Tails

#2 Updated by intrigeri 2019-12-15 09:26:38

I don’t need to be involved further in this conversation personally, just my 2cts:

I haven’t read the entire recent discussions on -l10n about it, but I was under the impression that folks there were complaining about the other CI results we’re already sending there, mostly because of:

• the noise it creates for many people who can’t do anything about the problem;
• the fact the people who can fix the problem may not be on that mailing list.

So sending more CI results there has potential to cause more such annoyance and CI alert fatigue.

Maybe we should take a step back, re-read those conversations, and figure out a good short-term + long-term strategy to give feedback to translators.
It could be that what you’re proposing here will be part of the short-term strategy — I don’t know :)

#3 Updated by hefee 2019-12-16 18:44:15

intrigeri wrote:
> I don’t need to be involved further in this conversation personally, just my 2cts:
>
> I haven’t read the entire recent discussions on -l10n about it, but I was under the impression that folks there were complaining about the other CI results we’re already sending there, mostly because of:
>
> * the noise it creates for many people who can’t do anything about the problem;
> * the fact the people who can fix the problem may not be on that mailing list.
>
> So sending more CI results there has potential to cause more such annoyance and CI alert fatigue.
>
> Maybe we should take a step back, re-read those conversations, and figure out a good short-term + long-term strategy to give feedback to translators. It could be that what you’re proposing here will be part of the short-term strategy — I don’t know :)

Well but for sanity-check-website people with a editor status can do something about and it is only triggered once a day. And can be solved, be removing or updating the problematic suggestion. That’s why I think tails-l10n would be a better audience. At least for me it is work of the Tranlation Platform maintainer to fix those suggestions.

General I’m on your page, that we really should rethink who is informed when, if CI is not happy, but maybe can can solve this issue when Feature #17358 get fixed.

#4 Updated by intrigeri 2019-12-17 12:15:31

Hi,

> Well but for sanity-check-website people with a editor status can do something about and it is only triggered once a day. And can be solved, be removing or updating the problematic suggestion.

Assuming you mean s/editor/reviewer/ (with the terminology we’re using in our design doc): that’s a good point.
Then, perhaps we should email these people specifically, instead of all -l10n subscribers. I guess it would not be too hard to have a cronjob keep an email alias up-to-date, pointing it to the list of reviewers extracted from the DB. But perhaps I’m over-optimistic :)

> At least for me it is work of the Tranlation Platform maintainer to fix those suggestions.

I don’t understand and this seems to contradict what you’re advocating for above.
Did you mean it is not the work of the Tranlation Platform maintainer? I would agree with that.

#5 Updated by hefee 2019-12-17 13:47:53

Hi,

> > Well but for sanity-check-website people with a editor status can do something about and it is only triggered once a day. And can be solved, be removing or updating the problematic suggestion.
>
> Assuming you mean s/editor/reviewer/ (with the terminology we’re using in our design doc): that’s a good point.

yep.

> Then, perhaps we should email these people specifically, instead of all -l10n subscribers. I guess it would not be too hard to have a cronjob keep an email alias up-to-date, pointing it to the list of reviewers extracted from the DB. But perhaps I’m over-optimistic :)

Just to make sure, that we are looking at every solution, we can also send mails to the authors of the suggestions. I can get the list of authors of the suggestions is easy. Just to have some numbers at the moment there are ~1900 suggestions, done by 37 persons, but I can’t tell who was doing the mistake and if a suggestion is not accepted for weeks, you will get this notice for weeks for things where you are not the source of. I think this is not a good way, as people adding the suggestion may be over strained by automatic failure mail, as this may be the first interaction with the Tails community.

That’s why I prefer to send this explicitly to the reviewers:

To get the list of reviewers email addresses is easy:

<code class="python">
from weblate.auth.models import Group
reviewers = Group.objects.get(name="Reviewers")
for i in reviewers.user_set.all():
    print(i.email)
</code>

> > At least for me it is work of the Translation Platform maintainer to fix those suggestions.
>
> I don’t understand and this seems to contradict what you’re advocating for above.
> Did you mean it is not the work of the Translation Platform maintainer? I would agree with that.

ACK - I meant not.