Forgot to say why I wanted to do this. I never quite get it, other people don't live in my head…

It is not hard for the developer to make a system whereby the user can select properties and values, but that puts the burden on the user. It is hard to acquire knowledge on how to create a correct query. Instead the system should make the difficult task of choosing the optimum question, and the user should only chose the one that is simple to answer.

On Wed, Jun 21, 2017 at 7:04 PM, John Erling Blad <jeblad@gmail.com> wrote:
I've been toying with a an idea about a dialog system where a search in Wikipedia could be analyzed for properties that optimally divides the result set by simple questions. Typically you will have a yes/no question on whether you look for a human if about half the results are instances of humans. Because it is not obvious which questions the user might be able to answer a small number of questions are posed simultaneously. After an answer is given it is not given that the remaining questions from the previous iteration are optimum, thus all of them must be reevaluated.

It forms a decision graph, and is not that unique. The only thing new is to use a semantic web to create the graph, and thus reduce the problem to a decision table.

The dialog system would be a small number of alternate questions, and each group would scale with the number of defined properties. It would not be of linear order, as some language needs alternate prose due to congruence (agreement).

My initial notes are pretty old, they are from 2010, long before Wikidata. Back then it seemed to be possible to use the categories, but with Wikidata it is much simpler to simply to build the necessary graph.

The initial problem posed at nowiki was "how to find article about a blue flower".

On Wed, Jun 21, 2017 at 2:38 PM, Jonas Kress <jonas.kress@wikimedia.de> wrote:
Hi,

that is a very nice tool! I really like that it is quite intuitive to use!

We also have been working on something like this at query.wikidata.org for a while.
It might not be obvious how to use it, so here you will find the documentation [1].

With the latest change we implemented context aware suggestions.
It will be deployed next week, but you can already try it out here [2].

Cheers,
Jonas

Inline-Bild 1



2017-06-21 13:37 GMT+02:00 Shani Evenstein <shani.even@gmail.com>:
Hay, I just published your email in the Facebook Group "The Wikidata Community", and someone has asked this: "can you please ask the author to create a sanitizable URL for queries? I can't quote them in wikitext."

Is that possible?

Thanks, 
Shani. 

On Wed, Jun 21, 2017 at 1:46 PM, Shani Evenstein <shani.even@gmail.com> wrote:
Thanks, Hay. This is awesome!

On Wed, Jun 21, 2017 at 1:38 PM, john cummings <mrjohncummings@gmail.com> wrote:
Muggles everywhere thank you :)

One suggestion which would be helpful for me is to be able to link to the Commons Category. Perhaps different options like this could be tick boxes at the top?

Thanks very much

John



On 21 June 2017 at 11:52, Hay (Husky) <huskyr@gmail.com> wrote:
Hey Thad,
thanks for the suggestions!

Concerning keywords in the description field of properties: i think
that is a shortcoming of the current autocomplete API that i'm using,
which is the same one Wikidata uses for the searchbox in the top-right
corner. Apparently they're not indexing those keywords, which makes it
difficult for me to also include them in my search.

I like the suggestion of using the 'that is' part of the rule as a way
to use filter operations, i'll take that into account when i'll be
adding that.

-- Hay

On Wed, Jun 21, 2017 at 2:13 AM, Thad Guidry <thadguidry@gmail.com> wrote:
> I like the idea of Layers of rules... like Photoshop in a way, and then you
> can probably surface quite a bit more.
>
> The one thing that is really lacking is finding properties that have a
> keyword in their description, and not just the label name.
> For instance, typing "person" in a property would give you all the
> properties that have a label containing "person", AND have a description
> containing the keyword "person".
>
> Also, the "that is" should be a drop down element to use other operations
> such as a "filter" or "containing" operation, etc.
>
> Imagine doing:
> "item" has a property equivalent class P1709  "containing"  "schema"
>
> instead of this:
>
> item
>
> has a property
>
> equivalent class P1709
>
> that is
>
> Schema.org Q3475322
>
>  Remove rule Add rule Remove all rules
> -Thad
> +ThadGuidry
>
> _______________________________________________
> Wikidata mailing list
> Wikidata@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wikidata
>

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


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




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




--
Jonas Kress
Software Developer

Wikimedia Deutschland e. V. | Tempelhofer Ufer 23-24 | 10963 Berlin
Phone: +49 (0)30 219 158 26-0
http://wikimedia.de

Imagine a world, in which every single human being can freely share in the sum of all knowledge. That‘s our commitment.

Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e. V. Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg unter der Nummer 23855 B. Als gemeinnützig anerkannt durch das Finanzamt für Körperschaften I Berlin, Steuernummer 27/681/51985.

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