[Freebase-discuss] Updating the business schema: UPDATE

Tue Apr 27 00:49:51 UTC 2010

me likey. comments:

- shouldn't "child" in the organization2 schema read "children". or
maybe child(ren). ugh, both are yuck.
- organization type: should this really be a free-for-all? w/o clarity
on how to use the field it could get ugly real quick
- business operation/locations: is this for locations (aka New York
City) or for specific business offices (aka Viacom NYC Headquarters).
there's been confusion on this in the past, and it isn't clear to me
what's happening with the new format (I just put in New York City for
Viacom and didn't see anything on the NYC topic so I'm not sure)
- divisions: how are we differentiating between a legal entity and a
non-legal division? they look to be treated the same now. or do we not
care about the legal difference?

I think that's it for me. awesome job. and if there's anything I can
do to help, please let me know (I finally have some more bandwidth
lately).

evening

On Mon, Apr 26, 2010 at 7:40 PM, Jeff Prucher <jeff at metaweb.com> wrote:
> Toby has added a bunch of types and properties to the Business2 base
> (http://business2.freebase.com/). Those of you who are interested in this
> schema should take a look at it.  I've updated the main wiki page about the
> refactoring to reflect the additions:
> http://wiki.freebase.com/wiki/Business_Commons/Updated_business_schema
>
> I've also created a page with the mappings between the properties in the old
> and new schemata:
> http://wiki.freebase.com/wiki/Business_and_Organization_Refactoring
>
> Let us know what you think.
>
> Jeff
>
>> -----Original Message-----
>> From: freebase-discuss-bounces at freebase.com
>> [mailto:freebase-discuss-bounces at freebase.com] On Behalf Of
>> Toby Segaran
>> Sent: Monday, April 19, 2010 4:43 PM
>> To: Freebase.com discussion list
>> Subject: Re: [Freebase-discuss] Updating the business schema
>>
>> Yep, working on a "Defunct Organization" type which will
>> replace the "Date dissolved" field and the "Defunct company" type.
>>
>> On Apr 19, 2010, at 12:48 PM, Tom Morris wrote:
>>
>> > On Thu, Apr 15, 2010 at 3:36 PM, Toby Segaran <toby at metaweb.com>
>> > wrote:
>> >> I'm going to stick with From and To dates for now and have the
>> >> meaning be slightly ambiguous where I don't have complete data.
>> >>
>> >> Does anyone have other strong objections to the schema?
>> I'd like to
>> >> start migrating data to it soon.
>> >
>> > I think it makes sense to move ahead sooner rather than
>> later.  Having
>> > the two different schemas visible will just cause confusion.
>> >
>> > I had it on my list to ask you about "place founded" but I see that
>> > got added in the last few days.
>> >
>> > It looks like "date dissolved" is not in the new schema.  Is the
>> > intent to move that to a Defunct Organization type or just
>> delete it?
>> >
>> > Do you have a table of old->new mappings?  I didn't see one in the
>> > wiki, but I imagine one will be needed anyway for the data
>> migration
>> > process.  That would help identify any remaining holes.
>> >
>> > Tom
>> >
>> >>
>> >> On Apr 9, 2010, at 11:15 AM, Tom Morris wrote:
>> >>
>> >>> On Thu, Apr 8, 2010 at 6:05 PM, Toby Segaran <toby at metaweb.com>
>> >>> wrote:
>> >>>>
>> >>>> I can see that I was a little too eager to adapt the
>> schema to my
>> >>>> particular data set :)
>> >>>>
>> >>>> The idea of upper and lower bounds is clever, in that it
>> actually
>> >>>> captures every case, but I think it will be way too
>> complicated for
>> >>>> most people entering or reading data, unless we made significant
>> >>>> changes to the UI and to MQL. Generally we don't have
>> any methods
>> >>>> for dealing with imprecise or bounded values in Freebase.
>> >>>>
>> >>>> I can really only see two possibilities:
>> >>>> - Include "From", "To" and "As of" dates, allowing
>> people to fill
>> >>>> in the data they have. This feels clunky and there's not
>> really any
>> >>>> precedent in Freebase for these kind of redundant models.
>> >>>
>> >>> This suffers from the same complexity issue that date
>> ranges does in
>> >>> the absence of enhanced query support.
>> >>>
>> >>>> - Go back to "From" and "To" fields, and set them to the
>> earliest
>> >>>> and latest as-of dates respectively. This will require
>> additional
>> >>>> parsing code for the SEC stuff that updates the CVTs,
>> but I'm happy
>> >>>> to write that. The obvious downside is that it doesn't tell you
>> >>>> when a tenure started or ended, just the earliest and
>> latest known
>> >>>> dates.
>> >>>
>> >>> That sounds like the best immediate term solution.  Pedantically
>> >>> it's recording (slightly) inaccurate data, but in most practical
>> >>> cases it'll be fine.
>> >>>
>> >>> Longer term I'd still like to see some thought put into a more
>> >>> sophisticated date model.  I think it's a tractable problem and
>> >>> solving it would provide some powerful capabilities.
>> >>>
>> >>> Tom
>> >>>
>> >>>>
>> >>>> I really don't want to just leave the date fields blank when
>> >>>> filling in directors, etc., as this destroys much of the utility
>> >>>> that Freebase can offer here (overlapping directorates,
>> conflicts
>> >>>> of interest, tracing career paths, etc.)
>> >>>>
>> >>>> Thoughts?
>> >>>>
>> >>>> Toby
>> >>>>
>> >>>> On Apr 8, 2010, at 7:11 AM, Tom Morris wrote:
>> >>>>
>> >>>>> On Thu, Apr 8, 2010 at 3:37 AM, Iain Sproat
>> <iainsproat at gmail.com>
>> >>>>> wrote:
>> >>>>>
>> >>>>>> The underlying issue is that we have as_of data from
>> an external
>> >>>>>> data source, but can't accurately imply the 'from' and 'to'
>> >>>>>> dates.  It would be great if we could enter uncertain dates in
>> >>>>>> the 'from'
>> >>>>>> date as
>> >>>>>> "before {as_of date}", which would solve the issue.
>> >>>>>
>> >>>>> Yes, I meant to tie this back in to our earlier
>> discussion of date
>> >>>>> ranges and imprecision.  There are many cases in researching
>> >>>>> historical people where you get evidence that only gives you a
>> >>>>> bound on the date for an event.  For example, you know
>> they were
>> >>>>> still alive when they wrote their will or that they
>> must have been
>> >>>>> born before date X because they'd reached their
>> majority by date
>> >>>>> Y.  For job titles, a reference to "the former of CEO
>> of BigCo" on
>> >>>>> date Z gives an upper bound on the "to" date for that
>> role.  The
>> >>>>> "as of" date is really just a closed period with a "from" of
>> >>>>> "before date D" and a
>> >>>>> "to" of "after date D" (with the same date for both ends).
>> >>>>> Being able
>> >>>>> to record dates in this way would allow you to combine multiple
>> >>>>> sets of evidence to derive a more tightly bounded, or
>> even exact,
>> >>>>> period.
>> >>>>>
>> >>>>> Tom
>> >>>>> _______________________________________________
>> >>>>> You are receiving this message because you are
>> subscribed to the
>> >>>>> Freebase-discuss mailing list.
>> >>>>> To post a message to the list: Freebase-discuss at freebase.com To
>> >>>>> unsubscribe, view archives, etc:
>> >>>>> http://lists.freebase.com/mailman/listinfo/freebase-discuss
>> >>>>
>> >>>>
>> >>>> _______________________________________________
>> >>>> You are receiving this message because you are subscribed to the
>> >>>> Freebase-discuss mailing list.
>> >>>> To post a message to the list: Freebase-discuss at freebase.com To
>> >>>> unsubscribe, view archives, etc:
>> >>>> http://lists.freebase.com/mailman/listinfo/freebase-discuss
>> >>>>
>> >>> _______________________________________________
>> >>> You are receiving this message because you are subscribed to the
>> >>> Freebase-discuss mailing list.
>> >>> To post a message to the list: Freebase-discuss at freebase.com To
>> >>> unsubscribe, view archives, etc:
>> >>> http://lists.freebase.com/mailman/listinfo/freebase-discuss
>> >>
>> >>
>> >> _______________________________________________
>> >> You are receiving this message because you are subscribed to the
>> >> Freebase-discuss mailing list.
>> >> To post a message to the list: Freebase-discuss at freebase.com To
>> >> unsubscribe, view archives, etc:
>> >> http://lists.freebase.com/mailman/listinfo/freebase-discuss
>> >>
>> > _______________________________________________
>> > You are receiving this message because you are subscribed to the
>> > Freebase-discuss mailing list.
>> > To post a message to the list: Freebase-discuss at freebase.com To
>> > unsubscribe, view archives, etc:
>> > http://lists.freebase.com/mailman/listinfo/freebase-discuss
>>
>>
>
> _______________________________________________
> You are receiving this message because you are subscribed to the Freebase-discuss mailing list.
> To post a message to the list: Freebase-discuss at freebase.com
> To unsubscribe, view archives, etc: http://lists.freebase.com/mailman/listinfo/freebase-discuss
>