Ivan,

Thanks for the reply. Yes as many people should be brought in about subject indexing as can be found. The UDC is a faceted classification; LCC (Lib of Congress Classification) and DDC (Dewey Decimal Classification) are not. UDC is a multilingual taxonomy (facet=language) and these other two are not. To me there's little to no competition with UDC.

But that's about what taxonomies are deployed -- I support many and all. It's the implementing technology that's key here and happily doensn't require expert panels and study groups. SKOS can handle multiple facets -- via its Collection objects. I view SKOS as a stepping stone to ISO Topic Maps; I suspect SKOS has been invented by the W3 as a response to Topic Map functionality not in OWL, but perhaps you can enlighten me further. Basically I have these questions.

1. Is WP going to be empowered with subject indexes as a core feature?

2. What's the relative priority of subject indexes over other wikidata requirements (ie after Semantic Infoboxes)

3. Should SKOS or a very close equivalent be subsumed into the WOM?

The answer to 3 probably should echo the realtionship between the WOM and the OWL, whatever that is intended to be.

Thanks,

John

Date: Sat, 31 Mar 2012 11:22:43 +0200
From: Ivan Herman <ivan@w3.org>
To: "Discussion list for the Wikidata project."
<wikidata-l@lists.wikimedia.org>
Subject: Re: [Wikidata-l] Wiki Subject Indexes
Message-ID: <F67C291D-CE31-4624-AD87-7B3ABCB9B612@w3.org>
Content-Type: text/plain; charset="us-ascii"

On Mar 31, 2012, at 02:00 , John McClure wrote:

> Hello,
>
> Can/should wiki subject indexes be a functional requirement of the wikidata
> project, or of some other? I think navigating a wiki today without a subject
> index is difficult, to say the least. A subject index seems such a critical
> component for information libraries! WP's portals are a nice step but still,
> users seem at the mercy of topical links inserted by authors of the portal.
> How much better it would be to have a taxonomy of subjects that can be
> associated with a page by ITS author so that the page can be found
> independently of portals.
>

I see a major issue with the user interface of this (though I agree with your assessment). It has to be very easy to set the right subject, otherwise people will not do it.

The dbpedia people extract some rough classification of the articles. It is not perfect, but may be worth looking at that, too.

> The semantics of SKOS, I suggest, should be baked in to wikis. I also
> suggest faceted UDC [1] or similar inter/national classification scheme be
> one of many that can be referenced by users when browsing any wiki. I
> envision that Subjects would be defined in a namespace as fundamental to a
> wiki as the Category namespace is. Basically, I can see requesting some
> software to correlate my own subject taxonomy with interwikis' (WP's)
> tasxonomies, so that I as a user don't have to manually search each
> interwiki for content relative to my personal list of subjects.

I think the project should reach out to libraries, ie, real experts. The library world is currently looking at the issues of how to redefine cataloging standards, how to make them Linked Data friendly, etc; because that work is still in flux, it may be the ideal time to talk to them. In view of the importance of WP, libraries cannot allow themselves to ignore this (I believe...)

Ivan

>
> Is this an idea before its time, something already considered, or something
> to consider now?
> Thanks for your thoughts. -- john
>
> [1] http://en.wikipedia.org/wiki/Universal_Decimal_Classification