#musicbrainz

I think my dad's (really old) CD player displayed fractional seconds, but it's been so long since we've turned it on, I can't recall.

it just bugs me when the total duration calculated by my player is off by 5 seconds or more :|

luks: that's the question, isn't it? Precision and consistency, or practicality?

windows MCI has a bug, so any data CD will show up the duration with +2 seconds

yeah, I think so

copper: I agree that it would be a closer approximation, but unless we actually displayed 1/75ths of a second it would probably not add up exactly.

(date CD = multi-session CD with data session)

MartinRudat: that's what I just said :)

MartinRudat: why not?

if you round all track durations and the CD duration, it should match

luks: not always

that's the point of rounding, isn't it?

well, not for tracks below 75 sectors

all the frames lost or added while rounding might add up to a full second lost or added

er, right

imagine ten tracks with XX minutes YY seconds and 37 frames

luks: 0.75 0.75 0.75 total 2.25 => 1 1 1 sum tracks = 3 vs round total = 2

...provided my math is correct.

the question now is if 0 0 0 and 2 is better than 1 1 1 and 2

lol, that's a question?

:)

if musicbrainz starts displaying rounded durations, users might start bitching to the developers of their players :P

hmm... could make a config option: truncate, round, or display fractions? =)

with javascript on-the-fly calculation?

could be done easily enough... though if it's worthwhile doing it that way or not...

I'm really just nitpicking here...

in any case, changing (int)(X/75) to (int)(X/75 + 0.5) should be easy enough

...but would you end up with more complaints or less? =)

haha, you tell me

probably more.

then again, that's what FAQs are for.

I did search the FAQ about that before asking :)

fixed - http://bugs.musicbrainz.org/changeset/9330

eh, cool :)

When does that go into production? Is there a periodic software update, or is it matter of software releases?

it will go live in the next release

which will be probably in a month and something

ok

hmm... of what use are changelog/edit notes on release annotations, given that they're autoedits? or is it to give later people a hint about reverting the edit or not?

... "The world's best ever beer songs 2005" released 2004-11-11 ? hmm... I suppose that makes sense more or less...

Like magazines appearing in March 2007 with "June 2007" on the cover. ;-)

hmm... the publishers have a time machine? =)

it's fashionable I guess

...better than '98 coming out in '99, I suppose.

Then again, it's currently 1989 and yet for some reason I'm getting albums released in "2007".

hmmm you don't seem like yourself, Marty

The 80s are 4-ever. ;-)

HMA: heh. =)

"who're you calling a chicken?" =)

luks/mo/Martin: Just typed up and sent out on style the results I found from going through and cross-comparing every single AR type we have (or don't have, in a few cases).

If that's what's on the cover, then that's what it should be.

'twould probably help if I was subscribed to the style list, I'd imagine. =)

I ran into that a while back. Check the ARs for Edison Phonograph Company, North American Phonograph Company, and National Phonograph Company

BrainFreud: I thought that for earliest release relationship, the audible content of both tracks in the relationship are supposed to be identical... in which case, the produced version can't possibly be the earliest release of the demo...

It's a little confusing, cause we have both a "earliest version of" and a "earliest release of" track-track AR

that one's talking about the "earliest version of" AR

we do? ... agh!

yeah

"earliest release of" covers the non-remastered track re-release situation. "remaster of" covers the remastered track release situation. "earliest version of" is the catchall for the rest

except... the wiki for the "earliest version of" incorrectly says that "remaster of" ought to be used, when it means "earliest release of" ought to be used. :P

oh, I see now.

hmm... 'She Will Have Her Way: The Songs of Tim and Neil Finn' is a cover of 'She Will Have Her Way: The Songs of Tim and Neil Finn (bonus disc: The Original Versions)'? =)

lol, I knew there had to be one example of it in there - but really, as a separate type of AR from just "cover of" on all tracks, is it really needed? Or does it just over-complexify?

even a single track from a different release on the cover album would invalidate that album-album AR. Just seems overly redundanty, and very limited in when it can be used, while essentially duplicating the much more easily used track-track AR

well... if someone tells me how I can add multiple relationships in a batch, I can see having release is a cover of release as being redundant... then again, how about track vs release performance relationships?

My opinion is that all ARs ought to be applied at the lowest possible level, since they don't inherit downwards

we need to slowly get rid of release performance ARs

I agree that 3) is needed.

#6 was the real reason I decided to write this up :)

luks: do you want to be lynched by people entering classical releases? where every performer in the release is billed?

no, I don't want

MartinRudat... I just did almost 14,000 track ARs... :P

that why I fixed the interface to make it easy to add track ARs

the problem is that if you will have one classical tracks on two releases, you can either have one set of track ARs or duplicate release ARs

MartinRudat: nothing has changed yet

it will be in the next release

ah. okay. I won't be getting my hands on my CDs until christmas time, anyhow. I don't see myself going to get a classical release just to add it to the database...

release level.

"Yet we have this AR only at the track level, and not the release level" ought to read "Yet we have this AR only at the release level, and not the track level"

BrianFreud: I don't expect you will have many comments on mb-style

if you want to get this done, you will need to split it into smaller changes

I figured it was strictly style related, but you're right

then say what exactly do you want to change

RFC -> (maybe discussion) -> RFV -> (maybe veto) -> a couple of days -> done

in other words, there is noone would you say: ok, let's do this

you will need to be the person

lol, easily said, hard to do - been through that twice :P

no, you haven't been there

even the suggestion of an RFV phase seems to kill all discussion on all proposals from anyone

people usually expect that they post something to mb-style and then somebody will decide

but it doesn't work that way anymore :/

Right - the difficulty I ran into is that some people mis-see the RFV phase as simply "I can say I veto it".

there is no problem if the discussion stops

And once they throw out the veto without an alternate solution, that's when I find discussion normally stops.

that just means it's time for RFV

that's why the DQ one stopped - a pre-emptive veto was sent before it ever even hit RFV, and everyone clammed up

preemptive veto means nothing

people like to argue, that's all

you know that and I know that, but not a lot of people on style seemed to recognize that - and I wasn't quite sure how to get it through that RFV into actually having something be done...

that's where even the wiki admits the process gets very hazy

but what I'm trying to say, if it stops, it means that you should pick it up, not that it's dead

imagine you are the one running this particular part of MB

you are not asking somebody else to do a change

hmm... all interesting suggestions, but I'd say that you want to have sent out (at least) 7 different emails for that lot.

So, "Ok, this is now an RFV. Any veto must propose an alternate solution. If no such vetos occur in the next two weeks, we'll consider this RFV to be finalized"?

you are asking for permission to change something yourself

a veto can mean that the problem you are trying to solve if not a problem for other people

I wasn't even aware you _could_ attach urls to tracks (given I've never had an occasion to want to)... so _that_'s where you stick information about where stuff got recorded...

then you can't have any alternative solution

MartinRudat: yeah, I only found it cause I was looking for any ARs I'd not seen before

and a track-url is not really anything like what you need for that type of info

dunno, there's at least one album I know I'm going to be adding that AR to, it's got tracks recorded in all sorts of different places in the UK...

yes, but you have to link the track to the official site of the recording location

...though, I wonder if the website of a church or a cathedral would count as a 'recording studio'... =)

Would make MUCH more sense to be able to set the recording location up as an entity, link the track(s) to that entity, then that entity be able to have a "has an official site at" AR of its own...

isn't there many other missing ARs to add?

for these kind of ARs I'd wait until we have proper support for them

Sure, but I was trying to deal with where the ones we have already are funky or just plain "huh? wtf???"

no, I meant martin's recording studio URLs

luks: Well, all we'd need is to add a new labeltype "Recording Location"

hell no

lol

totally agree - just looking at overloading it into Label as a shorter term solution

labels are abused enough, already

it could be migrated to Location easily, once separate support for "Location" exists

Just point "location" to that overloaded Label field, and take it back out of the Label subtype list...

it would clutter labels listings too much

has to be a better solution that each track linking randomly to various websites for venues...

*s/that/than

not really

you can easily track the studios from URLs

I agree that waiting for proper locations is better than abusing 'Label' for it.

luks: That assumes the studios have websites

Even assuming every track using that AR were to link to the exat same url every time someone wanted that same location, it limits that AR's use to all locations which have a web presence, thus mostly eliminating everything prior to 1996 or so

BrianFreud: my primary school's church has a website. it's been around for somewhat longer than since 1996; failing that, you could point at google maps, I'd suppose...