Proposal: splitting blog into separate site, translations go direct via merge request



You imply here the logic of people seeing strings as what ultimately matters. Fine, things seen less, see less work, on every level.

What we disagree over, is that given importance being lower, granted, the importance is not there… That does not compute, and this is where the bottom falls out of the bucket. If that were the case, it would have to be seen never, which means it could be removed. For posterity, and the low volume of traffic it sees, it should shine. It is part of a whole. Some of it is very enjoyable reading.

This argument short-circuited into end-users, a consideration that foregoes the effort translators put in for it to get there. Translators use the search function, and memory, to achieve consistency. Relinquishing consistency is tantamount to opting for auto-translation. That is not good enough for the exact same reason.

When “translating” if I see an error, I try to fix it, or mark it for later. The latter is more taxing not only for me, but for every other translator. Now every translator has to figure it out, now every translator having translator having done it right or wrong has to do something again once someone gets around to fixing it the right way.

It is stressful enough having to read the blogposts when translating them. Incomprehensible yet, though I don’t go so far as calling it one way or the other, is that it can’t be fixed, because I have to care about how important someone who doesn’t think it is important to begin with think it is. I don’t know if it is over how much time it takes to proofread, possibly at that not considering how much time it takes to do?

A lot of times it is non-translators complaining about how much effort it will be to translate. An further oxymoronic conundrum always lost on people that actually want to translate.

If everyone with commit access finds it to be a waste of time, instead consider this, it gets done, and it isn’t dangerous, and easily reverted. If not those that don’t want to, how about the person bringing it to the attention by, doing the actual work. Doing that work from the Weblate editor just takes a lot less time. No source string change gets past me on Hosted Weblate, so why not let everyone at it? It stands to reason that if F-Droid trusts me to screw things up, it could also trust me to ensure others don’t either. If one shows itself to be beneficial, why not put more eggs in the basket?

It wasn’t proofread when it went in, so why care that someone changes it now?
If fewer people see the low volume ones, trusting a system that also has unknown actors is about the same level of risk. Lets make those connections by eyeballing it.

Through the process of translating anything, I see, search for and remember a lot of other strings, and don’t need the added mental activity of figuring out whether or not something can be fixed if I see or know that it is broken. What I think about it matters, because it is feasible for me to fix it all, alone.

I assumedly need to translate all blogposts to know when there are new ones, and where the new ones are. That way I can label them by priority, which I am hoping does something for someone else. Or we do one better and fix things before they land.

In any event this smoothes out the curve of effort of having to do the bulk of the work over for the next person. If something is easier to understand for the anglo end user, it is easier to understand for the translator. In turn, the fewer mistakes made in translation, the fewer people with a subpar experience.

The Debian Handbook can arguably be said to be translated into no more than 9/10 languages. One of which is Norwegian Bokmål. That is not good enough for something that calls itself “The universal operating system”.
I would like to fix more of its content, but I forget what my login is for the repo they use. They make it hard for me, my time is limited, and there we are. Maybe in 2025 it will see as much work as I would do in 30 mins if it had direct editing? That cuts it pretty short for year of Linux on the desktop if you ask me.

Right now we need people to even do the work, nevermind well or perfect, because not translated is not good. Not translated over time is the volume of how not good it has been. The longer a string goes untranslated, the more problematic it is, on every level. Why is that?

The longer the strings are, the harder they are to translate. You ask of someone to translate one sentence, how many of those will you get to translate a page, or a book?
One concept per string is the goal. A wall of text is at one point insurmountable.

The last thing holding up effort is having to translate 1000+(?) blogposts to see any of that effort go live. If one blogpost is translated, that is one blogpost better than no blogposts translated. You want to send someone a link and they can’t read English, and automated translation services are just not going to cut it? Why not translate even the one blogpost, and see results immediately. It works on Wikipedia, and fortunately F-Droid enjoys the spoils of being a big project, so with open doors, even the unfavourable spirit meets an end.

TL;DR: Everything should be fixed, and improved, in the easiest manner possible.


Everything should be fixed. The hard part is prioritizing what is the most important to fix first, and who will do the work. We will never have enough contributions to fix everything, we must prioritize. And we must find out how to make the contributors as a whole as efficient as possible. Sometimes that means that one group has to do something annoying in order to save lots of effort elsewhere in the process. I have to do a lot of things like that when managing translations, others do too.