Hey everyone,
On Thu, Aug 19, 2021 at 9:42 PM Ed Summers <ehs(a)pobox.com> wrote:
First, I just wanted to say it is *awesome* to see
this level of
transparency and clarity about the state of the service.
+1. Really glad that
we're hearing about this now, before the problem
becomes so seriously that measures are taken without informing the
community beforehand.
2. The Query service is open to the public, with no
authentication,
which means the Wikidata team have very little idea who/what depends on
the service.
For 2 I wonder if it might make sense to start requiring registration &
authentication? Or is this not the wiki way?
Right now, the Wikimedia Commons Query
Service (in beta) is doing
something like that. You need to login and authenticate before you can
use it. I'm not sure if there have been any insights on the points you
mention (who uses it in what capacity). But if i'm reading Mike's
original email i think the main problem is inputting the ever growing
graph of data into Blazegraph, not people (ab)using the query service
in such a capacity that it doesn't scale. I understand that perhaps it
would make sense to add some kind of authentication layer, but it
would be a shame. The Wikimedia projects API's (i'm sharing the query
service among those) are one of the very few large-scale API's with
useful data that don't need any kind of authentication. Making it
really useful to teach to students or people just dabbling in writing
software.
Kind regards,
-- Hay / Husky