Grants talk:Project/Rapid/SD0001/Twinkle localisation

Some feedback

Looks great! Fantastic initiative, SD0001. Hopefully Twinkle finally gets i18n/l10n done right.

Morebits.wiki was written in 2011 using what was at the time a reasonably modern approach put in place by the long-since-vanished UncleDouggie, but I agree with your sentiment that it needs an upgrade. However, tinkering with this in any substantial way is a risk given how fundamentally it underpins Twinkle's functionality. The grant proposal could probably use some more detail on how you plan to mitigate this risk (e.g. build a new set of functions that use mw.Api and promises from scratch alongside the existing Morebits.wiki.* components, which remain untouched).

Also, just to be clear, it sounds like you plan to work on the following modules:

• First month
  • Fluff
  • Diff
  • Batch modules
  • Tag
  • XFD
• Second month
  • Protect
  • Warn
  • Block

Protect and ARV are real mongrels; I think it would be unwise to expect to deal with all these modules in the timeframe. I might suggest tackling CSD, a more mission-critical module for other wikis, before unravelling Protect. This, that and the other (talk) 05:59, 10 October 2020 (UTC)

It sounds nice also. We also need to build a multilingual module that can use for many different projects. Alphama (talk) 09:27, 10 October 2020 (UTC)

Thanks for the feedback This, that and the other. While Morebits.wiki.* is indeed very sensitive, I have gained quite a bit of experience in working on changes in Morebits, including some big PRs like #911 (though this hasn't been merged and would be redundant to the rewrite). Earlier this year, I also wrote a bot framework in JavaScript which does a lot of things like Morebits, but with code that's a lot simpler.

In any case, since it isn't possible to switch over all modules to the new library in one go, I plan to leave the existing Morebits untouched -- though it means that for some time we'll have to load two Morebits-like libraries, until all modules have been switched to the new version.

Regarding the modules, yes I indeed plan to tackle CSD -- inadvertently missed that earlier. It would probably be the first thing I do in the 2nd month. Protect is lower down on the priority list. SD0001 (talk) 17:10, 10 October 2020 (UTC)

Happy to endorse in that case! This, that and the other (talk) 23:34, 10 October 2020 (UTC)

WMF Comments

Hello SD0001, thank you for taking time to work on this proposal for developing the framework for Twinkle localization to different wikis and for answering the community feedback over the discussion page. We have reviewed your proposal and have the following follow up questions and comments:

Can you clarify how many hours you expect to work over this project. In compensation grants, we usually expect the grantees to report an overview of their hours. Can you break the hours expected in the modules you have mentioned in the talk page. Looking forward to your response. Best regards. RSharma (WMF) (talk) 12:56, 29 October 2020 (UTC)

@RSharma (WMF): I would estimate about 600 hours – 10 hours a day for 2 months. If I'm unable to make space for 10 hours/day in my schedule (which does include a couple of other small activities but not much) or feel a need to take a break in the middle, the project may extend into January. I would think a tentative break-up would be as follows:

• Morebits library: 5 days (in addition to the time I have already spent on this)
• Tag module: 4 days
• XFD module: 1 week
• CSD (speedy) module: 1 week
• RPP (protect) module: 1 week
• Block: 3 days
• Warn: 4 days
• Fluff/Diff: 2 days
• Batch modules: 1 day
• Thinking about the overall infrastructure and setting up internationalization of messages: 1 week
• Buffer: 2 weeks. Of course a lot of things above may not go well as planned given the complexity of the project. This buffer time will be used to revisit tasks which could not be completed in the above-mentioned timeframes. If I have time left, then I will spend this time working on the less important modules like Welcome, SharedIP, talkback, etc.

SD0001 (talk) 12:58, 30 October 2020 (UTC)

@RSharma (WMF): Any status update on this? No rush. Thanks, SD0001 (talk) 07:02, 6 November 2020 (UTC)

Approved

Hello HSD0001, thanks for sharing the additional context. I am happy to approve your grant request. Thank you for your contributions to the movement. I would request you to push back your grant start date to November 10 so we have enough time to process the grant. Regards, RSharma (WMF) (talk) 12:05, 9 November 2020 (UTC)

Requesting extension

Requesting an extension of the original grant period by two months.

The work is pretty much coming off very well. The most difficult phase of the project – figuring out a good architecture for the project – is over. Since localisation of a highly wiki-specific gadget like Twinkle has never been done before, there weren't any models to replicate or take inspiration from. The work on adding a string i18n library (#1) is pending but easy now that the plan is in place.

The start of the project was delayed by 10 days per above section, and afterwards it took me quite a bit to properly begin (end of November) due to some COVID-induced real-life uncertainties. Ideally, I didn't intend for this project to go on into January – I have just started a new job earlier this week. So from now on I have to considerably slow down due to lesser time available. Under these circumstances, a 2-month extension would be perfect to finish the project. A shorter extension would also probably be fine, but since it isn't clear to me how much time I will have alongside my job, a 2-month extension is a safer option.

Thanks, SD0001 (talk) 16:49, 7 January 2021 (UTC)

Hello SD0001, thanks for posting your extension request. It is approved and the project end date is extended by two months. The new grant end date is March 10, 2021, with the reporting due date on April 10, 2021. Best regards, DSaroyan (WMF) (talk) 12:09, 8 January 2021 (UTC)