Updated templates for master

l10n
#1

Sometimes 'cosmetic' change could mean a difference. To me is indicative if string is 1) in all caps, if 2) first letter is capitalized and if 3) all letters are lowercased.

Not a good reason Kruno. For one, not all scripts even entertain the idea of caps. CJK, Nagari, Amhara, Arabic-based ... non of them have caps. Yet somehow they manage to make their UI work.

Secondly, quite a lot of locales have strict lowercasing. Like ours. So caps in en-US don't make a blind bit of difference to our translations.

Thirdly, there is largely no rhyme, no reason, and apparently no season. In en-US, caps follow whatever the flavour of the month is. Evidence? Sure. Just as Microsoft was abandoning sentence case, LibreOffice went for it, or the other way round, I can't remember. Either way, round about the same time, their respective UI went *exactly* the opposite way.

If LO does not want to start treating en-US as a translatable locale, it really needs to introduce a new change category into Pootle of "cap change" which does not void existing translations but introduces a QA flag like "missing placeholder".

Right now, it feels like every time en-US play with their caps, they give 100 or so other locales the finger, as if our time didn't matter two hoots. And it's NOT like this is a novel issue, we've debated this ad nauseum and STILL you dump these on us volunteers. Do you know how much I'd charge for this kind of time?

Michael

#2

Translating LibreOffice bought me a house.

Kruno

#3

I'm not sure if this is meant funny or serious. It's not funny. I don't expect paying for translating LO but I would consider it courteous if a project I'm donating *lifetime* to would not behave as if my donated lifetime is worth nothing.

If serious, then hooray for you, a lot of us are not getting paid for our time.

Michael

Sgrìobh Krunose na leanas 13/10/2017 aig 07:00:

#4

Seconding this. The project just throws away the l10n people man-hours (but just you try to get them dev guys to fix something in the code!).

That's not right, and in OSS you'd definitely want to pay attention to putting things right (the crypto-currency here!).

...

It's not funny. I don't expect paying for

...

#5

What are you talking about? Developers are committing many patches per
day, even on holidays (see e.g. our GitHub Pulse or our dashboard).
This is a ridiculous accusation.

Also, I translate for Mozilla and cosmetic changes are something that
happens every day. Why is it so controversial here to try to fix the
strings? It’s user experience and perception, which in my book matters
a lot.

Everyone please have this in mind: for better or worse, many of our
developers aren’t native speakers and can’t contribute high-quality
English strings the first time they land a patch, so we inevitably
have to iterate. While I’m conscious that our tooling isn’t the best
and does lead to wasted time over repetitive actions, I would—as a
translator who contributes to Mozilla, Ubuntu, Fedora and many smaller
projects—appreciate a little understanding that any software in
continuous development IS BOUND TO CHANGE.

Adolfo

#6

A little clarification: I don’t intend to convey that your time isn’t
valued, guys and gals. And you, of course, don’t have to agree with
me. After all, this is just an idiot guy in his twenties…

#7

You are talking wisely

Seconding this. The project just throws away the l10n people man-hours (but
just you try to get them dev guys to fix something in the code!).

What are you talking about? Developers are committing many patches per
day, even on holidays (see e.g. our GitHub Pulse or our dashboard).
This is a ridiculous accusation.

No. And you are turning things upside down.

Developers do much, providing they like what they do (are about to do). On the other hand, the important and visible functionality may sit there *virtually* broken for years (like, hey, ms word import/export involving formulas).

Returning to question at hand: people here are expressing their frustration because the actual workflow model allows for cascading workloads for everybody with very little rationale behind.

BTW, does developer side suffer from such kind of problem? Never, I believe. Everything is coordinated (and rightly so!)

Also, I've translated mozilla stack in my day, and I know it allows for *economising* translator's effort -- much of the content is replicated. And I've NEVER seen a 25-30% 'update' of the kind we are talking about here.

-Yury

#8

Because they have a better model in which they update templates every
couple of days or so [1], contrary to our six-month delay where many
source code changes have time to accumulate and then get presented in
bulk to translators. This changes the perception, IMO, because in my
experience, even when there’s more material to translate at Mozilla,
it’s always more manageable…

That is something we need over here.

It might be helpful to know that there are plans to build automated
Pootle pushes [2], which in turn should give us much more frequent
updates and less overhead to worry about (assuming that implementing
it would be a huge time saver for our one overstretched sysadmin… I
might be thinking wishfully of course…)

[1] https://groups.google.com/forum/#!starred/mozilla.dev.l10n/_K2j7Sg0Orw
[2] https://redmine.documentfoundation.org/issues/1488

#9

Hi Yury, all,

You are talking wisely

Seconding this. The project just throws away the l10n people
man-hours (but
just you try to get them dev guys to fix something in the code!).

What are you talking about? Developers are committing many patches per
day, even on holidays (see e.g. our GitHub Pulse or our dashboard).
This is a ridiculous accusation.

No. And you are turning things upside down.

Developers do much, providing they like what they do (are about to do).
On the other hand, the important and visible functionality may sit there
*virtually* broken for years (like, hey, ms word import/export involving
formulas).

Returning to question at hand: people here are expressing their
frustration because the actual workflow model allows for cascading
workloads for everybody with very little rationale behind.

BTW, does developer side suffer from such kind of problem? Never, I
believe. Everything is coordinated (and rightly so!)

Also, I've translated mozilla stack in my day, and I know it allows for
*economising* translator's effort -- much of the content is replicated.
And I've NEVER seen a 25-30% 'update' of the kind we are talking about
here.

Did you read my mail reporting ESC minutes we had last week? Wasn't it
about scripting and the time we have to find to do it?
I'm weekly reporting l10n issues to the ESC, and if you want to join,
you are welcome too.

Cheers
Sophie

#10

...

Because they have a better model [...]

...

Yes.

It might be helpful to know that there are plans to build automated
Pootle pushes [2], which in turn should give us much more frequent

...

I'd say things ought to be organised so that changes like case changes would NEVER create an *obligatory* workload, but... oh, well, let's hope for this, at least.

-Yury

#11

Well, I'm 'out of loop' ATM, but thank you for your effort, the organising team, *really*. I understand all this's a difficult undertaking.

The problem (on which guys are commenting) is -- certain kinds of changes should never generate/cause an obligatory workload.
In your place (I know, it's easy to advise 'from sidelines'), I'd try to look at ways to automate this.

-Yury