The 4 hour intervention of Matt in the international #polyglots channel last Thursday came as a surprise to all of the people connected at that moment. Since that moment the different local slacks are full of question marks and guesses in the different translator channels around. Is it because some major plugin builder has difficulties with the current structure? Or is it that Matt wants to apply Mode 2 of the so-called BiModal IT to bring innovation?
The current way polyglots are working to get the translations ready for the end-user does not really differ from e.g. developing WordPress Core: Anybody can propose changes (Core e.g. via Trac, Translation directly in GlotPress), others are validating (Guest committers are called PTE for translation) and if you show what you’re worth and doing good, you can become a Core committer or GTE.
The main question was about the bulk import for top plugins that have a professional translation service (or outside of the centralized one). So why a single account would not be allowed to upload translations cross-language. Matt specified that “There are no code audits, so why would we have them for translations. We don’t need everything to be process and rule driven — there can be exceptions when the context and situation warrants”. The proposed way forward was “as in the PTE and GTE model, there’s not a *TE for someone who can approve across locales, I think that is what’s worth trying, in a whitelist approach. But we’d know pretty quickly, and just like in code you can roll things back pretty easily.”
The reply of the Polyglots was that “Quality is the main concern here. Users (clients) will blame poor translation on their developer and the platform. They’ll say that WordPress is low quality, unprofessional. Mistakes will be much harder to fix than prevent. The native speakers that WordPress has as staff, are the PTEs for plugins.” There might still be some bottle necks, but the polyglots community has expanded enormously in the last year(s). Certain locale are adding PTEs almost at a weekly basis so translations get validated and become available quickly. The polyglots team has performed an incredible work so far, and to stay on top of things, they would now benefit from development help to e.g. get notifications, easier process for assigning PTE, translation memory, global glossaries, etc.
But maybe the intervention was all about “rocking the boat”, challenge the polyglots and give them an opportunity to become the most agile part of WordPress by switching from pre-moderation to “trust and then verify” to keep the rules agile? In this innovation model it’s okay to mess up and fail (fail is First Attempt In Learning, right?). As Matt put it: “It’s cathedral vs bazaar, a fear of mistakes outweighing agility and speed. We can get really locked in past decisions. All of the concerns are based on hypotheticals rather than actual experience trying it”. This view seems to be globally shared with other companies that don’t want to just improve or expand existing processes. Try something completely new and different and see what comes out of it. If 2 out of 10 things succeed in this new way, you can already call it a success! If it’s not as expected, roll back.
For sure whatever the reason behind the sudden appearance of Matt was, Polyglots will now sit together, setup a trial phase with a limited number of participants, probably for a limited time and then evaluate if the goals like better quality, more coverage and/or better user experience are reached.
To be continued…