Hello all,

Thanks for the discussion! I will give my future MnM feedback on Magnus' Wikidata page, as suggested.

Best,
Tuomas / National Library of Finland

Lähettäjä: Wikidata <wikidata-bounces@lists.wikimedia.org> käyttäjän Magnus Manske <magnusmanske@googlemail.com> puolesta
Lähetetty: perjantai 8. toukokuuta 2020 17.10
Vastaanottaja: Discussion list for the Wikidata project <wikidata@lists.wikimedia.org>
Aihe: Re: [Wikidata] Questions about Mix'n'match tool
 
These days, all Mix'n'match edits should contain a backlink to the Mix'n'match entry in the edit summary (I think).

MnM will never "overwrite" human edits. It might add a second statement for a property to an item, but for some catalogs/properties that's a valid thing to do, so I'm hesitant to block such an option.

Also, the value on Wikidata might be wrong, and the MnM edit is correct. In which case, the constraint violation will highlight the issue, and it can be resolved on both Wikidata and MnM.

In related news, I have been collecting issues between Wikidata and MnM, in case someone wants to help resolve them (some might be resolved already):
https://tools.wmflabs.org/mix-n-match/#/issues


On Thu, May 7, 2020 at 7:16 PM Tom Morris <tfmorris@gmail.com> wrote:
Speaking of workflow, this Mix'n'Match report page says "If you fix something from this list on Wikidata, please fix it on Mix'n'match as well, if applicable" without giving any directions or hints as to how one might accomplish that. I just cleaned up "Multiple Wikidata items with external ID OL4859603A : Pierre Vidal (Q18002076)Pierre Vidal (Q3387281)".

How would I tell Mix'n'Match about that? Or, better yet, why can't the tool be taught not to overwrite human edits (particularly reverts)? To be fair, the original confusion is understandable, because the two entries/gentlemen are very confusable.

Tom

On Wed, May 6, 2020 at 4:28 AM Magnus Manske <magnusmanske@googlemail.com> wrote:
Hi,

I am the author of Mix'n'match, so I hope I can answer your questions.

Match mode:
By default, "match mode" only shows unmatched entries, example:

You can force pre-matched entries, but currently they won't show the automatic predictions:

If you/others like it, I can have predictions show for auto-matched as well, and/or mix unmatched and pre-matched in results.

"Next entry" (aka skip) will never change the bucket (pre-matched/unmatched).


Mobile game:
The mobile game shows unmatched only. There is currently no override.
This would be easy to change as well, if desired, though I'd probably have to show/highlight the pre-matched entry somehow.

"Skip" will never change the bucket (pre-matched/unmatched).


Visual tool:
The visual tool will show entries from both pre-matched and unmatched.

"Load another one" (aka "skip") will never change the bucket (pre-matched/unmatched).

I hope that answers your questions, please let me know if there is anything else I can do!

Cheers,
Magnus

On Wed, May 6, 2020 at 7:26 AM Palonen, Tuomas Y E <tuomas.palonen@helsinki.fi> wrote:
Hello,

I am an information specialist at the National Library of Finland, where I am linking our General Finnish Ontology YSO (30,000+ concepts with terms in Finnish-Swedish-English) to Wikidata at the moment. I just joined this mailing list. I am currently using Mix'n'match tool and would have a couple of questions. I would be very happy for any answers or contact info to someone who might have the answers. Thank you very much! Here are my questions:
  • Is there a connection in Mix'n'match between the Preliminarily matched/Unmatched division and any of the three: Match mode, Mobile matching, Visual tool?
  • More precisely, are the link suggestions provided by Match mode / Mobile matching / Visual tool created only from Preliminary matched list, only from Unmatched list or from both?
  • Also, if I reject a link suggestion in Match mode / Mobile matching / Visual tool, will that concept be added to the Unmatched list?
This would be important for my own (and possibly anybody else's) Mix'n'match work flow / method. If it was up to me, I would suggest rejected concepts not be added to the Unmatched list, but that's just me. I am planning to write a report of my linking project in the future, where I will most likely include some development suggestions / wishes. One immediate wish is that it would be great to tag or somehow put aside link suggestions that require more research and cannot be decided on the spot (for example, Wikidata items may include poor Finnish/Swedish terms and may require some corrections before I can do the linking).

Thanks for your help!
Best regards,
Tuomas / National Library of Finland
_______________________________________________
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