22:13:11 <ianychoi> #startmeeting openstack_i18n_meeting
22:13:12 <openstack> Meeting started Thu May  2 22:13:11 2019 UTC and is due to finish in 60 minutes.  The chair is ianychoi. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:13:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
22:13:15 <openstack> The meeting name has been set to 'openstack_i18n_meeting'
22:13:19 <ianychoi> Let's do :)
22:13:30 <eumel8> aeh ... jepp, had the same idea ;)
22:13:35 <ianychoi> #topic Weblate
22:14:09 <ianychoi> #link http://lists.openstack.org/pipermail/openstack-i18n/2019-April/003397.html
22:14:35 <ianychoi> #link https://www.redhat.com/archives/zanata-users/2019-April/msg00014.html
22:15:08 <ianychoi> Although there were two replies from Red Hat recently, it definitely seems that Zanata's health status is not good..
22:16:05 <ianychoi> I heard from ianw that Fedora is investigating with Weblate for translation platform, and from Andreas that some Suse's translation parts are using Weblate
22:16:31 <ianychoi> (not directly, Clark shared)
22:17:41 <ianychoi> Yesterday, I downloaded translation memory from Zanata: File name is "zanata-allProjects-allVersions-en-US-allLocales.tmx" and the size is about 127 MB.
22:18:11 <ianychoi> There might be several considerations for Weblate migration, but I think:
22:18:44 <ianychoi> - infra side: installation of Weblate instance (can we use Weblate's product?)
22:19:02 <ianychoi> - infra side: syncing with Gerrit, and git
22:19:19 <ianychoi> - I18n side: is translation memory migration possible?
22:19:34 <ianychoi> - i18n side: would usage be easy or difficult?
22:20:14 <ianychoi> eumel8: Any other consideration from your research? amotoki: your thoughts?
22:20:40 <eumel8> can you add https://etherpad.openstack.org/p/docs-i18n-ptg-train ?
22:21:14 <eumel8> clarkb wants to ask for hosting
22:21:31 <eumel8> but I'm not sure. We need a place to host pot files
22:23:06 <ianychoi> eumel8, Andrea's idea is that POT is just a kind of intermediate - what I mean, PO files are actually more important from my understanding, and POT file would be a kind of intermediate files. Would you share more why and how posting pot files would be needed in which step in Weblate?
22:23:41 <ianychoi> eumel8, QQ: https://etherpad.openstack.org/p/docs-i18n-ptg-train or https://etherpad.openstack.org/p/I18n-weblate-migration ? :)
22:24:10 <ianychoi> Will add https://etherpad.openstack.org/p/docs-i18n-ptg-train first
22:24:24 <eumel8> sounds good
22:24:46 <eumel8> translation memory is supported in any way: https://docs.weblate.org/en/latest/admin/memory.html
22:25:09 <ianychoi> And openid authentication from infra side
22:25:45 <eumel8> yeah, want's to talk with Jimmy, but no appointment so far
22:26:02 <eumel8> Weblate supports openid but not openstackid auth
22:26:16 <ianychoi> Then it would be fine IMO
22:26:25 <eumel8> https://github.com/python-social-auth/social-core/pull/359
22:26:28 <ianychoi> Zanata also just supports openid authentication
22:27:13 <eumel8> lets go through the list: https://etherpad.openstack.org/p/I18n-weblate-migration
22:27:42 <ianychoi> Sorry what is python-social-auth/social-core?
22:29:31 <eumel8> thats the tool that Weblate uses for OpenID
22:29:54 <ianychoi> Aha thanks
22:30:04 <eumel8> there are many backends so far: https://github.com/python-social-auth/social-core/tree/master/social_core/backends
22:30:54 <ianychoi> +1 :)
22:30:54 <eumel8> with that patch we can configure openstackid and openstackid-dev as backend
22:31:30 <eumel8> there are some questions for me about username so Jimmy with his web guys should take a look
22:31:44 <eumel8> second thing is the openstack logo inside
22:31:55 <eumel8> I addressed both
22:33:19 <ianychoi> Okay - then let's ask for inviting Jimmy and an openstackid developer (maybe Sebastian) to IRC meeting or asking via e-mail. Would it be fine?
22:34:02 <ianychoi> #action Weblate-openstackid/openid follow-up discussion with I18n team and openstackid developers
22:34:32 <ianychoi> amotoki, any other comments on Weblate migration from your perspective?
22:34:48 <amotoki> sorry, I am in the neutron session.
22:35:05 <ianychoi> Fine fine I am in home :(
22:35:49 <ianychoi> #topic Train priority
22:35:50 <eumel8> I'm in #208 ;)
22:36:06 <ianychoi> #link https://review.opendev.org/#/c/656047/
22:36:47 <ianychoi> I agree with Andreas - if there are not too much translation on stable branches, I think not extending to more stable branches would be fine. eumel8 your thoughts?
22:37:42 <ianychoi> Also, how about creating another version "releasenotes" to only manage release notes in Zanata?
22:37:47 <ianychoi> Please see a comment in
22:37:51 <ianychoi> #link https://review.opendev.org/#/c/643085/
22:38:15 <ianychoi> For other things, I wanna set the same priority as we have done previously
22:38:33 <ianychoi> amotoki, quick question: are you still translating Horizon releasenotes?
22:39:09 <amotoki> ianychoi: personally not. Some Japanese translators seems to be translate them.
22:40:04 <eumel8> not sure if we should working on this further, because it's Zanata related
22:40:18 <ianychoi> amotoki, thanks a lot for sharing :)
22:40:20 <amotoki> what is the background of having another branch for relnotes?
22:40:53 <ianychoi> amotoki, the mix of percent in master branch from rsimai: "I have no easy way to see where I can contribute without going down to the source files. I'm OK with any way that allows me to quickly understand, just the way how Zanata presents it is "tainted" with the release notes percentages, for me."
22:42:15 <amotoki> ah, i see. it is about the progress rate in Zanata dashboardm right?
22:42:29 <ianychoi> Yep
22:44:45 <ianychoi> eumel8, how long Weblate migration would take you think? :)
22:45:54 <eumel8> a day 2 operation. Depends on the architecture for syncs
22:46:29 <eumel8> interesting point is to work without pot files. is this possible?
22:46:58 <ianychoi> If we can do within Train cycle, then I think not touching anymore on Zanata would be fine
22:47:20 <eumel8> that's the plan, I think
22:47:55 <eumel8> we have only 2-3 lower issues to fix
22:49:07 <ianychoi> #topic I18n sync flow clarification
22:49:48 <ianychoi> eumel8, doesn't weblate support po files? :)
22:50:18 <ianychoi> #link https://docs.weblate.org/en/latest/formats.html
22:50:46 <eumel8> for sure, it's only the topic for updates
22:51:16 <eumel8> thats the point for pot files, I think
22:51:33 <ianychoi> Aha.. so are you mentioning syncing from the source changes on git projects to translation source?
22:52:04 <ianychoi> We can propose that each repo stays pot files then
22:52:20 <eumel8> I'm not sure, if sync of po files from upstream would be enough. What about new projects without po files?
22:52:40 <ianychoi> Just note that no storing pot files in each repo would not be possible without the power of Zanata. Previously, using Transifex, pot files were in project repos
22:53:05 <eumel8> interesting news, thx
22:53:34 <ianychoi> Then one issue were solved? :)
22:54:16 <eumel8> storing pot files in the repo?
22:54:20 <ianychoi> yep
22:54:35 <eumel8> who builds that?
22:54:39 <ianychoi> Infra
22:56:07 <ianychoi> We can ship generating POT files like https://review.opendev.org/#/c/306915/
22:56:18 <ianychoi> (although the patch deletes pot files)
22:57:33 <eumel8> seems like a required zuul job for that
22:59:15 <ianychoi> Using weblate, upstream-translation-update and propose-translation-update jobs would be also needed I thought
22:59:40 <ianychoi> Such generating POT files can be accomplished within upstream-translation-update job
23:00:59 <eumel8> It would be nice if we can skip that jobs. I'm not sure if it required.
23:01:29 <ianychoi> eumel8, how about doing POT update daily with  upstream-translation-update ?
23:01:52 <ianychoi> For other stuff, if possible, leaving would be fine :)
23:02:00 <eumel8> Weblate has the possibility to propose things in Gerrit and receive updates with Webhook
23:02:06 <adiantum> hi all
23:02:11 <ianychoi> adiantum, hi!
23:02:43 <ianychoi> eumel8, hmm, then such job should execute after every source change on Gerrit/Git ?
23:02:56 <ianychoi> If it is fine, then I am okay :)
23:03:33 <ianychoi> (what I mean, if it very simplifies I18n work and does not require too much overhead)
23:03:48 <ianychoi> I am telling that we have solution on POT file generation :)
23:04:07 <ianychoi> eumel8, any more clarification is needed now? :)
23:05:14 <ianychoi> (I want to move to open discussion not to much more exceed meeting time slots, although I am fine with time now)
23:05:30 <ianychoi> adiantum, are you in Denver? :)
23:05:54 <adiantum> ianychoi, unfortunately, no
23:06:21 <eumel8> ianychoi: I'm fine, thx
23:06:29 <ianychoi> #topic open discussion
23:06:38 <ianychoi> adiantum, neither do I :(
23:06:59 <adiantum> ianychoi, heh
23:07:09 <ianychoi> Really sorry for this Summit + PTG but I think we can meet on the next Summit + PTG :)
23:07:18 <ianychoi> in China!
23:07:25 <adiantum> =)
23:07:29 <ianychoi> I have never been to China, although it is not too far
23:08:26 <ianychoi> eumel8, if it is possible, then staying I18n team status as an official team by this year, and discussing I18n team's future in China would be good
23:08:46 <adiantum> ianychoi, small question, just for clarification: why are we planning to get rid of Zanata?
23:09:09 <eumel8> ianychoi: sure
23:09:12 <ianychoi> adiantum, there is no more Zanata open source development from September last year
23:09:28 <adiantum> wow
23:09:44 <eumel8> ianychoi: I listed the issues with the SIG on the Etherpad
23:09:49 <ianychoi> adiantum, active OpenStack folks are finding alternatives for our bright future. If there is active Zanata development, we wanna stay as-is status :)
23:09:50 <eumel8> what I see so far
23:10:24 <adiantum> eumel8, are you in Denver?
23:10:36 <eumel8> yes, I'm in Denver
23:10:46 <ianychoi> eumel8, now, just acknowledging how to change team status from official to sig and identifying how such change impacts in OpenStack world would be fine IMO
23:11:29 <ianychoi> I more think that final decision can be made from I18n team, since we still do translation activities and are actively finding alternative translation platform
23:13:18 <eumel8> yes, I also love the project updates to bring the work to the light. SIGs are not so visible
23:13:25 <adiantum> ianychoi, eumel8, and one more question: why we’re going to change from team to sig?
23:13:35 <ianychoi> eumel8, just I am surprised that Zanata translation memory "zanata-allProjects-allVersions-en-US-allLocales.tmx" is not so big - 127 MB
23:14:02 <eumel8> adiantum: organization overhead, like PTL election, team meetings etc.
23:14:16 <ianychoi> adiantum, lack of regular meeting rhythm, much less translation activities, and PTL election
23:14:37 <adiantum> ok
23:14:49 <eumel8> ianychoi: I don't care so many about translation memory. I would bring more machine translations into.
23:15:40 <ianychoi> eumel8, I have another view - still manual translation results can be basement of machine translation data, and I strongly think that translation memory itself can be open source
23:16:04 <ianychoi> OpenStack Foundation can announce that OpenStack translation memory is available as open source
23:16:17 <ianychoi> for example
23:17:02 <ianychoi> Currently, K8s has L10n team, Ceph team also have translation team. Fedora , Suse, many organizations have translation team.