AFAIK the hold-up is because the "numbers" data type should also be able to
carry an uncertainty, and a unit. Not really necessary for "number of
floors", but for plenty of other applications.



On Mon, Apr 22, 2013 at 11:21 AM, Michael Hale <[email protected]>wrote:

> Well, there is a trade-off. If we have too few types then it is harder for
> clients to use the data and inconsistencies might slip in. But if we delay
> the approval of too many properties because we want more types then we risk
> unnecessarily slowing the growth of the database. It's all binary
> information ultimately. Datatypes just add some constraints that block
> certain binary sequences from being stored in certain areas. The list of
> properties that are awaiting specific datatypes are standard datatypes
> found in many databases, but it is just an artifact of history that we
> consider those datatypes to be standard. For example, you see that "number
> of floors" is awaiting the number datatype. That's nice because it will
> prevent someone from entering "abc" as the number of floors in the Burj
> Khalifa. But just constraining the information to be a number doesn't
> prevent someone from entering -10 as the number of floors. So should we add
> an additional datatype for positive integers? If we try to specify too many
> server-side constraints on the information then our efforts to be supremely
> responsible become irresponsible because, for example, we are just slowing
> the ability of Wikidata to provide better templates to less popular
> language Wikipedias.
>
> ------------------------------
> Date: Mon, 22 Apr 2013 11:56:37 +0200
>
> From: [email protected]
> To: [email protected]
> Subject: Re: [Wikidata-l] phase 2 live on the first 11 Wikipedias
>
> Why are there so many properties waiting on
> http://www.wikidata.org/wiki/Wikidata:Property_proposal/Pending then???
>
>
> 2013/3/29 Michael Hale <[email protected]>
>
> It would be nice to never have to do any additional validation on the
> client, but in general I'm pretty sure it's impossible to specify server
> validation for every type in advance. Let's assume I want to upload gene
> sequences. I could go through the long process of waiting to get a new type
> approved and code added on the website just for the minor check of making
> sure this is a string of just As, Gs, Cs, and Ts, or I could just use
> generic strings and start right now.
>
> ------------------------------
> Date: Fri, 29 Mar 2013 17:42:29 +0100
>
> From: [email protected]
> To: [email protected]
> Subject: Re: [Wikidata-l] phase 2 live on the first 11 Wikipedias
>
> > parse it into a date on the client-side.
>
> this is exactly why I do not want to do. we should validate inputs, not
> (try to) parse outputs
>
>
> 2013/3/29 Michael Hale <[email protected]>
>
> It doesn't hinder the expansion of Wikidata because the current types are
> flexible. Added types will improve performance and simplify client access
> for some scenarios, but every scenario I can think of is possible with the
> currently available types. If you have data that you would like to enter, I
> suggest doing so and then you can switch some properties to the new types
> when they are added if you want. The three types currently supported for
> properties are lists of items, media, and strings. You can group different
> types together for a single property by creating a new item to hold them.
> Soon you will be able to add an extra property that can be used to
> differentiate values in a list, called a qualifier. Other types are coming,
> but for example, if you want to use a date now you can just use a string
> and then parse it into a date on the client-side.
>
> ------------------------------
> Date: Fri, 29 Mar 2013 16:54:23 +0100
> From: [email protected]
> To: [email protected]
> Subject: Re: [Wikidata-l] phase 2 live on the first 11 Wikipedias
>
>
> Contrats! Anyways, I still miss various data types to be available for
> properties - unavailability of these greatly hinders the expansion of
> wikidata database. Any plans on how fast new (and comprehensive) data types
> will be introduced to wikidata.org?
>
>
> 2013/3/28 Daniel Kinzler <[email protected]>
>
> On 28.03.2013 11:24, Lydia Pintscher wrote:
> >> In the announcement blog, {{#property:chief executive officer}} is not
> >> ready, do you know the rough plan on this? maybe end of Q2?
> >
> > It should come soon. Hopefully already in the next deployment but no
> > promises. We're currently figuring out the schedule for that with
> > Easter complicating things a bit.
>
> To give some idea in addition to what Lydia said:
>
> I wrote the code for that just now, while on the train from Berlin. Now it
> needs
> to be reviewed, tested, and, if all goes well, deployed. When that would
> be I'm
> not sure, with people taking vacations over the holidays, but I'd speak of
> weeks, not months. But it's quite a hunk of code, so delays are possible.
>
> cheers,
> daniel
>
>
> --
> Daniel Kinzler, Softwarearchitekt
> Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e. V.
>
>
> _______________________________________________
> Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
>
>
> _______________________________________________ Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
> _______________________________________________
> Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
>
>
> _______________________________________________ Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
> _______________________________________________
> Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
>
>
> _______________________________________________ Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
> _______________________________________________
> Wikidata-l mailing list
> [email protected]
> https://lists.wikimedia.org/mailman/listinfo/wikidata-l
>
>
_______________________________________________
Wikidata-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata-l

Reply via email to