2.0 alpha release

[marfox]

All the tool API calls now use the back end version 2, deployed in a Wikimedia VPS machine, with base URI: https://pst.wmflabs.org/pst/

This is the last step towards the 2.0 alpha release:

• it is mainly meant to test the new back end;
• the front end remains a Wikidata gadget version;
• the front end integrates currently implemented features:
  • reference preview;
  • property browser sidebar;
  • no duplicate statements.
• the dataset selection dialog now displays additional dataset info

[Kiailandi]

These are surely a lot of commits.

If I understand correctly the main points of this PR are:

• calling the new backend services

• avoid duplicates (d849f54)

• remove the "approve claim", allowing to approve references only. I suppose this has to do with the new backend services (9f3443a)

• a new way of selecting the datasets (332d8c9)

Did I miss something?

[marfox]

Hi @Kiailandi , I think you got the main points of this request.
The current title of this PR is misleading, since it refers to only one point. I'll rename it.

[Kiailandi]

Calling this a new alpha release seems a lot more fitting, thanks for the clarification!

[Tpt]

Thank you very much for the update. Sorry for the late comments. I have one concern with one commit of this PR "remove the "approve claim", allowing to approve references only. I suppose this has to do with the new backend services (9f3443a)"

Having this button was a strong request we had when deploying the first version of Primary Sources. Sometime (especially from Freebase) the claim is good but the reference come from a not authoritative website and, so, the user prefers not to add it. I believe it would be better add it back in order to allow the user to still do a "1 click" addition instead of having to enter it again using the regular UI

[marfox]

Hi @Tpt , here is the main rationale behind the decision of removing the claim curation buttons:

all statements in the tool must be referenced.

This is in line with the general Wikidata data quality goals:

1. to fill the reference gap in existing statements;
2. to ensure that new statements are already referenced.

This will be a hard requirement for future dataset uploads.

As a side note, I believe that the addition of new unreferenced statements is what stalled the migration of Freebase, due to a high amount of blacklisted references.
We should definitely relaunch the discussion on this topic with the community, maybe through a request for comment in Wikidata. Feel free to open it if you want!

Still, I believe it makes sense to implement such hard requirement on references, starting from now.

[Tpt]

We should definitely relaunch the discussion on this topic with the community, maybe through a request for comment in Wikidata.

Yes, definitely. I am quite busy now and you are the main person working on Primary Sources so you are probably in a better position than me to start this discussion.

As a side note, I believe that the addition of new unreferenced statements is what stalled the migration of Freebase, due to a high amount of blacklisted references.

My feeling is that the reasons are more 1) the speed of Primary Sources and the bugs in the UI 2) the lack of good references 3) the important amount of statements that are not false but that are the Freebase way and not the Wikidata way of encoding things.

This will be a hard requirement for future dataset uploads.

Just to check: have you talked with Lydia about it? And a practical question: what about mappings and similar data that are very hard to reference?

[marfox]

We should definitely relaunch the discussion on this topic with the community, maybe through a request for comment in Wikidata.

Yes, definitely. I am quite busy now and you are the main person working on Primary Sources so you are probably in a better position than me to start this discussion.

Fair enough. I can open the RfC, but please keep an eye on it, since you are certainly the one who can answer questions on Freebase.

This will be a hard requirement for future dataset uploads.

Just to check: have you talked with Lydia about it?

I have certainly mentioned this, but it would be best to schedule a call the 3 of us to discuss further.

what about mappings and similar data that are very hard to reference?

I don't have an answer on this. Another reason to chat with Lydia together.