#musicbrainz

it solves the problem of languages where it doesn't make any sense to have   at all, but translators go "wtf is that" and include it

isn't there some css stuff for avoiding wrapping anyway?

ianmcorvidae: do you think that would be less of a problem with   ? or about the same?

the latter is clearly a space and won't be included by people who know there's no space. nbsp is clearer in that people have more chance of recognising what it is

less, because people have some idea what nbsp is

(i.e. is it likely that translators do not know html character references at all, or is it likely they don't know the numeric ones but would be familiar with the named ones?)

the latter, though ndash vs 70695238345 or whatever it is benefits more

if it was &nonbreakingspace; or something they'd be comporable

comparable*

ianmcorvidae: would it be possible for the tools which extract strings to perform a translation from &xa0; to   ?

it'd be possible, I suppose

ianmcorvidae: and do the reverse transformation in the build step?

I almost feel like that's the wrong place though, it'd be better to just have   in our source and have the translation system (whether or not it's actually translating to another language) turn   to  

ocharles didn't want to have these transformations applied at runtime, but I think there is value in doing them, and I expect ocharles will be less opposed to doing this at build time.

ah

runtime is a bit iffy, yeah, it makes things nicest for contributors and translators though

yes. I'm trying to find a workable compromise between the two extremes :)

understood :)

My vote'd be for runtime.

honestly, my initial vote is for "do we need the damn things at all?"

(it's a shame HTML5 is so very lenient in many ways yet doesn't allow the bits needed to specify character entities when parsing it as xml)

Developer-translators occasionally grep for translations string sin the code.

*strings in

If the string is being changed between source and .pot, that won't work.

ianmcorvidae: yes. I like the mdashes and nbsps where we use them correctly.

v6lur, ping

intgr would work too

" which of the two "ja" should become "ning"?

anyway also, to start the task will be making sure we don't lose translations with what we currently have, which is independent of fixing the underlying problem

which hopefully I'll get a chance to do, heh

(taking into account "orchestra and choir" are a package inside the full credit)

reosarevok: Doesn't matter, "ja" and "ning" are the same word. You can leave them both "ja".

I thought I was supposed to use one of each when there are several levels in an enumeration

It wouldn't clarify anything in this context.

reosarevok: some would say that ning is a stronger separator, i.e. similarthing1 ja similarthing2 ning differentthing1 ja differentthing2

i.e. Ülle Ulla ning RAT ...

but "ja" for both is ok, too

Hmm, ok

> navap: General FYI, according to google webmaster tools, Susumu Hirasawa is the highest ranked artist by number of incomming links from unique domains < He is linked on every WP template page for an muscbrainz artist, guess thats giving him a nice boost

I'll leave ja then

reosarevok: re: "use one of each when there are several levels" -- in literary style, you would :)

i'm not sure which style is approppriate for MB

Neither am I :)

intgr: it would carify a bit -- namely that the choir is that of the RAT Estonia

(i mean using "Ülle Ulla ning RAT Estonia ork. ja koor" would)

but like you said, no big difference

I see

This artist http://musicbrainz.org/artist/8a20c051-2920-4bc... has nothing linked to it and cannot be removed

maybe http://musicbrainz.org/edit/19525464 has to fail first?

meh, why isn't that on the artist edits list?

Or is it and I'm just blind? :)

hum, is there a rule about "..." should we use three . or can we use … ?

The same as every other typographic issue I guess

mat_: you can afair. there was some discussion to convert it via guess case automatically i remember (it does not right now)

reosarevok: That rule is "Use of basic ASCII punctuation characters [...] is allowed, but typographically-correct punctuation is preferred" ?

Yeah, so I guess that'd count as … being preferred

intgr, ok, I like that rule very much :-)

mat_: It's a guideline, not a rule, so you could ignore it anyway ;)

jacobbrett, well, in the case of …, ’, ‘ or even « and », my fingers know where they are :-)

*blinks* https://musicbrainz.org/work/8b73b6a9-aa1d-4d3f...

(not saying it's wrong, I'm just staring at it in disbelief)

Morning all

How goes it?

morning

kovacsur, what part feels the worst ? old english ? ballet ? or the half dozen iswcs ?

The iswcs actually exist and they're indeed for this stuff http://www.reverbnation.com/beatswizard/songs

if CallerNo6 reads the chatlogs: IIRC I deleted the Glenn Gould wiki page because every info on it is on the release page. Unfortunately I cannot verify because this release still causes a 502. And it was a box set not a series anyway, so I think this was ok

MClemo: about the Glen Gould, that makes sense, thanks

kovacsur: That's awesome! I want that release!

kovacsur: believe! http://www.united-nations-of-beer.com/images/ca...

what a scam, that looks a lot more like Middle English

I assume someone has already linked to this: http://labs.echonest.com/Uploader/index.html

It’s now beyond Gangnam Style

I seem to recall there being a weird issue a while ago with locales (and multiple names for the same artist). Was any action ever taken on that, to move everything to a "script" rather than "language"?

No.

Also, it's not as simple as that, as sometimes you do want a locale.

(locales are not languages.)

Shouldn't there be a way to specify that multiple languages are the same?

You mean, that one name applies for multiple locales?

Yeah

Like, say en_US and en_GB

Would be nice, wouldn’t it?

But no, there’s no way to do that, other than entering it many many many times

Ugh

There has to be a better way...

Not at the moment, no.

you should only really use en_US and en_GB if the names are specific to those countries

belak: "en_GB" and "en_US" does denote that the languages are the same.

belak: That's the "en" part.

*do

applications are generally designed such that if a user requests en_GB and only en is available, they'll get en.

we really need to replace our locale system in order to have more control over it

there's a number of issues; we're just using a list from some perl module at present

Re: what nikki just mentioned, it'd be nice to have a way to target e.g. an alias at a country regardless of language.

(E.g, Wedlock is known as 105 Owl something in the GB/UK, apparently. And while this may seem to map easily to en_GB - what about sco, cy, ga, ...?)

I liked freso's mul-latn suggestion

hmm, I think mul-Latn actually fits in with my idea fine.

what's your idea then? :P

use IETF language tags with the general format en-GB-Latn

country and script are each optional

that's more or less what we have now

(the script tag is optional in the case where it's the same as the language's standard script)

well, *some* of our entries already have scripts, like chinese

kepstin-laptop: But what if something applies to one script or one country across different languages?

the problem is more the interface

freso: I said that I like your mul idea. so mul-GB-Latn and mul-Latn work there :)

right now we generate a list and it's not that flexible

Heh. True.

but in the db it's just a text string

yeah, interface would be tricky.

Also, when did I propose that idea? >_>

you commented on a ticket a while back

Oh.

7me is teh clever 1

best thing I can think of for the ui is to allow selecting language, country, script separately.

where country, script would default to "any country" and "language's standard script"

a "commonly used complete locales" interface would be good too

Yeah, that Freso guy did make a good suggestion there.

kepstin-laptop: +1