[Data-modeling] Current heads of state
Tom Morris
tfmorris at gmail.com
Fri Jun 27 21:01:04 UTC 2008
I think the denormalization will cause more trouble than it's worth.
Another advantage to the dated approach, perhaps with some UI
enhancements that can help users query "current," is that it's fault
resilient. If I ask for the mayor as of 1 July 2008 and the latest
entry is Joe Smith 2004-2006, then I know I don't have a valid value
to work with.
Doing the "current" processing during the query is more robust than
storing this information in the database.
Tom
On Tue, Jun 24, 2008 at 6:35 PM, Ed Laurent <spatial.db at gmail.com> wrote:
> I still lean towards a simpler model with start and end dates that can be
> used to infer whether something is current or not but I also see the value
> of this denormalization. An empty "current head of state" would be easy to
> find for the initial data entry. However, I don't think it is any more
> useful in the long term unless someone keeps on top dynamic topic
> properties. Even then, the denormalization could make for really clunky data
> entry unless there was a way to automatically update both the current and
> historical properties when changes occur. Empty fields and a lack of
> updating can throw a monkey wrench into either approach.
>
> -Ed
>
> On Tue, Jun 24, 2008 at 5:53 PM, Jeff Prucher <jeff at metaweb.com> wrote:
>>
>> There's a discussion thread started by user joguinn on the /location type
>> that I'd like to get more input on. The original thread is here:
>>
>> <http://www.freebase.com/discuss/threads/location#/guid/9202a8c04000641f8000
>> 0000086ea1f8> (it's the last few comments in the thread "City Parks/Golf
>> Courses").
>>
>> The request is to include a property for "current head of state" on the
>> type
>> Country. This would be a partial denormalization of the Governing
>> Officials
>> property on the Governmental Jurisdiction type, but consistent with a
>> pattern that we've seen several times lately. It would also have some
>> additional semantic value, since we have no way in the current model to
>> assert that any particular office represents the head of state. The
>> reason
>> I'm bringing it up here is that I can imagine this snowballing into a slew
>> of "Current X" properties on various location types ("Current Mayor",
>> "Current Legislators", "Current Insurance Commissioner", etc.) and I'd
>> like
>> to hear what people think about either the proposed property as it stands,
>> or the potential proliferation of similar properties.
>>
>> Thanks,
>> Jeff
More information about the Data-modeling
mailing list