The main reason why Queries are not done yet is because in the beginning of 2013 I deprioritized them compared to the original plan. Only a single developer kept working on them, instead of a major part of the team, as was originally planned.
I made this decision because it became clear to me that we will likely be able to continue the Wikidata development beyond the original 12-month plan (as was indeed the case) and that, in the medium run, rushing this functionality would only hurt the project. I thus decided to increase the priorities on tasks which had a higher short-term benefit and were more immediate, e.g. many smaller things, but also more datatypes, ranks, and clean-ups, but also reactions to the roll-outs which had begun back then. This made us highly responsive to the current needs of the community, and lead to a sustained growth of Wikidata.
If it would be needed, queries could be rushed. But that would have a negative impact on the longer sustainability of the project. If it would be deemed a higher priority, the development of queries could be sped up. But this comes with a sacrifice regarding other functionalities. Thus yes, more resources would lead to a faster development of queries (if it were decided that this would be the appropriate priority).
The latter especially means that a sustained contribution from external developers can also lead to a faster development of the query functionality. We have seen with the sustained support of Benestar for the Badges functionality that this is feasible and possible. So instead of simply expressing complaints about features not being developed fast enough, how about actually helping with making them real? it is Open Source after all. Or at least simply make a case for the importance of this functionality? The development team keeps listening to the community like no other that I know of, and prioritizes their effort with respect to that.
So, in short, blame me.
Cheers, Denny
On Tue, Jan 7, 2014 at 2:08 PM, Jan Kučera kozuch82@gmail.com wrote:
Hm,
nice to read all the reasoning why queries are yet still not possible, but I think we live in 2014 not and not 1914 actually... seems like the problem is too small budget or bad management... can not really think of another reason. How much do you think would it cost to make queries reality for production at Wikidata?
Regards, Jan
2013/11/29 Gerard Meijssen gerard.meijssen@gmail.com
Hoi, Please understand that providing functionality like query is something that has to fit into a continuously live environment. This is an environment where the Wikidata functionality is used all the time and where some of the underlying functionality is changed as well. The Wikidata development is not happening in a vacuum.
Given that we hope to get next week a new type of property, it should be obvious that Wikidata is not feature complete. When you add on extra functionality like a query engine, you add extra complications while the work is ongoing to get to the stage where Wikidata is feature complete for the data types.
Another aspect is that it is NOT the Wikidata team to decide what goes into production on Wikipedia projects. The Ask 1.0 functionality for instance is at its release level. It is now for other people to determine if they want to include it in. They have their own road maps and, it is not obvious for an observer what the rationales are. NB Ask 1.0 is also used in Semantic MediaWiki and it provides a query kind of functionality. Query does require some performance <grin> and what is too much </grin>.
So in one aspect there is a query functionality to be used in Wikipedia ea. What the query functionality will deliver that is still being build is not clear to me.
On another note, there are other projects that have lingering before they were implemented. Nothing new here. There have been other projects that had to change because of external pressures. Nothing new here.
If you want query functionality on the existing data now, there is a hack that works quite nicely. It makes use of data replicated to the labs environment. The replication is broken and given the holidays it has not been picked up for the third day now so the data is three days old. Thanks, GerardM
On 29 November 2013 14:03, Martynas Jusevičius martynas@graphity.orgwrote:
Jan,
my suspicion is that my predictions from last year hold true: it is a far more complex task to design a scalable and performant data model, query language and/or query engine solely for Wikidata than the designers of this project anticipated - unless they did anticipate and now knowingly fail to deliver.
You can check some threads from december last year, and they relate to even older ones: http://www.mail-archive.com/wikidata-l@lists.wikimedia.org/msg01415.html
Martynas
On Fri, Nov 29, 2013 at 1:47 PM, Jan Kučera kozuch82@gmail.com wrote:
Ok.
One is a bit disappointed seeing various projects to fail to deliver according to their original timelines... seems like there is not enough money in? Do you need more developers to perform better?
2013/11/26 Lydia Pintscher lydia.pintscher@wikimedia.de
On Mon, Nov 25, 2013 at 9:55 PM, Jan Kučera kozuch82@gmail.com
wrote:
Hi,
so how things are going? Anything for testing already?
Nothing to test yet. As soon as there is I will send an email to this list. The current status is that we still need to make some final adjustments on the database schema and finish the java script part of the user interface as well as ranks.
Cheers Lydia
-- Lydia Pintscher - http://about.me/lydia.pintscher Product Manager for Wikidata
Wikimedia Deutschland e.V. Obentrautstr. 72 10963 Berlin www.wikimedia.de
Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e.
V.
Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg unter der Nummer 23855 Nz. Als gemeinnützig anerkannt durch das Finanzamt für Körperschaften I Berlin, Steuernummer 27/681/51985.
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
Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l