediated?
>>>>>>>>
>>>>>>> Mediators, or compound value types, are not especially different
>>>>>>> from any other type; the UI treats them a little differently,  
>>>>>>> but
>>>>>>> that?s it.
>>>>>>>
>>>>>>> But I think you are asking about instances rather than types,
>>>>>>> correct? In other words, Arnold Schwarzenegger connects to a
>>>>>>> marriage, and that marriage connects to two people, both Arnold
>>>>>>> Schwarzenegger and Maria Shriver.  How can you tell which of the
>>>>>>> spouses is the one you started from?
>>>>>>>
>>>>>>> The only way to do that, really, is by looking at the IDs.
>>>>>>>
>>>>>>> {
>>>>>>>    "query": {
>>>>>>>      "id": null,
>>>>>>>      "name": "Arnold Schwarzenegger",
>>>>>>>      "spouse_s": [
>>>>>>>        {
>>>>>>>          "spouse": [
>>>>>>>            {
>>>>>>>              "id":null,
>>>>>>>              "name":null
>>>>>>>            }
>>>>>>>          ],
>>>>>>>          "type":"/people/marriage"
>>>>>>>        }
>>>>>>>      ],
>>>>>>>      "type":"/people/person"
>>>>>>>    }
>>>>>>> }
>>>>>>>
>>>>>>> {
>>>>>>>    "result": {
>>>>>>>      "spouse_s": [
>>>>>>>        {
>>>>>>>          "spouse": [
>>>>>>>            {
>>>>>>>              "id": "#9202a8c04000641f80000000001da07e",
>>>>>>>              "name": "Maria Shriver"
>>>>>>>            },
>>>>>>>            {
>>>>>>>              "id": "#9202a8c04000641f8000000000006567",
>>>>>>>              "name": "Arnold Schwarzenegger"
>>>>>>>            }
>>>>>>>          ],
>>>>>>>          "type": "/people/marriage"
>>>>>>>        }
>>>>>>>      ],
>>>>>>>      "type": "/people/person",
>>>>>>>      "id": "#9202a8c04000641f8000000000006567",
>>>>>>>      "name": "Arnold Schwarzenegger"
>>>>>>>    }
>>>>>>> }
>>>>>>>
>>>>>>> You can see that one of the spouses in the marriage has the
>>>>>>> same ID
>>>>>>> as the person with whom you started, and that the other is
>>>>>>> different.
>>>>>>>
>>>>>>> Unfortunately, there?s not really an automatic way to say  
>>>>>>> ?get all
>>>>>>> of the values of these properties except the ones that point  
>>>>>>> back
>>>>>>> to the parent within the scope of this query.?
>>>>>>>
>>>>>>> HTH,
>>>>>>> Chris
>>>>>>>
>>>>>
>>>
>>> -- 
>>> Steven G Sak
>>> Applied Minds, Inc.
>>> 11718 Bowman Green Drive
>>> Reston, Virginia 20190
>>> o: (703) 483-2207
>>> c: (703) 626-2557
>>>
>>> _______________________________________________
>>> 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
>
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From acallejas at appliedminds.com  Thu Jul 19 21:37:54 2007
From: acallejas at appliedminds.com (Augusto Callejas)
Date: Thu, 19 Jul 2007 14:37:54 -0700
Subject: [Developers] mediated properties
In-Reply-To: <9E5AD426-0D3E-4748-A010-144118BC332F@metaweb.com>
Message-ID: <C2C527C2.E94%acallejas@appliedminds.com>

hi-

i ran that query but all properties have value "true"
for the "disambiguator" property hint.

from the type documentation:

  Musical Group Membership is a compound value type which defines the
  relationship between a Musical Artist (a band or collaboration) and its
  Musical Group Members (musicians).

is there a difference when you say it mediates for all of the properties
versus when the documentation says it defines relationships between
musical artist and group member?  i'm looking for how that relationship
is captured in the type information.  since the disambigurator property
hint is the same for all properties, it doesn't define the relationship
i'm looking for.

thanks,
augusto.


On 7/19/07 10:52 AM, "Darin Wilson" <darin at metaweb.com> wrote:

> Hi Augusto-
> 
> A mediating type, or complex value type (CVT), actually mediates for
> all of the properties that the type contains; however, the type
> designer can choose which of those properties should be displayed
> when an instance of the type linked from another topic. If you look
> at the type definition of /music/group_membership, you'll see that
> each of the properties has the "Display as disambiguator" box
> checked. This means that the property will be displayed, if the data
> is available.
> 
> If you look at the instance list for /music/group_membership, you'll
> see that some of the instances have years and other don't. That's
> because the "Period (start)" and/or "Period (end)" properties haven't
> been filled out yet. If the data were available, it would be displayed.
> 
> http://www.freebase.com/view/filter?id=%2Fmusic%2Fgroup_membership
> 
> To see this programmatically, you can use the following query. The /
> freebase/type_hints/mediator property on the type will tell you if
> it's a CVT or not. The /freebase/property_hints/disambiguator on the
> properties will tell you if the value of the property should be
> displayed.
> 
> {
>    "query":[{
>      "id":"/music/group_membership",
>      "type":"/type/type",
>      "/freebase/type_hints/mediator":null,
>      "properties":[{
>        "id":null,
>        "/freebase/property_hints/disambiguator":null
>      }]
>    }]
> }
> 
> HTH,
> Darin
> 
> 
> On Jul 19, 2007, at 10:05 AM, Augusto Callejas wrote:
> 
>> hi-
>> 
>> regarding mediating types, how do i determine what two other types
>> it is mediating for?  for instance, there is the complex type:
>> 
>>   http://freebase.com/view/filter?id=/music/group_membership
>> 
>> reading the type description reveals that it relates these two types:
>> 
>>   /music/group_member (via property /music/group_membership/member)
>>   /music/musical_group (via property /music/group_membership/group)
>> 
>> how do i determine this programmatically?  at first, i was going to
>> use that fact that these two properties have "unique":true, but
>> so is the case for two other properties:
>> 
>>   /music/group_membership/start
>>   /music/group_membership/end
>> 
>> how can i make the distinction?
>> 
>> thanks,
>> augusto.
>> 
>> 
>> On 6/28/07 3:46 PM, "Dae Park" <daepark at metaweb.com> wrote:
>> 
>>> We perform a "schema query" for each type a topic is an instance
>>> which includes a query for each property and their expected_type. If
>>> the expected_type is marked as a mediator then we go another ply out
>>> and ask for the property's expected_type's disambiguating properties.
>>> 
>>> {
>>>    id: #topic_id
>>>    type: [{
>>>      id: null,
>>>      properties: [{
>>>        id: null,
>>>        master_property: null,
>>>        reverse_property: null,
>>>        unique: null,
>>>        expected_type: {
>>>          id: null,
>>>          "/freebase/type_hints/mediator": null,
>>>          properties: [{
>>>            master_property: null,
>>>            reverse_property: null,
>>>            id: null,
>>>            unique: null,
>>>            "/freebase/type_hints/disambiguator": true
>>>          }}
>>>        }
>>>      }]
>>>    }]
>>> }
>>> 
>>> When displaying these mediator properties (sub_properties) from a
>>> topic's property:
>>> 
>>> if the topic_property and sub_property are reciprocal of one another:
>>> if sub_property is unique:
>>> skip
>>> else:
>>> foreach value in sub-property:
>>> if value == topic we're viewing:
>>> continue:
>>> display value
>>> 
>>> 
>>> Two properties (prop1, prop2) are reciprocal's of one another if:
>>> 
>>> prop1.reverse_property == prop2
>>> or
>>> prop2.master_property == prop1
>>> 
>>> 
>>> 
>>> 
>>> 
>>> 
>>> On Jun 28, 2007, at 1:15 PM, Steve Sak wrote:
>>> 
>>>> 
>>>> How does the freebase.com "person" web pages "know" to display the
>>>> name
>>>> of the spouse/sibling or the name of the organization a person
>>>> worked
>>>> for under "employment history"? Hard-coded?
>>>> 
>>>> Nick Thompson wrote:
>>>>> in general, i don't think your question has a well-defined answer -
>>>>> there's no reason for the system to believe that
>>>>> /business/employment_tenure/company is more relevant than
>>>>> /business/employment_tenure/title, for example.  you could be
>>>>> looking
>>>>> for a list of people at a particular company, in which case the job
>>>>> title would more interesting.
>>>>> 
>>>>> so, you might rule out /business/employment_tenure/person from
>>>>> context
>>>>> as chris described, but that doesn't determine a unique "other
>>>>> side".
>>>>> that's the nature of compound value types - if a relationship only
>>>>> had two sides, it would probably be modeled as a single link
>>>>> instead
>>>>> of being mediated by a compound value object.
>>>>> 
>>>>>     nick
>>>>> 
>>>>> Steve Sak wrote:
>>>>>> So, if I'm looking at /people/person/employment_history how
>>>>>> would I
>>>>>> determine that the relevant property is
>>>>>> /business/employment_tenure/company?
>>>>>> 
>>>>>> 
>>>>>> Alec Flett wrote:
>>>>>>> I want to add that checking by is only valid if you know you're
>>>>>>> coming
>>>>>>> in and out on the same property (as is the case with marriage)
>>>>>>> - if
>>>>>>> the properties are different, you need to look at master/reverse
>>>>>>> properties because you can imagine a director who also produced a
>>>>>>> movie - if you were just going by ids, looking at the
>>>>>>> relationship of
>>>>>>> 
>>>>>>> "Who is the /director/ of this movie /produced by/ Martin
>>>>>>> Scorsese?" -
>>>>>>> you might get Martin Scorsese back, but that doesn't mean that
>>>>>>> "director" is the reverse of "produced by".
>>>>>>> 
>>>>>>> Alec
>>>>>>> 
>>>>>>> Christopher R. Maden wrote:
>>>>>>>> Steve Sak wrote:
>>>>>>>> 
>>>>>>>>> To restate the question: given a mediator object how do I
>>>>>>>>> determine
>>>>>>>>> which property is the one it's mediating?  For example, type A
>>>>>>>>> has a
>>>>>>>>> mediator M that has a bunch of properties mediates between A
>>>>>>>>> and
>>>>>>>>> another
>>>>>>>>> object.  How do I determine which property of M is the one that
>>>>>>>>> points
>>>>>>>>> to the other object being mediated?
>>>>>>>>> 
>>>>>>>> Mediators, or compound value types, are not especially different
>>>>>>>> from any other type; the UI treats them a little differently,
>>>>>>>> but
>>>>>>>> that?s it.
>>>>>>>> 
>>>>>>>> But I think you are asking about instances rather than types,
>>>>>>>> correct? In other words, Arnold Schwarzenegger connects to a
>>>>>>>> marriage, and that marriage connects to two people, both Arnold
>>>>>>>> Schwarzenegger and Maria Shriver.  How can you tell which of the
>>>>>>>> spouses is the one you started from?
>>>>>>>> 
>>>>>>>> The only way to do that, really, is by looking at the IDs.
>>>>>>>> 
>>>>>>>> {
>>>>>>>>    "query": {
>>>>>>>>      "id": null,
>>>>>>>>      "name": "Arnold Schwarzenegger",
>>>>>>>>      "spouse_s": [
>>>>>>>>        {
>>>>>>>>          "spouse": [
>>>>>>>>            {
>>>>>>>>              "id":null,
>>>>>>>>              "name":null
>>>>>>>>            }
>>>>>>>>          ],
>>>>>>>>          "type":"/people/marriage"
>>>>>>>>        }
>>>>>>>>      ],
>>>>>>>>      "type":"/people/person"
>>>>>>>>    }
>>>>>>>> }
>>>>>>>> 
>>>>>>>> {
>>>>>>>>    "result": {
>>>>>>>>      "spouse_s": [
>>>>>>>>        {
>>>>>>>>          "spouse": [
>>>>>>>>            {
>>>>>>>>              "id": "#9202a8c04000641f80000000001da07e",
>>>>>>>>              "name": "Maria Shriver"
>>>>>>>>            },
>>>>>>>>            {
>>>>>>>>              "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>              "name": "Arnold Schwarzenegger"
>>>>>>>>            }
>>>>>>>>          ],
>>>>>>>>          "type": "/people/marriage"
>>>>>>>>        }
>>>>>>>>      ],
>>>>>>>>      "type": "/people/person",
>>>>>>>>      "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>      "name": "Arnold Schwarzenegger"
>>>>>>>>    }
>>>>>>>> }
>>>>>>>> 
>>>>>>>> You can see that one of the spouses in the marriage has the
>>>>>>>> same ID
>>>>>>>> as the person with whom you started, and that the other is
>>>>>>>> different.
>>>>>>>> 
>>>>>>>> Unfortunately, there?s not really an automatic way to say
>>>>>>>> ?get all
>>>>>>>> of the values of these properties except the ones that point
>>>>>>>> back
>>>>>>>> to the parent within the scope of this query.?
>>>>>>>> 
>>>>>>>> HTH,
>>>>>>>> Chris
>>>>>>>> 
>>>>>> 
>>>> 
>>>> -- 
>>>> Steven G Sak
>>>> Applied Minds, Inc.
>>>> 11718 Bowman Green Drive
>>>> Reston, Virginia 20190
>>>> o: (703) 483-2207
>>>> c: (703) 626-2557
>>>> 
>>>> _______________________________________________
>>>> 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
>> 
>> 
>> _______________________________________________
>> 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 robert at metaweb.com  Thu Jul 19 21:59:49 2007
From: robert at metaweb.com (Robert Cook)
Date: Thu, 19 Jul 2007 14:59:49 -0700
Subject: [Developers] mediated properties
In-Reply-To: <C2C527C2.E94%acallejas@appliedminds.com>
References: <C2C527C2.E94%acallejas@appliedminds.com>
Message-ID: <4DE8CA24-F8A1-4D09-BC6F-C48E759CD719@metaweb.com>

Augusto --

You are asking a very good question.  A "mediator" is a data pattern  
that is represented as a Compound Value Type in Metaweb.  In other  
words, mediators are CVTs that act like an "annotated relationship"  
between two types.  One example is "Marriage", which joins two people  
together that also stores add start and end dates as metadata.

Currently, we don't have a way to annotate which of the two  
properties represent the "relationship" between two types.  That is,  
a mediator is more of a data idiom that makes sense to the modeler,  
but has no formal representation in the data store.

We've received feedback similar to yours in the past, and we're  
currently working on a way of making that distinction (while  
simultaneously solving a few other problems.)  We don't yet have an  
ETA, but know that it's a real priority for us.

Robert

On Jul 19, 2007, at 2:37 PM, Augusto Callejas wrote:

> hi-
>
> i ran that query but all properties have value "true"
> for the "disambiguator" property hint.
>
> from the type documentation:
>
>   Musical Group Membership is a compound value type which defines the
>   relationship between a Musical Artist (a band or collaboration)  
> and its
>   Musical Group Members (musicians).
>
> is there a difference when you say it mediates for all of the  
> properties
> versus when the documentation says it defines relationships between
> musical artist and group member?  i'm looking for how that  
> relationship
> is captured in the type information.  since the disambigurator  
> property
> hint is the same for all properties, it doesn't define the  
> relationship
> i'm looking for.
>
> thanks,
> augusto.
>
>
> On 7/19/07 10:52 AM, "Darin Wilson" <darin at metaweb.com> wrote:
>
>> Hi Augusto-
>>
>> A mediating type, or complex value type (CVT), actually mediates for
>> all of the properties that the type contains; however, the type
>> designer can choose which of those properties should be displayed
>> when an instance of the type linked from another topic. If you look
>> at the type definition of /music/group_membership, you'll see that
>> each of the properties has the "Display as disambiguator" box
>> checked. This means that the property will be displayed, if the data
>> is available.
>>
>> If you look at the instance list for /music/group_membership, you'll
>> see that some of the instances have years and other don't. That's
>> because the "Period (start)" and/or "Period (end)" properties haven't
>> been filled out yet. If the data were available, it would be  
>> displayed.
>>
>> http://www.freebase.com/view/filter?id=%2Fmusic%2Fgroup_membership
>>
>> To see this programmatically, you can use the following query. The /
>> freebase/type_hints/mediator property on the type will tell you if
>> it's a CVT or not. The /freebase/property_hints/disambiguator on the
>> properties will tell you if the value of the property should be
>> displayed.
>>
>> {
>>    "query":[{
>>      "id":"/music/group_membership",
>>      "type":"/type/type",
>>      "/freebase/type_hints/mediator":null,
>>      "properties":[{
>>        "id":null,
>>        "/freebase/property_hints/disambiguator":null
>>      }]
>>    }]
>> }
>>
>> HTH,
>> Darin
>>
>>
>> On Jul 19, 2007, at 10:05 AM, Augusto Callejas wrote:
>>
>>> hi-
>>>
>>> regarding mediating types, how do i determine what two other types
>>> it is mediating for?  for instance, there is the complex type:
>>>
>>>   http://freebase.com/view/filter?id=/music/group_membership
>>>
>>> reading the type description reveals that it relates these two  
>>> types:
>>>
>>>   /music/group_member (via property /music/group_membership/member)
>>>   /music/musical_group (via property /music/group_membership/group)
>>>
>>> how do i determine this programmatically?  at first, i was going to
>>> use that fact that these two properties have "unique":true, but
>>> so is the case for two other properties:
>>>
>>>   /music/group_membership/start
>>>   /music/group_membership/end
>>>
>>> how can i make the distinction?
>>>
>>> thanks,
>>> augusto.
>>>
>>>
>>> On 6/28/07 3:46 PM, "Dae Park" <daepark at metaweb.com> wrote:
>>>
>>>> We perform a "schema query" for each type a topic is an instance
>>>> which includes a query for each property and their  
>>>> expected_type. If
>>>> the expected_type is marked as a mediator then we go another ply  
>>>> out
>>>> and ask for the property's expected_type's disambiguating  
>>>> properties.
>>>>
>>>> {
>>>>    id: #topic_id
>>>>    type: [{
>>>>      id: null,
>>>>      properties: [{
>>>>        id: null,
>>>>        master_property: null,
>>>>        reverse_property: null,
>>>>        unique: null,
>>>>        expected_type: {
>>>>          id: null,
>>>>          "/freebase/type_hints/mediator": null,
>>>>          properties: [{
>>>>            master_property: null,
>>>>            reverse_property: null,
>>>>            id: null,
>>>>            unique: null,
>>>>            "/freebase/type_hints/disambiguator": true
>>>>          }}
>>>>        }
>>>>      }]
>>>>    }]
>>>> }
>>>>
>>>> When displaying these mediator properties (sub_properties) from a
>>>> topic's property:
>>>>
>>>> if the topic_property and sub_property are reciprocal of one  
>>>> another:
>>>> if sub_property is unique:
>>>> skip
>>>> else:
>>>> foreach value in sub-property:
>>>> if value == topic we're viewing:
>>>> continue:
>>>> display value
>>>>
>>>>
>>>> Two properties (prop1, prop2) are reciprocal's of one another if:
>>>>
>>>> prop1.reverse_property == prop2
>>>> or
>>>> prop2.master_property == prop1
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Jun 28, 2007, at 1:15 PM, Steve Sak wrote:
>>>>
>>>>>
>>>>> How does the freebase.com "person" web pages "know" to display the
>>>>> name
>>>>> of the spouse/sibling or the name of the organization a person
>>>>> worked
>>>>> for under "employment history"? Hard-coded?
>>>>>
>>>>> Nick Thompson wrote:
>>>>>> in general, i don't think your question has a well-defined  
>>>>>> answer -
>>>>>> there's no reason for the system to believe that
>>>>>> /business/employment_tenure/company is more relevant than
>>>>>> /business/employment_tenure/title, for example.  you could be
>>>>>> looking
>>>>>> for a list of people at a particular company, in which case  
>>>>>> the job
>>>>>> title would more interesting.
>>>>>>
>>>>>> so, you might rule out /business/employment_tenure/person from
>>>>>> context
>>>>>> as chris described, but that doesn't determine a unique "other
>>>>>> side".
>>>>>> that's the nature of compound value types - if a relationship  
>>>>>> only
>>>>>> had two sides, it would probably be modeled as a single link
>>>>>> instead
>>>>>> of being mediated by a compound value object.
>>>>>>
>>>>>>     nick
>>>>>>
>>>>>> Steve Sak wrote:
>>>>>>> So, if I'm looking at /people/person/employment_history how
>>>>>>> would I
>>>>>>> determine that the relevant property is
>>>>>>> /business/employment_tenure/company?
>>>>>>>
>>>>>>>
>>>>>>> Alec Flett wrote:
>>>>>>>> I want to add that checking by is only valid if you know you're
>>>>>>>> coming
>>>>>>>> in and out on the same property (as is the case with marriage)
>>>>>>>> - if
>>>>>>>> the properties are different, you need to look at master/ 
>>>>>>>> reverse
>>>>>>>> properties because you can imagine a director who also  
>>>>>>>> produced a
>>>>>>>> movie - if you were just going by ids, looking at the
>>>>>>>> relationship of
>>>>>>>>
>>>>>>>> "Who is the /director/ of this movie /produced by/ Martin
>>>>>>>> Scorsese?" -
>>>>>>>> you might get Martin Scorsese back, but that doesn't mean that
>>>>>>>> "director" is the reverse of "produced by".
>>>>>>>>
>>>>>>>> Alec
>>>>>>>>
>>>>>>>> Christopher R. Maden wrote:
>>>>>>>>> Steve Sak wrote:
>>>>>>>>>
>>>>>>>>>> To restate the question: given a mediator object how do I
>>>>>>>>>> determine
>>>>>>>>>> which property is the one it's mediating?  For example,  
>>>>>>>>>> type A
>>>>>>>>>> has a
>>>>>>>>>> mediator M that has a bunch of properties mediates between A
>>>>>>>>>> and
>>>>>>>>>> another
>>>>>>>>>> object.  How do I determine which property of M is the one  
>>>>>>>>>> that
>>>>>>>>>> points
>>>>>>>>>> to the other object being mediated?
>>>>>>>>>>
>>>>>>>>> Mediators, or compound value types, are not especially  
>>>>>>>>> different
>>>>>>>>> from any other type; the UI treats them a little differently,
>>>>>>>>> but
>>>>>>>>> that?s it.
>>>>>>>>>
>>>>>>>>> But I think you are asking about instances rather than types,
>>>>>>>>> correct? In other words, Arnold Schwarzenegger connects to a
>>>>>>>>> marriage, and that marriage connects to two people, both  
>>>>>>>>> Arnold
>>>>>>>>> Schwarzenegger and Maria Shriver.  How can you tell which  
>>>>>>>>> of the
>>>>>>>>> spouses is the one you started from?
>>>>>>>>>
>>>>>>>>> The only way to do that, really, is by looking at the IDs.
>>>>>>>>>
>>>>>>>>> {
>>>>>>>>>    "query": {
>>>>>>>>>      "id": null,
>>>>>>>>>      "name": "Arnold Schwarzenegger",
>>>>>>>>>      "spouse_s": [
>>>>>>>>>        {
>>>>>>>>>          "spouse": [
>>>>>>>>>            {
>>>>>>>>>              "id":null,
>>>>>>>>>              "name":null
>>>>>>>>>            }
>>>>>>>>>          ],
>>>>>>>>>          "type":"/people/marriage"
>>>>>>>>>        }
>>>>>>>>>      ],
>>>>>>>>>      "type":"/people/person"
>>>>>>>>>    }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> {
>>>>>>>>>    "result": {
>>>>>>>>>      "spouse_s": [
>>>>>>>>>        {
>>>>>>>>>          "spouse": [
>>>>>>>>>            {
>>>>>>>>>              "id": "#9202a8c04000641f80000000001da07e",
>>>>>>>>>              "name": "Maria Shriver"
>>>>>>>>>            },
>>>>>>>>>            {
>>>>>>>>>              "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>>              "name": "Arnold Schwarzenegger"
>>>>>>>>>            }
>>>>>>>>>          ],
>>>>>>>>>          "type": "/people/marriage"
>>>>>>>>>        }
>>>>>>>>>      ],
>>>>>>>>>      "type": "/people/person",
>>>>>>>>>      "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>>      "name": "Arnold Schwarzenegger"
>>>>>>>>>    }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> You can see that one of the spouses in the marriage has the
>>>>>>>>> same ID
>>>>>>>>> as the person with whom you started, and that the other is
>>>>>>>>> different.
>>>>>>>>>
>>>>>>>>> Unfortunately, there?s not really an automatic way to say
>>>>>>>>> ?get all
>>>>>>>>> of the values of these properties except the ones that point
>>>>>>>>> back
>>>>>>>>> to the parent within the scope of this query.?
>>>>>>>>>
>>>>>>>>> HTH,
>>>>>>>>> Chris
>>>>>>>>>
>>>>>>>
>>>>>
>>>>> -- 
>>>>> Steven G Sak
>>>>> Applied Minds, Inc.
>>>>> 11718 Bowman Green Drive
>>>>> Reston, Virginia 20190
>>>>> o: (703) 483-2207
>>>>> c: (703) 626-2557
>>>>>
>>>>> _______________________________________________
>>>>> 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
>>>
>>>
>>> _______________________________________________
>>> 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
>
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers


From acallejas at appliedminds.com  Thu Jul 19 22:16:56 2007
From: acallejas at appliedminds.com (Augusto Callejas)
Date: Thu, 19 Jul 2007 15:16:56 -0700
Subject: [Developers] mediated properties
In-Reply-To: <4DE8CA24-F8A1-4D09-BC6F-C48E759CD719@metaweb.com>
Message-ID: <C2C530E8.E99%acallejas@appliedminds.com>

robert-

thanks that does clear it up. going by the definition of the meditator
pattern, does this type qualify as a mediator:

  http://www.freebase.com/view/schema?id=/measurement_unit/rect_size

i wouldn't think so since it only has two properties that are of value
types.

another way to state the question is: are all compound value types mediators
(e.g. opposed to all mediators being compund value types)?

thanks,
augusto.



On 7/19/07 2:59 PM, "Robert Cook" <robert at metaweb.com> wrote:

> Augusto --
> 
> You are asking a very good question.  A "mediator" is a data pattern
> that is represented as a Compound Value Type in Metaweb.  In other
> words, mediators are CVTs that act like an "annotated relationship"
> between two types.  One example is "Marriage", which joins two people
> together that also stores add start and end dates as metadata.
> 
> Currently, we don't have a way to annotate which of the two
> properties represent the "relationship" between two types.  That is,
> a mediator is more of a data idiom that makes sense to the modeler,
> but has no formal representation in the data store.
> 
> We've received feedback similar to yours in the past, and we're
> currently working on a way of making that distinction (while
> simultaneously solving a few other problems.)  We don't yet have an
> ETA, but know that it's a real priority for us.
> 
> Robert
> 
> On Jul 19, 2007, at 2:37 PM, Augusto Callejas wrote:
> 
>> hi-
>> 
>> i ran that query but all properties have value "true"
>> for the "disambiguator" property hint.
>> 
>> from the type documentation:
>> 
>>   Musical Group Membership is a compound value type which defines the
>>   relationship between a Musical Artist (a band or collaboration)
>> and its
>>   Musical Group Members (musicians).
>> 
>> is there a difference when you say it mediates for all of the
>> properties
>> versus when the documentation says it defines relationships between
>> musical artist and group member?  i'm looking for how that
>> relationship
>> is captured in the type information.  since the disambigurator
>> property
>> hint is the same for all properties, it doesn't define the
>> relationship
>> i'm looking for.
>> 
>> thanks,
>> augusto.
>> 
>> 
>> On 7/19/07 10:52 AM, "Darin Wilson" <darin at metaweb.com> wrote:
>> 
>>> Hi Augusto-
>>> 
>>> A mediating type, or complex value type (CVT), actually mediates for
>>> all of the properties that the type contains; however, the type
>>> designer can choose which of those properties should be displayed
>>> when an instance of the type linked from another topic. If you look
>>> at the type definition of /music/group_membership, you'll see that
>>> each of the properties has the "Display as disambiguator" box
>>> checked. This means that the property will be displayed, if the data
>>> is available.
>>> 
>>> If you look at the instance list for /music/group_membership, you'll
>>> see that some of the instances have years and other don't. That's
>>> because the "Period (start)" and/or "Period (end)" properties haven't
>>> been filled out yet. If the data were available, it would be
>>> displayed.
>>> 
>>> http://www.freebase.com/view/filter?id=%2Fmusic%2Fgroup_membership
>>> 
>>> To see this programmatically, you can use the following query. The /
>>> freebase/type_hints/mediator property on the type will tell you if
>>> it's a CVT or not. The /freebase/property_hints/disambiguator on the
>>> properties will tell you if the value of the property should be
>>> displayed.
>>> 
>>> {
>>>    "query":[{
>>>      "id":"/music/group_membership",
>>>      "type":"/type/type",
>>>      "/freebase/type_hints/mediator":null,
>>>      "properties":[{
>>>        "id":null,
>>>        "/freebase/property_hints/disambiguator":null
>>>      }]
>>>    }]
>>> }
>>> 
>>> HTH,
>>> Darin
>>> 
>>> 
>>> On Jul 19, 2007, at 10:05 AM, Augusto Callejas wrote:
>>> 
>>>> hi-
>>>> 
>>>> regarding mediating types, how do i determine what two other types
>>>> it is mediating for?  for instance, there is the complex type:
>>>> 
>>>>   http://freebase.com/view/filter?id=/music/group_membership
>>>> 
>>>> reading the type description reveals that it relates these two
>>>> types:
>>>> 
>>>>   /music/group_member (via property /music/group_membership/member)
>>>>   /music/musical_group (via property /music/group_membership/group)
>>>> 
>>>> how do i determine this programmatically?  at first, i was going to
>>>> use that fact that these two properties have "unique":true, but
>>>> so is the case for two other properties:
>>>> 
>>>>   /music/group_membership/start
>>>>   /music/group_membership/end
>>>> 
>>>> how can i make the distinction?
>>>> 
>>>> thanks,
>>>> augusto.
>>>> 
>>>> 
>>>> On 6/28/07 3:46 PM, "Dae Park" <daepark at metaweb.com> wrote:
>>>> 
>>>>> We perform a "schema query" for each type a topic is an instance
>>>>> which includes a query for each property and their
>>>>> expected_type. If
>>>>> the expected_type is marked as a mediator then we go another ply
>>>>> out
>>>>> and ask for the property's expected_type's disambiguating
>>>>> properties.
>>>>> 
>>>>> {
>>>>>    id: #topic_id
>>>>>    type: [{
>>>>>      id: null,
>>>>>      properties: [{
>>>>>        id: null,
>>>>>        master_property: null,
>>>>>        reverse_property: null,
>>>>>        unique: null,
>>>>>        expected_type: {
>>>>>          id: null,
>>>>>          "/freebase/type_hints/mediator": null,
>>>>>          properties: [{
>>>>>            master_property: null,
>>>>>            reverse_property: null,
>>>>>            id: null,
>>>>>            unique: null,
>>>>>            "/freebase/type_hints/disambiguator": true
>>>>>          }}
>>>>>        }
>>>>>      }]
>>>>>    }]
>>>>> }
>>>>> 
>>>>> When displaying these mediator properties (sub_properties) from a
>>>>> topic's property:
>>>>> 
>>>>> if the topic_property and sub_property are reciprocal of one
>>>>> another:
>>>>> if sub_property is unique:
>>>>> skip
>>>>> else:
>>>>> foreach value in sub-property:
>>>>> if value == topic we're viewing:
>>>>> continue:
>>>>> display value
>>>>> 
>>>>> 
>>>>> Two properties (prop1, prop2) are reciprocal's of one another if:
>>>>> 
>>>>> prop1.reverse_property == prop2
>>>>> or
>>>>> prop2.master_property == prop1
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> On Jun 28, 2007, at 1:15 PM, Steve Sak wrote:
>>>>> 
>>>>>> 
>>>>>> How does the freebase.com "person" web pages "know" to display the
>>>>>> name
>>>>>> of the spouse/sibling or the name of the organization a person
>>>>>> worked
>>>>>> for under "employment history"? Hard-coded?
>>>>>> 
>>>>>> Nick Thompson wrote:
>>>>>>> in general, i don't think your question has a well-defined
>>>>>>> answer -
>>>>>>> there's no reason for the system to believe that
>>>>>>> /business/employment_tenure/company is more relevant than
>>>>>>> /business/employment_tenure/title, for example.  you could be
>>>>>>> looking
>>>>>>> for a list of people at a particular company, in which case
>>>>>>> the job
>>>>>>> title would more interesting.
>>>>>>> 
>>>>>>> so, you might rule out /business/employment_tenure/person from
>>>>>>> context
>>>>>>> as chris described, but that doesn't determine a unique "other
>>>>>>> side".
>>>>>>> that's the nature of compound value types - if a relationship
>>>>>>> only
>>>>>>> had two sides, it would probably be modeled as a single link
>>>>>>> instead
>>>>>>> of being mediated by a compound value object.
>>>>>>> 
>>>>>>>     nick
>>>>>>> 
>>>>>>> Steve Sak wrote:
>>>>>>>> So, if I'm looking at /people/person/employment_history how
>>>>>>>> would I
>>>>>>>> determine that the relevant property is
>>>>>>>> /business/employment_tenure/company?
>>>>>>>> 
>>>>>>>> 
>>>>>>>> Alec Flett wrote:
>>>>>>>>> I want to add that checking by is only valid if you know you're
>>>>>>>>> coming
>>>>>>>>> in and out on the same property (as is the case with marriage)
>>>>>>>>> - if
>>>>>>>>> the properties are different, you need to look at master/
>>>>>>>>> reverse
>>>>>>>>> properties because you can imagine a director who also
>>>>>>>>> produced a
>>>>>>>>> movie - if you were just going by ids, looking at the
>>>>>>>>> relationship of
>>>>>>>>> 
>>>>>>>>> "Who is the /director/ of this movie /produced by/ Martin
>>>>>>>>> Scorsese?" -
>>>>>>>>> you might get Martin Scorsese back, but that doesn't mean that
>>>>>>>>> "director" is the reverse of "produced by".
>>>>>>>>> 
>>>>>>>>> Alec
>>>>>>>>> 
>>>>>>>>> Christopher R. Maden wrote:
>>>>>>>>>> Steve Sak wrote:
>>>>>>>>>> 
>>>>>>>>>>> To restate the question: given a mediator object how do I
>>>>>>>>>>> determine
>>>>>>>>>>> which property is the one it's mediating?  For example,
>>>>>>>>>>> type A
>>>>>>>>>>> has a
>>>>>>>>>>> mediator M that has a bunch of properties mediates between A
>>>>>>>>>>> and
>>>>>>>>>>> another
>>>>>>>>>>> object.  How do I determine which property of M is the one
>>>>>>>>>>> that
>>>>>>>>>>> points
>>>>>>>>>>> to the other object being mediated?
>>>>>>>>>>> 
>>>>>>>>>> Mediators, or compound value types, are not especially
>>>>>>>>>> different
>>>>>>>>>> from any other type; the UI treats them a little differently,
>>>>>>>>>> but
>>>>>>>>>> that?s it.
>>>>>>>>>> 
>>>>>>>>>> But I think you are asking about instances rather than types,
>>>>>>>>>> correct? In other words, Arnold Schwarzenegger connects to a
>>>>>>>>>> marriage, and that marriage connects to two people, both
>>>>>>>>>> Arnold
>>>>>>>>>> Schwarzenegger and Maria Shriver.  How can you tell which
>>>>>>>>>> of the
>>>>>>>>>> spouses is the one you started from?
>>>>>>>>>> 
>>>>>>>>>> The only way to do that, really, is by looking at the IDs.
>>>>>>>>>> 
>>>>>>>>>> {
>>>>>>>>>>    "query": {
>>>>>>>>>>      "id": null,
>>>>>>>>>>      "name": "Arnold Schwarzenegger",
>>>>>>>>>>      "spouse_s": [
>>>>>>>>>>        {
>>>>>>>>>>          "spouse": [
>>>>>>>>>>            {
>>>>>>>>>>              "id":null,
>>>>>>>>>>              "name":null
>>>>>>>>>>            }
>>>>>>>>>>          ],
>>>>>>>>>>          "type":"/people/marriage"
>>>>>>>>>>        }
>>>>>>>>>>      ],
>>>>>>>>>>      "type":"/people/person"
>>>>>>>>>>    }
>>>>>>>>>> }
>>>>>>>>>> 
>>>>>>>>>> {
>>>>>>>>>>    "result": {
>>>>>>>>>>      "spouse_s": [
>>>>>>>>>>        {
>>>>>>>>>>          "spouse": [
>>>>>>>>>>            {
>>>>>>>>>>              "id": "#9202a8c04000641f80000000001da07e",
>>>>>>>>>>              "name": "Maria Shriver"
>>>>>>>>>>            },
>>>>>>>>>>            {
>>>>>>>>>>              "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>>>              "name": "Arnold Schwarzenegger"
>>>>>>>>>>            }
>>>>>>>>>>          ],
>>>>>>>>>>          "type": "/people/marriage"
>>>>>>>>>>        }
>>>>>>>>>>      ],
>>>>>>>>>>      "type": "/people/person",
>>>>>>>>>>      "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>>>      "name": "Arnold Schwarzenegger"
>>>>>>>>>>    }
>>>>>>>>>> }
>>>>>>>>>> 
>>>>>>>>>> You can see that one of the spouses in the marriage has the
>>>>>>>>>> same ID
>>>>>>>>>> as the person with whom you started, and that the other is
>>>>>>>>>> different.
>>>>>>>>>> 
>>>>>>>>>> Unfortunately, there?s not really an automatic way to say
>>>>>>>>>> ?get all
>>>>>>>>>> of the values of these properties except the ones that point
>>>>>>>>>> back
>>>>>>>>>> to the parent within the scope of this query.?
>>>>>>>>>> 
>>>>>>>>>> HTH,
>>>>>>>>>> Chris
>>>>>>>>>> 
>>>>>>>> 
>>>>>> 
>>>>>> -- 
>>>>>> Steven G Sak
>>>>>> Applied Minds, Inc.
>>>>>> 11718 Bowman Green Drive
>>>>>> Reston, Virginia 20190
>>>>>> o: (703) 483-2207
>>>>>> c: (703) 626-2557
>>>>>> 
>>>>>> _______________________________________________
>>>>>> 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
>>>> 
>>>> 
>>>> _______________________________________________
>>>> 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
>> 
>> 
>> _______________________________________________
>> Developers mailing list
>> Developers at freebase.com
>> http://lists.freebase.com/mailman/listinfo/developers
> 



From alecf at metaweb.com  Thu Jul 19 22:18:10 2007
From: alecf at metaweb.com (Alec Flett)
Date: Thu, 19 Jul 2007 15:18:10 -0700
Subject: [Developers] mediated properties
In-Reply-To: <C2C527C2.E94%acallejas@appliedminds.com>
References: <C2C527C2.E94%acallejas@appliedminds.com>
Message-ID: <469FE322.1030901@metaweb.com>

It's probably useful to describe the two use cases behind compound value 
types, because that's not always obvious, and the two cases aren't 
technically modeled any different from each other right now.

*Use case 1: two-valued CVT*
  This case is generally when there is a relationship between two 
topics, but there is metadata to be associated with the relationship 
itself. A good example of this is "film performance" - really this is a 
relationship between an Actor, and a Film, and the "role" is really just 
extra metadata. What you'd really want in MQL is to be able to do 
something like:

  {"name": "The Departed",
    "type": "/film/film",
   "starring": [] }

and get this result:
  {"name": "The Departed",
   "type": "/film/film/",
    "starring": ["Martin Sheen", "Leonardo DiCaprio", ...]}

And if you expanded this out further, you might learn more:
  {"name": "The Departed",
    "type": "/film/film",
   "starring": [{"*": null}] }

would give you:
{"name": "The Departed",
   "type": "/film/film/",
    "starring": [
    {"actor": "Martin Sheen", "role": "Captain Queenan"},
    {"actor": "Leonardo DiCaprio", "role": "Billy Costigan"},...]}

The key here is that there are two obvious endpoints in the simpler 
query. When visiting the mediator object and "coming in from the film" 
you want the actor, and when you're "coming in from the actor" you want 
the film.

*Use case 2: Multi-valued CVT*
A good example of this is education - a compound value which might 
connect a person to a school and a degree.

{"name": "Alec Flett"
  "type": "/people/person",
   "education": []}

{"name": "Alec Flett"
  "type": "/people/person",
   "education": [null, null, ...]}

Because in this case it isn't really obvious what you meant - did you 
want the school or the degree? Would you expect ["Carnegie Mellon 
University", ...] or ["Bachelor of Science", ...] ?

If I expand it out, you still get the right info
{"name": "Alec Flett"
  "type": "/people/person",
   "education": [{"*": null}]}

This is kinda what we'd want:
{"name": "Alec Flett"
  "type": "/people/person",
   "education": [{"school": "Carnegie Mellon University", "degree": 
"Bachelors of Science", "}]}


At the moment the freebase.com user interface looks at all the 
properties which are "disambiguating" and uses them. Here are the 
disambiguating properties of /education/education:
{
      "id":"/education/education",
      "properties":[{
        "/freebase/property_hints/disambiguator":true,
        "key":"degree"
      },{
        "/freebase/property_hints/disambiguator":true,
        "key":"institution"
      },{
        "/freebase/property_hints/disambiguator":true,
        "key":"start_date"
      },{
        "/freebase/property_hints/disambiguator":true,
        "key":"end_date"
      },{
        "/freebase/property_hints/disambiguator":true,
        "key":"student"
      },{
        "/freebase/property_hints/disambiguator":true,
        "key":"major_field_of_study"
      },{
        "/freebase/property_hints/disambiguator":true,
        "key":"dissertation_title"
      }]
    }

So at the moment what you've pointed out is that we currently model ALL 
compound values as "case 2" - there's no obvious left or right. 
Ultimately what we'd like is for MQL to understand case 1, and that 
would make compound values much easier to use.

Alec

Augusto Callejas wrote:
> hi-
>
> i ran that query but all properties have value "true"
> for the "disambiguator" property hint.
>
> from the type documentation:
>
>   Musical Group Membership is a compound value type which defines the
>   relationship between a Musical Artist (a band or collaboration) and its
>   Musical Group Members (musicians).
>
> is there a difference when you say it mediates for all of the properties
> versus when the documentation says it defines relationships between
> musical artist and group member?  i'm looking for how that relationship
> is captured in the type information.  since the disambigurator property
> hint is the same for all properties, it doesn't define the relationship
> i'm looking for.
>
> thanks,
> augusto.
>
>
> On 7/19/07 10:52 AM, "Darin Wilson" <darin at metaweb.com> wrote:
>
>   
>> Hi Augusto-
>>
>> A mediating type, or complex value type (CVT), actually mediates for
>> all of the properties that the type contains; however, the type
>> designer can choose which of those properties should be displayed
>> when an instance of the type linked from another topic. If you look
>> at the type definition of /music/group_membership, you'll see that
>> each of the properties has the "Display as disambiguator" box
>> checked. This means that the property will be displayed, if the data
>> is available.
>>
>> If you look at the instance list for /music/group_membership, you'll
>> see that some of the instances have years and other don't. That's
>> because the "Period (start)" and/or "Period (end)" properties haven't
>> been filled out yet. If the data were available, it would be displayed.
>>
>> http://www.freebase.com/view/filter?id=%2Fmusic%2Fgroup_membership
>>
>> To see this programmatically, you can use the following query. The /
>> freebase/type_hints/mediator property on the type will tell you if
>> it's a CVT or not. The /freebase/property_hints/disambiguator on the
>> properties will tell you if the value of the property should be
>> displayed.
>>
>> {
>>    "query":[{
>>      "id":"/music/group_membership",
>>      "type":"/type/type",
>>      "/freebase/type_hints/mediator":null,
>>      "properties":[{
>>        "id":null,
>>        "/freebase/property_hints/disambiguator":null
>>      }]
>>    }]
>> }
>>
>> HTH,
>> Darin
>>
>>
>> On Jul 19, 2007, at 10:05 AM, Augusto Callejas wrote:
>>
>>     
>>> hi-
>>>
>>> regarding mediating types, how do i determine what two other types
>>> it is mediating for?  for instance, there is the complex type:
>>>
>>>   http://freebase.com/view/filter?id=/music/group_membership
>>>
>>> reading the type description reveals that it relates these two types:
>>>
>>>   /music/group_member (via property /music/group_membership/member)
>>>   /music/musical_group (via property /music/group_membership/group)
>>>
>>> how do i determine this programmatically?  at first, i was going to
>>> use that fact that these two properties have "unique":true, but
>>> so is the case for two other properties:
>>>
>>>   /music/group_membership/start
>>>   /music/group_membership/end
>>>
>>> how can i make the distinction?
>>>
>>> thanks,
>>> augusto.
>>>
>>>
>>> On 6/28/07 3:46 PM, "Dae Park" <daepark at metaweb.com> wrote:
>>>
>>>       
>>>> We perform a "schema query" for each type a topic is an instance
>>>> which includes a query for each property and their expected_type. If
>>>> the expected_type is marked as a mediator then we go another ply out
>>>> and ask for the property's expected_type's disambiguating properties.
>>>>
>>>> {
>>>>    id: #topic_id
>>>>    type: [{
>>>>      id: null,
>>>>      properties: [{
>>>>        id: null,
>>>>        master_property: null,
>>>>        reverse_property: null,
>>>>        unique: null,
>>>>        expected_type: {
>>>>          id: null,
>>>>          "/freebase/type_hints/mediator": null,
>>>>          properties: [{
>>>>            master_property: null,
>>>>            reverse_property: null,
>>>>            id: null,
>>>>            unique: null,
>>>>            "/freebase/type_hints/disambiguator": true
>>>>          }}
>>>>        }
>>>>      }]
>>>>    }]
>>>> }
>>>>
>>>> When displaying these mediator properties (sub_properties) from a
>>>> topic's property:
>>>>
>>>> if the topic_property and sub_property are reciprocal of one another:
>>>> if sub_property is unique:
>>>> skip
>>>> else:
>>>> foreach value in sub-property:
>>>> if value == topic we're viewing:
>>>> continue:
>>>> display value
>>>>
>>>>
>>>> Two properties (prop1, prop2) are reciprocal's of one another if:
>>>>
>>>> prop1.reverse_property == prop2
>>>> or
>>>> prop2.master_property == prop1
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Jun 28, 2007, at 1:15 PM, Steve Sak wrote:
>>>>
>>>>         
>>>>> How does the freebase.com "person" web pages "know" to display the
>>>>> name
>>>>> of the spouse/sibling or the name of the organization a person
>>>>> worked
>>>>> for under "employment history"? Hard-coded?
>>>>>
>>>>> Nick Thompson wrote:
>>>>>           
>>>>>> in general, i don't think your question has a well-defined answer -
>>>>>> there's no reason for the system to believe that
>>>>>> /business/employment_tenure/company is more relevant than
>>>>>> /business/employment_tenure/title, for example.  you could be
>>>>>> looking
>>>>>> for a list of people at a particular company, in which case the job
>>>>>> title would more interesting.
>>>>>>
>>>>>> so, you might rule out /business/employment_tenure/person from
>>>>>> context
>>>>>> as chris described, but that doesn't determine a unique "other
>>>>>> side".
>>>>>> that's the nature of compound value types - if a relationship only
>>>>>> had two sides, it would probably be modeled as a single link
>>>>>> instead
>>>>>> of being mediated by a compound value object.
>>>>>>
>>>>>>     nick
>>>>>>
>>>>>> Steve Sak wrote:
>>>>>>             
>>>>>>> So, if I'm looking at /people/person/employment_history how
>>>>>>> would I
>>>>>>> determine that the relevant property is
>>>>>>> /business/employment_tenure/company?
>>>>>>>
>>>>>>>
>>>>>>> Alec Flett wrote:
>>>>>>>               
>>>>>>>> I want to add that checking by is only valid if you know you're
>>>>>>>> coming
>>>>>>>> in and out on the same property (as is the case with marriage)
>>>>>>>> - if
>>>>>>>> the properties are different, you need to look at master/reverse
>>>>>>>> properties because you can imagine a director who also produced a
>>>>>>>> movie - if you were just going by ids, looking at the
>>>>>>>> relationship of
>>>>>>>>
>>>>>>>> "Who is the /director/ of this movie /produced by/ Martin
>>>>>>>> Scorsese?" -
>>>>>>>> you might get Martin Scorsese back, but that doesn't mean that
>>>>>>>> "director" is the reverse of "produced by".
>>>>>>>>
>>>>>>>> Alec
>>>>>>>>
>>>>>>>> Christopher R. Maden wrote:
>>>>>>>>                 
>>>>>>>>> Steve Sak wrote:
>>>>>>>>>
>>>>>>>>>                   
>>>>>>>>>> To restate the question: given a mediator object how do I
>>>>>>>>>> determine
>>>>>>>>>> which property is the one it's mediating?  For example, type A
>>>>>>>>>> has a
>>>>>>>>>> mediator M that has a bunch of properties mediates between A
>>>>>>>>>> and
>>>>>>>>>> another
>>>>>>>>>> object.  How do I determine which property of M is the one that
>>>>>>>>>> points
>>>>>>>>>> to the other object being mediated?
>>>>>>>>>>
>>>>>>>>>>                     
>>>>>>>>> Mediators, or compound value types, are not especially different
>>>>>>>>> from any other type; the UI treats them a little differently,
>>>>>>>>> but
>>>>>>>>> that?s it.
>>>>>>>>>
>>>>>>>>> But I think you are asking about instances rather than types,
>>>>>>>>> correct? In other words, Arnold Schwarzenegger connects to a
>>>>>>>>> marriage, and that marriage connects to two people, both Arnold
>>>>>>>>> Schwarzenegger and Maria Shriver.  How can you tell which of the
>>>>>>>>> spouses is the one you started from?
>>>>>>>>>
>>>>>>>>> The only way to do that, really, is by looking at the IDs.
>>>>>>>>>
>>>>>>>>> {
>>>>>>>>>    "query": {
>>>>>>>>>      "id": null,
>>>>>>>>>      "name": "Arnold Schwarzenegger",
>>>>>>>>>      "spouse_s": [
>>>>>>>>>        {
>>>>>>>>>          "spouse": [
>>>>>>>>>            {
>>>>>>>>>              "id":null,
>>>>>>>>>              "name":null
>>>>>>>>>            }
>>>>>>>>>          ],
>>>>>>>>>          "type":"/people/marriage"
>>>>>>>>>        }
>>>>>>>>>      ],
>>>>>>>>>      "type":"/people/person"
>>>>>>>>>    }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> {
>>>>>>>>>    "result": {
>>>>>>>>>      "spouse_s": [
>>>>>>>>>        {
>>>>>>>>>          "spouse": [
>>>>>>>>>            {
>>>>>>>>>              "id": "#9202a8c04000641f80000000001da07e",
>>>>>>>>>              "name": "Maria Shriver"
>>>>>>>>>            },
>>>>>>>>>            {
>>>>>>>>>              "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>>              "name": "Arnold Schwarzenegger"
>>>>>>>>>            }
>>>>>>>>>          ],
>>>>>>>>>          "type": "/people/marriage"
>>>>>>>>>        }
>>>>>>>>>      ],
>>>>>>>>>      "type": "/people/person",
>>>>>>>>>      "id": "#9202a8c04000641f8000000000006567",
>>>>>>>>>      "name": "Arnold Schwarzenegger"
>>>>>>>>>    }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> You can see that one of the spouses in the marriage has the
>>>>>>>>> same ID
>>>>>>>>> as the person with whom you started, and that the other is
>>>>>>>>> different.
>>>>>>>>>
>>>>>>>>> Unfortunately, there?s not really an automatic way to say
>>>>>>>>> ?get all
>>>>>>>>> of the values of these properties except the ones that point
>>>>>>>>> back
>>>>>>>>> to the parent within the scope of this query.?
>>>>>>>>>
>>>>>>>>> HTH,
>>>>>>>>> Chris
>>>>>>>>>
>>>>>>>>>                   
>>>>> -- 
>>>>> Steven G Sak
>>>>> Applied Minds, Inc.
>>>>> 11718 Bowman Green Drive
>>>>> Reston, Virginia 20190
>>>>> o: (703) 483-2207
>>>>> c: (703) 626-2557
>>>>>
>>>>> _______________________________________________
>>>>> 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
>>>>         
>>> _______________________________________________
>>> 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
>>     
>
>
> _______________________________________________
> 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/20070719/fec23c26/attachment-0001.htm 

From robert at metaweb.com  Thu Jul 19 22:27:26 2007
From: robert at metaweb.com (Robert Cook)
Date: Thu, 19 Jul 2007 15:27:26 -0700
Subject: [Developers] mediated properties
In-Reply-To: <C2C530E8.E99%acallejas@appliedminds.com>
References: <C2C530E8.E99%acallejas@appliedminds.com>
Message-ID: <116BC840-EE9C-4AC8-8165-F0B0FBC714C1@metaweb.com>

On Jul 19, 2007, at 3:16 PM, Augusto Callejas wrote:

> robert-
>
> thanks that does clear it up. going by the definition of the meditator
> pattern, does this type qualify as a mediator:
>
>   http://www.freebase.com/view/schema?id=/measurement_unit/rect_size
>
> i wouldn't think so since it only has two properties that are of value
> types.

You are right, this is *not* a mediator, just a CVT.  One way to tell  
in this case, there is nothing 'passing through' it.  Put another  
way, it has no properties that are reciprocated from the other end.


> another way to state the question is: are all compound value types  
> mediators
> (e.g. opposed to all mediators being compund value types)?

Actually, it's the opposite -- Compound Value Types are *both* the  
unnamed example you just described and mediators.  I think the  
problem here is that we don't have a more specific name for your  
example, and "compound value" is pretty descriptive for it.

I think you've highlighted a problem.  We need some better terminology.

R


> thanks,
> augusto.
>
>
>
> On 7/19/07 2:59 PM, "Robert Cook" <robert at metaweb.com> wrote:
>
>> Augusto --
>>
>> You are asking a very good question.  A "mediator" is a data pattern
>> that is represented as a Compound Value Type in Metaweb.  In other
>> words, mediators are CVTs that act like an "annotated relationship"
>> between two types.  One example is "Marriage", which joins two people
>> together that also stores add start and end dates as metadata.
>>
>> Currently, we don't have a way to annotate which of the two
>> properties represent the "relationship" between two types.  That is,
>> a mediator is more of a data idiom that makes sense to the modeler,
>> but has no formal representation in the data store.
>>
>> We've received feedback similar to yours in the past, and we're
>> currently working on a way of making that distinction (while
>> simultaneously solving a few other problems.)  We don't yet have an
>> ETA, but know that it's a real priority for us.
>>
>> Robert
>>
>> On Jul 19, 2007, at 2:37 PM, Augusto Callejas wrote:
>>
>>> hi-
>>>
>>> i ran that query but all properties have value "true"
>>> for the "disambiguator" property hint.
>>>
>>> from the type documentation:
>>>
>>>   Musical Group Membership is a compound value type which defines  
>>> the
>>>   relationship between a Musical Artist (a band or collaboration)
>>> and its
>>>   Musical Group Members (musicians).
>>>
>>> is there a difference when you say it mediates for all of the
>>> properties
>>> versus when the documentation says it defines relationships between
>>> musical artist and group member?  i'm looking for how that
>>> relationship
>>> is captured in the type information.  since the disambigurator
>>> property
>>> hint is the same for all properties, it doesn't define the
>>> relationship
>>> i'm looking for.
>>>
>>> thanks,
>>> augusto.
>>>
>>>
>>> On 7/19/07 10:52 AM, "Darin Wilson" <darin at metaweb.com> wrote:
>>>
>>>> Hi Augusto-
>>>>
>>>> A mediating type, or complex value type (CVT), actually mediates  
>>>> for
>>>> all of the properties that the type contains; however, the type
>>>> designer can choose which of those properties should be displayed
>>>> when an instance of the type linked from another topic. If you look
>>>> at the type definition of /music/group_membership, you'll see that
>>>> each of the properties has the "Display as disambiguator" box
>>>> checked. This means that the property will be displayed, if the  
>>>> data
>>>> is available.
>>>>
>>>> If you look at the instance list for /music/group_membership,  
>>>> you'll
>>>> see that some of the instances have years and other don't. That's
>>>> because the "Period (start)" and/or "Period (end)" properties  
>>>> haven't
>>>> been filled out yet. If the data were available, it would be
>>>> displayed.
>>>>
>>>> http://www.freebase.com/view/filter?id=%2Fmusic%2Fgroup_membership
>>>>
>>>> To see this programmatically, you can use the following query.  
>>>> The /
>>>> freebase/type_hints/mediator property on the type will tell you if
>>>> it's a CVT or not. The /freebase/property_hints/disambiguator on  
>>>> the
>>>> properties will tell you if the value of the property should be
>>>> displayed.
>>>>
>>>> {
>>>>    "query":[{
>>>>      "id":"/music/group_membership",
>>>>      "type":"/type/type",
>>>>      "/freebase/type_hints/mediator":null,
>>>>      "properties":[{
>>>>        "id":null,
>>>>        "/freebase/property_hints/disambiguator":null
>>>>      }]
>>>>    }]
>>>> }
>>>>
>>>> HTH,
>>>> Darin
>>>>
>>>>
>>>> On Jul 19, 2007, at 10:05 AM, Augusto Callejas wrote:
>>>>
>>>>> hi-
>>>>>
>>>>> regarding mediating types, how do i determine what two other types
>>>>> it is mediating for?  for instance, there is the complex type:
>>>>>
>>>>>   http://freebase.com/view/filter?id=/music/group_membership
>>>>>
>>>>> reading the type description reveals that it relates these two
>>>>> types:
>>>>>
>>>>>   /music/group_member (via property /music/group_membership/ 
>>>>> member)
>>>>>   /musi