You mean a https://en.wikipedia.org/wiki/Decision_tree ? Yeah, we probably already have enough material to build some ...


2013/11/21 Gerard Meijssen <gerard.meijssen@gmail.com>
Hoi Tom²,
I agree with you for the motivational stuff. However ... the  Reasonator depends on what triggers its special functionality. I do understand what makes a human to the current Reasonator. Bottom line, "is a" "human" is the best option for this. However, when we are to make special functionality for specific cases (I intentionally do not use big words here) it could be argued that we need to go the way of the upper ontologies (this sounds to me like the way of the Dodo).. One way to circumvent this is to associate particular properties with associated questions... eg.. date of birth "Is this a human or an animal?"

What is also relevant is that we are working with 280+ natural languages at this time. The language committee changed its policy so that any recognised language can be included... There are over 7000 of those.. This means that we have to consider how to express things without using words to much.
Thanks,
        GerardM

PS as far as I am concerned we do need the high brow instrumentation but it needs to be out of the way as much as possible.


On 21 November 2013 13:14, Thomas Douillard <thomas.douillard@gmail.com> wrote:
Hi Gerard, I mostly agree with you, just something : saying "we are not very good to know what an item is" is not really constructive in this usecase : if we want to add edit capabilities to reasonator, we got a end user that knows what he want to edit and who knows what the item he wants to edit with, so he'll make the identification and the edition at the same time, there is no need to "first" identify all the items (it couls even be counter productive as just identifiying the items has limited immediate use so it might not be really a good motivational for community as a whole.)

Tom²


2013/11/21 Gerard Meijssen <gerard.meijssen@gmail.com>
Hoi,

At this time the Reasonator has specific functionality for people, organisms and places.there are many ways it can be improved. It can even become a layer on top of Wikidata with proper caching and all the works that make it perform for the world. Currently it runs on the toolserver..

The most important upgrade from my perspective is that it becomes easy to add missing labels for existing statements in any and all languages. Consider that only XX % of all the Wikidata items have a label in English.

I do like the idea of getting suggestions about what statements to add. However, it starts with identifying something for what it is. That is something we are not good at. At this stage over 50% of all our items have no statements and consequently for these items this approach will not work.

The most potent additions from a Reasonator point of view is adding visual things like signatures, flags, coat of arms, and COMMONS CATEGORIES. Particularly the Commons categories are powerful because they already provide you with search capabilities for images in "your" language. Yes, there will be a big Wikidata-Commons project but this functionality we already have.

So yes, the Reasonator code is GPL licensed. so go wild and do great things.
Thanks,
      GerardM


On 21 November 2013 12:41, Thomas Douillard <thomas.douillard@gmail.com> wrote:
Hi all, speaking of reasonator, I think it is a potential solution if it does not leave only by the hand of Magnus (who is very productive but I imagine has a very long schedule ;) and that we start to extend it colaboratively.

What I would like to see in it or in a gadget based on it :
* More models, useful for obvious reasons
* Edit capabilities : there is one or two messages on Wikidata starting to ask "what property should I use", I personnaly would very happy to answer them "install the reasonator gadget and see what it suggests to you for this kind of items !" with a good chance it actually helps.


Tom²

_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l



_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l



_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l



_______________________________________________
Wikidata-l mailing list
Wikidata-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-l