On 08.05.2015 08:50, Lydia Pintscher wrote:
On Fri, May 8, 2015 at 7:15 AM, Stas Malyshev smalyshev@wikimedia.org wrote:
I am worried that having two different data sets within the same instance would be a problem for tools working with the data, and for humans too. And frankly, I don't see too much benefit - virtually all added value Wikidata has now is working with the assumption of the semantics of Wikidata values and properties. Everything that pertains to lexemes, forms, etc. will have to be built separately, so why do it within the same site and have all the mechanics act as a split brain? I would think having parallel instance of Wikibase would serve the same goal much better, while preserving all the benefits of using the Wikibase toolkit and basic data model. Ultimately, it's the same as having separate databases vs. having one huge database (or even one huge table) with columns marking virtual partitions - the former is much easier to handle if the sets are completely disjoint, as we'd have between Wikidata and Wiktionary, as far as I can see. Maybe I am missing some benefit joint structure would produce?
The benefits of having it in one instance are huge imho. Our community exists and knows how to handle structured data by now. Processes/documentation/etc are set up. The world outside is starting to realize that Wikidata is the place to go to for structured data around Wikimedia now. And we probably do want easy connecting between items/properties/lexems etc. As we're talking about different entity types the data is easy enough to keep apart for those who want to.
+1
Other technical solutions can be found for keeping content apart when needed (e.g., separate dumps by entity types).
Cheers,
Markus