[Data-modeling] Adding a date/time CVT to organization members
Jeff Thompson
jeff at thefirst.org
Tue Feb 26 03:21:03 UTC 2008
I agree that handling the CVT indirection (as John pointed out) is a technical
problem that I trust you'll figure out. But if I may keep the candle lit for
that future day when these are worked out...
Jeff Prucher wrote:
> Most properties on creative works -- author of a poem, performer on a
> musical track, media of a painting, cover artist for an edition of a book,
> director of a film, etc.
> Contents of a magazine issue
> Vineyard a vintage of wine's grapes were from
These properties are fixed because they are exactly what the creator
of the work says they are. The author of a poem publishes a book of
poems and says "I'm the author", etc. Unlike natural phenomena, or
uncertain data, these tautologically track human-generated information which is what
the information source says they are, hence they are fixed. And also
arguably not that interesting for data mining.
> Parent/Child/siblings (of a person or other organism)
If I may niggle, sibling is already a relation CVT, I assume
because it needs to be an N-way relation, not just a binary relation
which is all that a "normal" property can handle.
> Maximum force of a (past) cyclone's winds
For measurements of natural phenomena like this, new techniques improve the
measurements even of past events. I understand that many just want Freebase
to have the "one correct approved consensus answer". But this would
be to decide up front (even before Freebase leaves Alpha) that no information
needs to be captured about the refinement of these measurements over time,
or the honest disagreements of different groups who use different measurement
techniques. Fortunately, I think the fate of Freebase will be better than that,
once those technical details are worked out.
> Jeff Prucher
>
>> -----Original Message-----
>> From: data-modeling-bounces at freebase.com
>> [mailto:data-modeling-bounces at freebase.com] On Behalf Of Jeff Thompson
>> Sent: Friday, February 22, 2008 11:24 PM
>> To: Freebase data modeling mailing list
>> Subject: Re: [Data-modeling] Adding a date/time CVT to
>> organization members
>>
>> Each time this comes up, it's my job to be the a**h*ole and
>> ask the converse question:
>> Can you say what properties on any type you will *not* need
>> to track changing values over time...
>>
>> Jeff Prucher wrote:
>>> We have had a request to insert a CVT between the types
>> "organization"
>>> and "organization member" in order to be able to track information
>>> about the periods over which members have belonged to
>> organizations.
>>> The CVT will simply add start and end dates. (The properties that
>>> would be affected are /organization/organization/members and
>>>
>> /organization/organization_member/member_of_these_organization
>> s.) This
>>> is a very straightforward fix to do, but doing so will break any
>>> applications using these properties. Is anybody using these
>>> properties, and if so, would this be a problem for you?
>> We'd probably
>>> make the change in the middle of March.
>>>
>>> Jeff Prucher
>>> Type Librarian & Ontologist
>>> Metaweb Technologies, Inc.
>>>
>>>
>>> _______________________________________________
>>> Data-modeling mailing list
>>> Data-modeling at freebase.com
>>> http://lists.freebase.com/mailman/listinfo/data-modeling
>>>
>>>
>>>
>> _______________________________________________
>> Data-modeling mailing list
>> Data-modeling at freebase.com
>> http://lists.freebase.com/mailman/listinfo/data-modeling
>>
>
> _______________________________________________
> Data-modeling mailing list
> Data-modeling at freebase.com
> http://lists.freebase.com/mailman/listinfo/data-modeling
>
>
>
More information about the Data-modeling
mailing list