FOSDEM

[JimMadge]

I'm speaking at the FOSDEM community devroom about the new maintainer team and our experience getting up to speed with the project.

I would like to acknowledge everyone who helped but also respect everyone wishes about share their names or photographs. If everyone could tell me if and how they would like to be acknowledged I would really appreciate it. If you'd like to say what name (real or alias) and what picture (if any) you'd like in the slides that would be great. Also, if you'd like to give me a pronunciation of your name I'll try my best to get it right 😅.

@lwasser @Berkmann18 @JoshuaKGoldberg @flpm @gr2m @da5nsy

Also please tag anyone who I might have forgotten, or comment if you feel I have overlooked you 🙏.

[da5nsy]

Awesome that you're doing this @JimMadge! 🎉

I'd be happy to be acknowledged (although I think my role here has been small so far).
Name: Danny Garside, or @da5nsy on Github (and most other things). Picture.

[flpm]

I am fine with it too. Thanks @JimMadge 🥳 You can use my github picture. The only thing for me is that since my contributions are personal and not part of my work I can't use my employer's name. You can find some more info here if you need https://flpm.dev

Felipe is pronounced as "fih-LEE-pay"

[JoshuaKGoldberg]

Ah it's so exciting you're speaking there! I've heard great things 🙂

You're welcome to use any of my public assets & information (I speak at conferences a good bit myself).

• Name: Josh Goldberg (common North American pronounciation)
• Links: https://www.joshuakgoldberg.com (primary); those linked on that page (secondary)
• Photo: https://www.joshuakgoldberg.com/img/josh.jpg
• Acknowledgements: you can say maintainer if you want, but I'll note I've done little beyond triage issues and give input this year. If you mark a distinction between tiers, I'd be on a less-involved tier.

[JimMadge]

My personal belief, and the spirit of this project, says, all contributions are valid and there isn't much benefit in trying to measure the size or significance of them 👍.

[lwasser]

Yay - happy to be included as well @JimMadge !!! That's awesome. One day I'll make it to that meeting!!

[JimMadge]

It is a great time. I'm sure you would feel very at home in the open research and Python devrooms 🚀.

[Berkmann18]

I'm happy to be included.
Name: Maximilian Berkmann
If you want to use a picture then my LinkedIn works: https://www.linkedin.com/in/mberkmann/

[JimMadge]

Is there another place, or could you email it to me? I can't see the profile without an account 😅

[JimMadge]

Working on the slides here.

Preview here.

[JimMadge]

Thanks for the extra context 🙏.

I absolutely didn't mean to call you out, so sorry if it felt like that.
What I really want to say is the "website has lost translations", so being honest that there is a regression with the site. That isn't the same as saying "we lost/deleted the translations" though, so I can look to clarify that.

My assumption (so please tell me if I'm wrong) is roughly,

• The translations are chunks of text from the site, translated from English to another language
• We have done some restructuring of the site, and change the content in the move from Docusaurus to Astro
• Therefore the translations don't necessarily match up with the English site anymore
• There is some work to be done to move the chunks and check they are still correct

Is that right?

[JimMadge]

Actually... looking at Crowdin, it seems to have done a pretty good job at managing with the changes in the repo like file.md --> /some_directory/file.mdx 🎉.

[JimMadge]

On your other point. I'm not really sure what the best idea with translations are. One one hand I think, "share everything you have translated, even if it is incomplete". On the other, "It isn't kind to say we have an X translation when it is only a few lines". 🤔

If it helps us get translations back sooner, I would say publish all of them, regardless of how complete they are.

[flpm]

Oh, I did not take it that way! I was feeling bad for not working more to get it back in working order 😅

The way translations (almost always) work is using a process defined a long time ago by the GNU gettext app used to localize GNU software.

Strings are extracted from a source code file and organized as a "database" of strings (normally stored in a set of text files with the .po extension) in folders corresponding to each language.

These strings are then translated inside those files. At build time (or just before), the original language strings are replaced by their translations.

Interpreted languages like Python make the string substitution in run time.

Extraction from code is often hard, developers need to "mark" which strings should be translate and which should not. But in our case it is markdown, so all strings are extracted to be translated.

Astro (and Docusaurus) handle translations by expecting a set of folders with the Markdown translated. So we need to use a tool to handle all the gettext process of extraction and reconstruction before the build. That is what CrowdIn does (plus it allow translation and management via the site).

The great advantage of using the whole gettext process instead of just manually translating the markdown directly is that if you translate the file by hand then change the layout of the English file (e.g., switch the order of sections) you need to manually make that change in each language. With gettext the layout does not matter, translated strings are inserted where they need to be no matter the layout of the original English file.

So as long as we don't lose the translated string files we should be able to reconstruct the process.

What would cause you to lose translations is decay, as the English text changes the translations slowly become obsolete, more and more strings don't get replaced and without maintenance eventually all is in English again.

[JimMadge]

Ah, of course. I have done a little with .po files before, reusing strings wherever they appear. I had imagined that for a project that is mostly prose the process was somehow different. Like we translated whole files rather than reusable strings 🎉.

[lwasser]

Hi Jim!! i hope your talk is coming along!! lemme know if i can be more helpful than cheering you on via emojis 🙃
pssss we have more accounts to hunt down ... i just lit up your notifications 😆

[JimMadge]

I think it is going well. Always happy for comments/feedback.

pssss we have more accounts to hunt down ... i just lit up your notifications 😆

Yes, that is going in the talk as part of "what haven't we fixed, what is next" and the lesson of "set up teams/large bus factor from the beginning"

[lwasser]

Yes please!! I keep finding new accounts and tools - the app has a list of them, none of which we have access to, but many of which we may not need? THIS is the funny part, too, and I have no idea which things we NEED vs. what is just extra... so documentation as you go, and then maybe teams for who has access to what things. And then, finally, don't forget to UPDATE UPDATE UPDATE.

Also similar to the documentation, but maybe more dangerous... It turns out the tests in the app used dated dependencies that also had some pretty significant security issues.

If I can be helpful to you in creating the talk, please ping me !! ✨