Swedish translation
zer...@gmail.com
I made a few small fixes to the Swedish translation. It all started
with the translation for "e-mail address" which in swedish was "e-
postadress:" with a : added, while the original string had none.
But I found that there were a few areas that could use improvement so
I messed around and I hope that the original translator isn't mad at
me for it ;)
(You did a great job Philip et. al! ;))
http://groups.google.com/group/Django-I18N/web/django+%282%29.po
This file is a derivative of ref 6453.
Malcolm Tredinnick
Please open a ticket in Trac and attach the PO file there. That way it
won't be forgotten (put it in the translation component so that I notice
it when I next go looking for translations that need committing).
Also, did you remember to check the djangojs.po file for any similar
problems?
Regards,
Malcolm
Dmitri Fedortchenko
http://code.djangoproject.com/ticket/5687
By the way I submitted it as a patch/diff, I hope that's acceptable.
//Dmitri
Malcolm Tredinnick
> Done on all counts. djangojs.po file also had a few minor issues.
>
> http://code.djangoproject.com/ticket/5687
>
> By the way I submitted it as a patch/diff, I hope that's acceptable.
It's the preferred method. :-)
Thanks,
Malcolm
>
ludvig.ericson
> Done on all counts. djangojs.po file also had a few minor issues.
>
> http://code.djangoproject.com/ticket/5687
>
> By the way I submitted it as a patch/diff, I hope that's acceptable.
>
> //Dmitri
>
>
>
>
> > On Thu, 2007-10-04 at 20:25 +0000, zera...@gmail.com wrote:
> > > Hi
>
> > > I made a few small fixes to the Swedish translation. It all started
> > > with the translation for "e-mail address" which in swedish was "e-
> > > postadress:" with a : added, while the original string had none.
> > > But I found that there were a few areas that could use improvement so
> > > I messed around and I hope that the original translator isn't mad at
> > > me for it ;)
> > > (You did a great job Philip et. al! ;))
>
> > >http://groups.google.com/group/Django-I18N/web/django+%282%29.po
>
> > > This file is a derivative of ref 6453.
>
> > Please open a ticket in Trac and attach the PO file there. That way it
> > won't be forgotten (put it in the translation component so that I notice
> > it when I next go looking for translations that need committing).
>
> > Also, did you remember to check the djangojs.po file for any similar
> > problems?
>
> > Regards,
> > Malcolm
There are a number of translations which I wonder why you changed.
For example, you changed the occurrences of "bytes," where the English
text does indeed say "bytes," in plural and all, to the Swedish
singular equivalent, "byte." I'd love to know why.
Also, you changed "lade" to "la," which is just incorrect.
Another one I'm really wondering is the change of "ett" to "en" on a
noun that is unknown, how could you possibly motivate that?
Additionally, on several places you've changed the translation to one
that doesn't exactly match the English version, for example "We're
sorry, but the requested page could not be found." doesn't comtain the
"We're sorry" part, and there are a lot of that type of retranslation.
You changed "This account is inactive." to "This account is
inactivated." which isn't the same either, when the previous
translation matched exactly, again, why?
I think I was the one who wrote "zipkod" instead of "postnummer,"
because "postnummer" sounds localized to Sweden, when, as far as I can
see, this is actually German, Italian, Japanese, ... zip codes.
Again with "static" vs. "flat," while indeed it should be "flatsida"
and "flatsidor," I don't think "statisk sida" and "statiska sidor" is
correct. It's not "static pages," it's "flat pages."
Other than that, good job, I suppose.
--
Ludvig Ericson
Dmitri Fedortchenko
In Swedish there is no plural versions of the words byte, megabyte, gram, kilogram etc.
It's all singular. You don't usually say "kilogrammer" or "megabytes"?
I also double-checked in several respectable enough places, among them wikipedia
where no plural form of the word byte is ever used.
http://sv.wikipedia.org/wiki/Byte_%28enhet%29
Also, you changed "lade" to "la," which is just incorrect.
I would not say it is "just incorrect". Plenty of searching around has not convinced me.
So I am not sure anymore. As I understand it, lade till is more old fashioned/rural, but I could be wrong.
Anyway I will yield to the original here since I am unsure.
Another one I'm really wondering is the change of "ett" to "en" on a
noun that is unknown, how could you possibly motivate that?
Well, thinking about it now I guess I am unsure.
I thought some more about it in context and I think the original is better.
Although I think that it just doesn't sound right either way. There should be a better way to formulate that...
Additionally, on several places you've changed the translation to one
that doesn't exactly match the English version, for example "We're
sorry, but the requested page could not be found." doesn't comtain the
"We're sorry" part, and there are a lot of that type of retranslation.
I felt that it's important that rather then being an exact copy of the English
wording, the Swedish translation should sound professional and true to the language.
Generally the politeness level in Swedish is lower then that of English and I felt that that should be reflected in the translation. If we are to translate every phrase exactly, we might as well just use Google's translator-bot.
I could be wrong, but I think this translation sounds more Swedish.
You changed "This account is inactive." to "This account is
inactivated." which isn't the same either, when the previous
translation matched exactly, again, why?
Thinking about it more clearly, the English version is a little ambiguous.
Since it can mean that the account is disabled, in which case "Inaktiverat" is more correct.
It can also mean that the account has not seen much activity, and in this case I guess the original translation is more correct.
So I changed it back to the original.
I think I was the one who wrote "zipkod" instead of "postnummer,"
because "postnummer" sounds localized to Sweden, when, as far as I can
see, this is actually German, Italian, Japanese, ... zip codes.
In Swedish the word "postnummer" does indeed mean post code and zip code, so I feel it is only natural to use it.
http://sv.wikipedia.org/wiki/Postnummer
http://lexin.nada.kth.se/cgi-bin/sve-eng
Again with "static" vs. "flat," while indeed it should be "flatsida"
and "flatsidor," I don't think "statisk sida" and "statiska sidor" is
correct. It's not "static pages," it's "flat pages."
Well the fact is that "flatsida" sounds more like "a lesbian's page" (flata is these days one way to refer to a lesbian).
I felt that for the sake of avoiding ridicule and jokes about this newest addition to django,
we should rather use translations which make sense rather then half-translating.
According to the documentation of the flatpages application, it is used to store HTML content in a database. This in my book constitutes static pages.
However, in my quest to find a better translation I did remember the word "platt" which in fact also means "flat".
I still think static sounds better, but I changed it to "Platt sida" to be more true to the original.
I uploaded the new patch, feel free to check it.
//Dmitri
Dmitri Fedortchenko
I also talked to some friends and came to the conclusion that the best translation for "We're sorry" seems to be "Vi beklagar", so I updated my translation to reflect this:
"We're sorry, but the requested page could not be found." in Swedish:
"Vi beklagar, men den efterfrågade sidan kunde tyvärr inte hittas."
More true to the original then both previous translations.
//D
ludvig.ericson
> > For example, you changed the occurrences of "bytes," where the English
> > text does indeed say "bytes," in plural and all, to the Swedish
> > singular equivalent, "byte." I'd love to know why.
>
> In Swedish there is no plural versions of the words byte, megabyte, gram,
> kilogram etc.
> It's all singular. You don't usually say "kilogrammer" or "megabytes"?
> I also double-checked in several respectable enough places, among them
> wikipedia
> where no plural form of the word byte is ever used.http://sv.wikipedia.org/wiki/Byte_%28enhet%29
Okay, I can live with that, though I'm rather used to saying "två
bytes" etc, but I've heard people just say byte there too.
> > Also, you changed "lade" to "la," which is just incorrect.
>
> I would not say it is "just incorrect". Plenty of searching around has not
> convinced me.
> So I am not sure anymore. As I understand it, lade till is more old
> fashioned/rural, but I could be wrong.
>
> Anyway I will yield to the original here since I am unsure.
Lade is more formal indeed, but I don't feel that we should use
colloquial language in a translation, as you note further down, it
should feel as if it was written in Swedish from the start, not
translated. I for one suggest we use formal language, because it's the
most neutral, and Svenska Akademiens ordbok lists lade as the proper
version, and la as the more colloquial one:
läg³ɑ², v.
-er, lade la³de², vard. äv. la la⁴
> > Another one I'm really wondering is the change of "ett" to "en" on a
> > noun that is unknown, how could you possibly motivate that?
>
> Well, thinking about it now I guess I am unsure.
> I thought some more about it in context and I think the original is better.
> Although I think that it just doesn't sound right either way. There should
> be a better way to formulate that...
>
> I will yield to the original here.
Yeah, there's no proper way to do it except for "en/ett," which is
shady in its own way. I find "ett" more proper since you say "ett
ting" and "ett objekt," but that might get thoroughly wrong in other
places.
The problem is that we can't adjust Django itself to work with the
language enough for it to feel completely natural, the only solution
here would be to let programmers flag a model's genus, which I don't
think will happen nor do I want to see it happen, so in conclusion
either one would do, but you having changed it stumped me a bit. :-)
> > Additionally, on several places you've changed the translation to one
> > that doesn't exactly match the English version, for example "We're
> > sorry, but the requested page could not be found." doesn't comtain the
> > "We're sorry" part, and there are a lot of that type of retranslation.
>
> I felt that it's important that rather then being an exact copy of the
> English
> wording, the Swedish translation should sound professional and true to the
> language.
>
> Generally the politeness level in Swedish is lower then that of English and
> I felt that that should be reflected in the translation. If we are to
> translate every phrase exactly, we might as well just use Google's
> translator-bot.
> I could be wrong, but I think this translation sounds more Swedish.
I understand, and agree. The question is where the border goes from
translating to rewriting, a few rewordings are fine, but yeah, depends
on how deep you want to go in translating. There are many more of
these issues, the comment score system is one, at least in my mind,
the whole thing feels awkward, and almost impossible to properly
translate.
> > You changed "This account is inactive." to "This account is
> > inactivated." which isn't the same either, when the previous
> > translation matched exactly, again, why?
>
> Thinking about it more clearly, the English version is a little ambiguous.
> Since it can mean that the account is disabled, in which case "Inaktiverat"
> is more correct.
> It can also mean that the account has not seen much activity, and in this
> case I guess the original translation is more correct.
>
> So I changed it back to the original.
I can't say I agree. If I see "Det här kontot är inaktivt," I'd
interpret that as my account having been inactivated by an
administrator, not that I haven't logged in for a while. Swedish
"inaktivt" has the same ambiguity that English has, so I don't get why
not just stick with it?
> > I think I was the one who wrote "zipkod" instead of "postnummer,"
> > because "postnummer" sounds localized to Sweden, when, as far as I can
> > see, this is actually German, Italian, Japanese, ... zip codes.
>
> In Swedish the word "postnummer" does indeed mean post code and zip code,
> so I feel it is only natural to use it.
> http://sv.wikipedia.org/wiki/Postnummer
> http://lexin.nada.kth.se/cgi-bin/sve-eng
Yes, it does mean that. However, when a form asks for a 'postnummer,'
my guess would be my Swedish zip code, i.e. something bound to Sweden,
not a general one.
Further, Lexin does actually agree with me:
Svenskt uppslagsord
post|nummer -numret -nummer -numren (el. -nummerna) subst.
ett femsiffrigt nummer före ortnamnet i en postadress
Note how it explicitly says five-digit number, and
Engelsk översättning
post code, (zip code US)
So "postnummer" isn't a proper translation, at least I don't think it
is, I've seen sites using "zipkod" as the translation, to avoid this
confusion and I thought that'd make most sense.
> > Again with "static" vs. "flat," while indeed it should be "flatsida"
> > and "flatsidor," I don't think "statisk sida" and "statiska sidor" is
> > correct. It's not "static pages," it's "flat pages."
>
> Well the fact is that "flatsida" sounds more like "a lesbian's page" (flata
> is these days one way to refer to a lesbian).
> I felt that for the sake of avoiding ridicule and jokes about this newest
> addition to django,
> we should rather use translations which make sense rather then
> half-translating.
>
> According to the documentation of the flatpages application, it is used to
> store HTML content in a database. This in my book constitutes static pages.
>
> However, in my quest to find a better translation I did remember the word
> "platt" which in fact also means "flat".
>
> I still think static sounds better, but I changed it to "Platt sida" to be
> more true to the original.
Yeah, I don't think flatsida is a smashingly good translation,
however, I don't think flatsida would get the lesbian connotations you
speak of, the English flat, when referring to the adjective, is only
properly translated into flat in Swedish, for example "handflata,"
"flat tallrikar," and so on.
Static also bears meaning of constant, unchanging, as Lexin suggests:
Svenskt uppslagsord
statisk statiskt statiska adj.
stillastående, oföränderlig
I'd say "platt" is the best suggestion so far, though, shouldn't it be
a compound? I mean, both compound and non-compound versions make sense
in this case, but since even the English version -- and English
doesn't do compounds very often -- uses a compound there, so it's more
of a name than a description, and hence, using a compound in Swedish
would both be at least as accurate and gain a little on staying true
to the original text.
> I uploaded the new patch, feel free to check it.
As soon as I get the time, though I'd bet some other Swedish
translator will come by any time soon.
It'd also be nice if we could get a list of Swedish translators, I
think we're about 3-4 now, those I know are me, mr. Fedortchenko and
mr. Lindborg.
--
Ludvig Ericson
Malcolm Tredinnick
view. Good to see some convergence on a single version and a discussion
of the rationale.
I have a couple of points to make about the process that might help
future translators. I'm clearly not Swedish, but a few of these points
are true in general for localisation work.
On Sat, 2007-10-06 at 14:14 +0000, ludvig.ericson wrote:
> > On Oct 4, 10:43 pm, "Dmitri Fedortchenko" <zera...@gmail.com> wrote:
[...]
> > > Also, you changed "lade" to "la," which is just incorrect.
> >
> > I would not say it is "just incorrect". Plenty of searching around has not
> > convinced me.
> > So I am not sure anymore. As I understand it, lade till is more old
> > fashioned/rural, but I could be wrong.
> >
> > Anyway I will yield to the original here since I am unsure.
> Lade is more formal indeed, but I don't feel that we should use
> colloquial language in a translation, as you note further down, it
> should feel as if it was written in Swedish from the start, not
> translated. I for one suggest we use formal language, because it's the
> most neutral, and Svenska Akademiens ordbok lists lade as the proper
> version, and la as the more colloquial one:
Different countries are going to have different standards here (and it's
good that there seems to be a good reference in Sweden). If in doubt,
it's reasonably standard practice to err towards a more formal tone in
message text. This makes the program(s) suitable for use in all
environments without causing unnecessary offense or discomfort at using
an inappropriately familiar tone.
Previously on this list (as Ludvig no doubt remembers), we've had the
discussion about formal or familiar versions of "you" (using the formal
version is preferred, appeared to be the consensus). Same logic applies
to these cases.
[...]
> The problem is that we can't adjust Django itself to work with the
> language enough for it to feel completely natural, the only solution
> here would be to let programmers flag a model's genus, which I don't
> think will happen nor do I want to see it happen,
Okay, so you realise the difficulties here. Just to point out difficult
this is, noun classification (by class or gender) isn't really possibly
to do in a truly multilingual fashion. The problem is that whilst, for
example, a lot of Indo-European languages are gender-based (masculine,
feminine, possibly a neutral or indeterminate class as well), this isn't
true in African languages and some Indo-Australian languages. There,
classification is a bit more role-based (e.g. animate vs. inanimate in
some cases) and that is what affects adjective endings and the like.
To make things worse, even in the European collection, there is
contradiction. For example, sun is masculine in Latin and languages
derived from that and feminine in Germanic languages. Moon is the
opposite. There are others, but they are the two easy ones to remember
whenever I need them.
The problem we (internationalisers and translators) are up against are
pretty large. :-)
[...]
> It'd also be nice if we could get a list of Swedish translators, I
> think we're about 3-4 now, those I know are me, mr. Fedortchenko and
> mr. Lindborg.
Agreed. :-)
Another thing that has proven very useful on other projects I've been
involved with (GNOME and KDE being two example where this works) is to
build up a glossary document. Any time you come across a word that is
slightly tricky to translate and you have to make some decision -- even
or especially if it's slightly imperfect -- note it down so that you and
future translators are consistent.
Ideally (and I don't know how to organise that for Django at the
moment), the different languages can at least share the common English
version, so that a new team doesn't have to first work out all the
technical terms.
It really does help in cases where you have multiple translators working
on a project to have a "style guide". Even if the first paragraph says
something like "our main reference is the Svenska Akademiens ordbok" and
then includes the results of anything that turned out trickier than you
hoped.
Of course, for bonus points (the lazy approach, but you can call it
"reuse") you could contact the translators from a big project like GNOME
or KDE and see if they already have a glossary or wordlist you can crib
off initially. :-)
Anyway, like I said, it's encouraging to see a bit of debate about the
precise wording. It gives me some confidence that the result will stand
up under scrutiny. This list is quiet enough in general that it's not a
bad place to have these discussions if you don't have your own
locale-based list.
Regards,
Malcolm
Dmitri Fedortchenko
You might say "2 gigabyteshårddisk" but that's a different story.
As for postnummer, the dictionary clearly states that the Swedish definition of the word is a 5 digit swedish post code, but the english translation is indeed "post code" or "zip code". So in that case I feel postnummer is the most correct.
The "page can not be found", I changed a little so it sounds quite a bit more like the original now.
And finally on the point of "Platt sida" becoming a compound... Well in this case I am not sure that a compound is a good idea... I can't really decide to be honest with you. Let's see if someone else has any ideas.
The other things that you mentioned (en/ett, lade and inaktivt), I changed back to their original versions since I agree with you there.
When it comes to the score system I actually have not seen it in context, so it's hard to translate. Depending on the context, I am not sure that Poäng works. Poäng feels like a score that is infinite, there is no roof. Betyg is more for a set scale like 1 to 5.
So I let it be for now.... hehe
//D
Dmitri Fedortchenko
Sure, static does mean that it is unchanging, but in the old days (and today still), HTML pages that are not backed by scripts are usually called "static pages". That does not mean that their content can not be changed, but rather that it is the opposite of Dynamic.
Which as far as I understand it is exactly the idea behind the flatpages application.
ludvig.ericson
> One note about the the "static" word.
> Sure, static does mean that it is unchanging, but in the old days (and today
> still), HTML pages that are not backed by scripts are usually called "static
> pages". That does not mean that their content can not be changed, but rather
> that it is the opposite of Dynamic.
> Which as far as I understand it is exactly the idea behind the flatpages
> application.
change the fact that the word does have several connotations.
(Posting a new mail after this one because you broke yours up.)
ludvig.ericson
> When it comes to the byte, I just stuck with what's used in most
> publications which is that no plural version exists.
> You might say "2 gigabyteshårddisk" but that's a different story.
kilo," "två kilo," "ett gram," "två gram," I guess "en byte," "två
byte" would be okay, but it still sounds odd to me.
> As for postnummer, the dictionary clearly states that the Swedish definition
> of the word is a 5 digit swedish post code, but the english translation is
> indeed "post code" or "zip code". So in that case I feel postnummer is the
> most correct.
Guess I'm not going to get anywhere with this, but I still think
postnummer leads the user to believe that this is a Swedish, 5-digit
postal code.
[...]
> And finally on the point of "Platt sida" becoming a compound... Well in this
> case I am not sure that a compound is a good idea... I can't really decide
> to be honest with you. Let's see if someone else has any ideas.
Well, just the fact that the English original uses a compound is
enough, IMO, but sure, let's see.
I wrote some to django-se about it.
[...]
> When it comes to the score system I actually have not seen it in context, so
> it's hard to translate. Depending on the context, I am not sure that Poäng
> works. Poäng feels like a score that is infinite, there is no roof. Betyg is
> more for a set scale like 1 to 5.
> So I let it be for now.... hehe
Yeah, betyg is "grades" or "marks," but poäng is "points," not score.
Lexin's best translation for score was "poängsiffra" or "poängtal."
Not very satisfying either, so I guess we'll only be able to decide
when we see it in use--I only use FreeComments, so I never get to see
that part.
--
Ludvig Ericson
ludvig.ericson
wrote:
> This has been an interesting thread to watch from a process point of
> view. Good to see some convergence on a single version and a discussion
> of the rationale.
translations, when there's more than one translator for given
language? That way we can ensure a much higher quality.
[...]
> > Lade is more formal indeed, but I don't feel that we should use
> > colloquial language in a translation, as you note further down, it
> > should feel as if it was written in Swedish from the start, not
> > translated. I for one suggest we use formal language, because it's the
> > most neutral, and Svenska Akademiens ordbok lists lade as the proper
> > version, and la as the more colloquial one:
>
> Different countries are going to have different standards here (and it's
> good that there seems to be a good reference in Sweden). If in doubt,
> it's reasonably standard practice to err towards a more formal tone in
> message text. This makes the program(s) suitable for use in all
> environments without causing unnecessary offense or discomfort at using
> an inappropriately familiar tone.
Lade vs. la is really only about written and spoken language, but like
every other word where the written and spoken language differs too
much, the spoken version starts to creep into being the written one as
well, and we have a revised language. Thusly, it is okay to write
'la,' i.e., it's accepted, but sooner or later somebody's going to
notice it and yeah, why not just fix it.
> Previously on this list (as Ludvig no doubt remembers), we've had the
> discussion about formal or familiar versions of "you" (using the formal
> version is preferred, appeared to be the consensus). Same logic applies
> to these cases.
There's a problem with that, because Swedish does have more polite
forms than what we're using, but it's hard to explain in English
because "you" means both you as in one second person, and you as in a
whole group of second person. Many other languages have separate words
for that, Swedish having "du" for one, "ni" for many. The issue is
that "ni," which is referring to a group, can be used to a single
person but then suddenly gets politeness bonuses. You used to do this
to royalties, and I'm sure you still do, but it isn't used very much.
My rule of thumb there goes, "Have you seen it in Microsoft's
translations?" - because quite frankly, I find it that Windows has a
very natural Swedish for the most of it. (And no, they use "du," the
less polite form.)
> Okay, so you realise the difficulties here. Just to point out difficult
> this is, noun classification (by class or gender) isn't really possibly
> to do in a truly multilingual fashion. The problem is that whilst, for
> example, a lot of Indo-European languages are gender-based (masculine,
> feminine, possibly a neutral or indeterminate class as well), this isn't
> true in African languages and some Indo-Australian languages. There,
> classification is a bit more role-based (e.g. animate vs. inanimate in
> some cases) and that is what affects adjective endings and the like.
>
> To make things worse, even in the European collection, there is
> contradiction. For example, sun is masculine in Latin and languages
> derived from that and feminine in Germanic languages. Moon is the
> opposite. There are others, but they are the two easy ones to remember
> whenever I need them.
>
> The problem we (internationalisers and translators) are up against are
> pretty large. :-)
Yeah, the problem is how widely languages vary, and is most likely
never going to be solved.
[...]
> > It'd also be nice if we could get a list of Swedish translators, I
> > think we're about 3-4 now, those I know are me, mr. Fedortchenko and
> > mr. Lindborg.
>
> Agreed. :-)
We've one now, django-se (mailing list) and #django.se on Freenode.
> Another thing that has proven very useful on other projects I've been
> involved with (GNOME and KDE being two example where this works) is to
> build up a glossary document. Any time you come across a word that is
> slightly tricky to translate and you have to make some decision -- even
> or especially if it's slightly imperfect -- note it down so that you and
> future translators are consistent.
>
> Ideally (and I don't know how to organise that for Django at the
> moment), the different languages can at least share the common English
> version, so that a new team doesn't have to first work out all the
> technical terms.
>
> It really does help in cases where you have multiple translators working
> on a project to have a "style guide". Even if the first paragraph says
> something like "our main reference is the Svenska Akademiens ordbok" and
> then includes the results of anything that turned out trickier than you
> hoped.
>
> Of course, for bonus points (the lazy approach, but you can call it
> "reuse") you could contact the translators from a big project like GNOME
> or KDE and see if they already have a glossary or wordlist you can crib
> off initially. :-)
Hohum, maybe that's a good idea. Perhaps a page on the trac wiki or
so?
As for borrowing from GNOME/KDE, I don't know how much we'd be able to
get out of that, but maybe.
> Anyway, like I said, it's encouraging to see a bit of debate about the
> precise wording. It gives me some confidence that the result will stand
> up under scrutiny. This list is quiet enough in general that it's not a
> bad place to have these discussions if you don't have your own
> locale-based list.
Yeah, we've now an IRC channel at #django.se as I noted earlier, for
those who want support with Django but speak Swedish better, or if you
just have general considerations -- though, the channel is in Swedish.
--
Sincerely,
Ludvig Ericson
Malcolm Tredinnick
> On Oct 6, 4:54 pm, Malcolm Tredinnick <malc...@pointy-stick.com>
> wrote:
> > This has been an interesting thread to watch from a process point of
> > view. Good to see some convergence on a single version and a discussion
> > of the rationale.
> Yeah, perhaps we should have some sort of triage process of
> translations, when there's more than one translator for given
> language? That way we can ensure a much higher quality.
Soon (for some definition of "soon"), we -- the core devs -- want to
move to a system where there are one or two designated maintainers for
each locale. All that really means is that those people will have direct
commit access to their directory under django/conf/locale/ and in that
way, commits to the given language go through them. Not every language
need necessarily have a maintainer immediately, because sometimes we get
a contribution and then the person disappears. A little bit of
persistence will be needed before moving to that step, but for most of
our common languages, you guys have already shown that level of
commitment.
I've been very aware of the lack of formal triage that goes on at the
moment. Any large translation change from somebody I don't recognise as
a regular translator, I'll usually bounce off the most recent regular
translator. This particular Swedish patch had already crossed my
boundary into needing another opinion before this thread started. Really
small changes that look like obvious typo fixes or adding missing
strings, I'll generally commit immediately, but even that's a bit risky.
So, yes, we're trying to move towards a system that provides a little
more double-checking for new contributors and remove the single point of
failure from me. The first step (allowing Trac registration) has already
happened. Jacob and I are both pretty busy at the moment, but I'll poke
him again about setting up the next step.
I think for most languages, it's fairly clear who the regular
translators are, so hopefully there won't be any real controversy when
then time comes.
Regards,
Malcolm
ludvig.ericson
On Oct 7, 5:24 pm, Malcolm Tredinnick <malc...@pointy-stick.com>
wrote:
> On Sun, 2007-10-07 at 07:52 -0700, ludvig.ericson wrote:
> > Yeah, perhaps we should have some sort of triage process of
> > translations, when there's more than one translator for given
> > language? That way we can ensure a much higher quality.
>
> Soon (for some definition of "soon"), we -- the core devs -- want to
> move to a system where there are one or two designated maintainers for
> each locale. All that really means is that those people will have direct
> commit access to their directory under django/conf/locale/ and in that
> way, commits to the given language go through them. Not every language
> need necessarily have a maintainer immediately, because sometimes we get
> a contribution and then the person disappears. A little bit of
> persistence will be needed before moving to that step, but for most of
> our common languages, you guys have already shown that level of
> commitment.
Mh, yeah, might help. Though I can see controversy arise when two
translators are in dispute of a translation, making their change and
committing, then the other translator changes back and commits, and so
on, and that'd be pretty nasty for everybody using trunk, as it'd bump
the repository revision each time.
> I've been very aware of the lack of formal triage that goes on at the
> moment. Any large translation change from somebody I don't recognise as
> a regular translator, I'll usually bounce off the most recent regular
> translator. This particular Swedish patch had already crossed my
> boundary into needing another opinion before this thread started. Really
> small changes that look like obvious typo fixes or adding missing
> strings, I'll generally commit immediately, but even that's a bit risky.
Ah yeah, but you've to watch out with these typo-look-alikes, they can
be quite deceptive in terms of breaking grammar, and as have been
shown, some translation updates aren't called for because there's just
is no proper way to translate. (e.g., the neutrum/utrum genus on an
unknown noun.)
A triaging process would really solve 99% of the controversy problems,
and I don't see why this isn't possible right now - I mean, simple as
this:
1) Translator adds patch to trac, and notifies this mailing list and/
or locale specific (e.g., django-se)
2) Other translators review changes, and this is where triaging takes
place, just not in terms of trac logics: One or more translators will
have to approve or disapprove.
3) A commit happens, be it through you or the locale-specific commit
man.
(p.s., anybody know why does Google Groups mashes up the newlines in
most of my messages?)
--
Ludvig Ericson