Currently, the primary sources tool maintenance and improvement
processes are fairly sub-optimal, as:
1. the core team at Google has not enough time to tackle bugs and
2. the pull request/merge flow is insufficient alone, since both the
back-end and the front-end must also be deployed in production,
eventually by someone else.
I would like to report here the requirements to make the tool sustainable:
A. a developer to understand the back-end code , written in C++;
C. access to the WMF Labs machine to deploy the back-end ;
D. a Wikidata administrator to deploy the front-end ;
E. centralized and exhaustive documentation.
As part of the StrepHit project goals , my team is striving to help
with A. (not exactly trivial) and C., but we really need B. and D. to be
Date: Tue, 31 May 2016 08:47:18 +0000 From: Sebastian
<schaffert(a)google.com> To: Thomas Steiner <tomac(a)google.com>om>,
"Discussion list for the Wikidata project."
<wikidata(a)lists.wikimedia.org> Subject: Re: [Wikidata] Primary sources
tool "reject claim" broken? Message-ID:
Content-Type: text/plain; charset="utf-8" Hi Thomas and all, there might
be a caching issue here. That part of the code is here:
and it still seems right to me, but I'll check again. I won't have much
time in the next days though :( I'll give it one hour no, maybe I
discover something. Cheers, Sebastian On Tue, May 31, 2016 at 9:17 AM
Thomas Steiner <tomac(a)google.com> wrote:
> >Hi Markus and Marco, all,
> >Thanks for your support of and caring for the Primary Sources Tool.
> >Please find my replies inline.
>> > >Dear <whoever is maintaining the PS gadget now>,
> >I guess the core team still does, with the caveat explained by Denny
> >in , the tl;dr is that we work on it on top of our regular jobs and
> >that we are happy to hand it over to folks with more time on their
>> > >The PS tool seems to break more and more. Besides the persisting issue
>> > >duplicated claims being offered (even if they are already stored),
>> > >now also the issue that claims cannot be rejected. If I reject a claim,
>> > >page reloads, but the suggestion still shows up after that.
> >I checked both problems. It seems the writes from the front-end
> >somehow do not make it to the back-end. I opened a random item Q632229
> >and approved and rejected claims. The approval went through just fine
> >(i) However, I could reproduce the duplicate claims being shown, the
> >reason is that the uniqueness comparison does not take references into
> >account , a known @ToDo up for grabs.
> >(ii) I could also in some cases reproduce the non-rejectable claims
> >issue. I repeated disapproved statement 868483 , but if you query
> >the back-end for incoming Freebase statements for Q632229, it keeps
> >coming back as "unapproved"  (search for "868483").
> >For (i), if someone wants to tackle this, happy to merge their Pull
> >Request. For (ii), Sebastian, do you have a suspicion why this could
> >be the case?
> >Dr. Thomas Steiner, Employee (http://blog.tomayac.com
> >Google Germany GmbH, ABC-Str. 19, 20354 Hamburg, Germany
> >Managing Directors: Matthew Scott Sucherman, Paul Terence Manicle
> >Registration office and registration number: Hamburg, HRB 86891
> >-----BEGIN PGP SIGNATURE-----
> >Version: GnuPG v2.0.29 (GNU/Linux)
> >-----END PGP SIGNATURE-----