#musicbrainz

catgroove: :D

damn.. i should get back to studying korean.

warp: I've thought about that too, but it doesn't seem to pose a huge problem, at least not much more than the natural melody of english limits english lyrics. even when singing the tones don't get completely mangeled and it's possible to understand

foolip: http://wiki.musicbrainz.org/ModeratorLanguage/

foolip: no, I mean if it states "sung together with" it's pretty much safe to use 'performed vocal'. If it only states 'featuring' but you know he only did the vocals imho you shouldn't use 'performed vocal'

which language list is that?

Prodoc: yep, certainly

Prodoc: why not?

foolip: a languagfe list of editrs

I've only added these relationships for songs I've actually listened to, so far

im in the 'norwegian' section

you ucnaaddyoursefl manually

ah, I though it was a mailing list

very well, will add

no no :)

yay!

and also with tonal lnaguages you can make puns

norwegian is also a tonal language

not as much as chinese.. but astill

yeah, the word puns are pretty horrible here, hehe

(I'm in beijing btw)

warp: because it's ambiguous. You would require additional proof for it to be valid

åpen bar -> åpenbar

open bar and.. well revelation

yeah, don't split the words in swedish/danish/norwegian

Prodoc: ambiguous?

'open to different interpretations'

split the words?

there was a kind of civil movement against "separating the words" in sweden a while ago

Prodoc: yes, you would need knowledge of the artists involved, certainly. but if you have that knowledge, and can possibly back it up, why not add the more detailed AR ?

lol never heard of it

hmm

catgroove: writing åpenbar as "åpen bar" is what I mean

like.. ifits one word its different than if its two

no because åpenbaar is revelation

like I said, with additional proof it would be ok

yeah, it doesn't mean the same, and it doesn't same the same

but just 'knowing' by listening isn't sufficient imo. What if he provided a minor sample as well?

!

grøntf\or vs grønt fór

eh

never mind

can't type today

hello all

I'm amazed by the "Is This Bad Terminology?" discussion in ModeratorLanguage

hello klokie

hi, please forgive me if I missed this point in the FAQ, but I am really puzzled as to why the MB initiative switched from RDF/XML to just XML?

yea I know foolip

wtf

it just seems to me like a step backwards - that it would be better to work toward integrating the amazing resource that is the MB db with other RDFS data sources cropping up almost daily

my $.02

hmm

ruaok? a Q for you?

klokie: I'm trying to locate the statement as to why MB stopt using RDF but I can't find it as fast as I would like to. It basically came down to RDF is not being used and it was too much effort to continue using it

oy :/

thanks for your reply

but I think things have changed quite a lot in the last 3 years WRT RDF

that's just another flavour of xml

Any pythoners here?

#python is so stupid debate about electricity and saving our planet

is some*

it is a flavor of XML, but one with a clear, extensible vocabulary when applied to MB, which would allow use of MB data in other contexts

the current xml can't be used in other contexts? :)

lol, sorry faux

I just don't wanna go asking widely off topic questions needlessy

acid2: i do bits of python, but i'm very new at it.

I'm having a bit of trouble with pickling... I pickle an object from one script, and wish to use the result in another module. When I dump the object, the pickled file starts with "(i__main__HeadlessDisplayer", but when I try to import the class in my consuming module, it appears as "(idTorrent.torrents.views HeadlessDisplayer"

That's the question

lol

of course it can :) but it seems to require more effort on the part of all other groups in order to align the namespaces, object hierarchies, etc. in order to get them to coexist

klokie: I tried giving rdf a go a while back. At it's current state it's a lot of trouble without much to gain imho

ok

cheers

of course, you could always create an XSL file to convert the XML output to RDF/XML ;-)

yep, working on that :]

hehe

I can tell you now that that's going to be slower than doing it by hand. :P

if you manage to create a proper format, it might not be too much trouble to let MB serve two format but don't push your luck ;-)

would you care to explain why you want it in RDF besides just for the sake of it being in RDF? how do you plan on using it in a better way than just XML? (trying hard not to sound too negative about RDF and discouraging you)

catgroove is now known as LOLgroove

HUHUHUHUHUHUHLOLHUHUHUH

AH

LOLgroove is now known as catgroove

well here's one example: make a query to show me all artists who released recordings in the 1970s

and this can't be done in the pure XML format?

right now, I would have to 1) use MB's server @ 1 query per second or set up a mirror, make a query to the MB-specific free text search engine (Lucene?) that is not completely up to date, and that returns a result set in XML, which I would then need to transform (via XSL?) into a displayable format

or maybe I am missing something

Just use the database, easy SQL query.

*Just mirror the database

ok

you would have to do the same for RDF

well I could just set up a SPARQL query against the RDF datastore, and I have my results

I have no idea how rdf works, but wouldn't that involve downloiading the whole rdf dump (or diffs thereof) frequently?

if a record label has two bands record the same song but in different languages, would the last one released count as a cover of the first one?

right, the database would need to update via diffs periodically, as is the case currently

the way I look at it is it's a limitation of the search functionality of MB, it wouldn't matter if you get your results in XML or RDF

I agree

Yeah, my problem is understanding why you want the data in a format that's not sql, assuming you can't get /just results/ from the mb server.

the idea behind RDF, and Dublin Core for that matter, is interesting. The only problem, the reason why I lost interest in it a while ago, is that it doesn't work. Like with Dublin Core you'll run into limitations really fast. As a result you'll start defining all your own things again

(tags, attributes, etc)

data in rdf is self-describing, which is something that sql can only do on a human level and XML can sometimes do but only on a project-by-project basis

Self-describing is a joke, get over it. :P

ok

indeed it's self-describing but it misses it's point if you have to add your own stuff again

</biased after having to use SOAP>

what exactly defines "cover" and "earlier version"?

SOAP is different story, really - it's no longer Simple is it?

anyway thanks a lot for your responses

would I be stupid to add a "earlier version" relationship between two different artists?

i'm not sure i follow

foolip: 'earlier version' is e.g. an extended version or a demo version of the track by the same artist

yarg thunderstorms

the reason I'm hesitating is because the same song writer rewrote the song in another language and gave it to another band, not exactly what I think of as a "cover"

momvisit

but perhaps cover is close enough...

that's a bit of a borderline case. I think cover would indeed be more suitable

very well, cover it is

though you might want to add that as an edit note in case others think different about it

yep, will do

what can be done when there are several releases of both the original song and the cover version? if relationships between all version are to be created that would be N*M relationships, which seems overkill

but since there is no concept of "song" in musicbrainz, what can be done?

the general consensus on that matter is to use only the earliest releases until we have a way to overcome the problem

sounds good, and happens to be what I've been doing

on a separate note, is there any relationship like "parodies", "references" or similar?

something which isn't a cover, but maybe borrows parts of the lyrics

not currently afaik. there is _some_ discussion on the .. earliest release of .. AR wiki page.

foolip: feel free to add such suggestions to the wiki, http://wiki.musicbrainz.org/OtherVersionRelatio...

thanks, will have a peek

Not at this stage as far as I know. I believe there was a decision on how we should handle parodies but it hasn't been implemented yet

foolip: or, if you feel more strongly about it, write a proposal for the musicbrainz style mailinglist.

maybe I'm not that enthusiastic

:)

foolip: then adding the wiki is best, that way when someone else does start such a discussion again... it's sure not to be forgotten.

sure