jacobbrett: re performance links for remasters: adding original composer credits to derived work is also redundant, but we do it
2012-03-01 06133, 2012
rswarbrick
amino_: Hopefully one day, this will be inferred by the software and then we needn't have it in the style guide.
2012-03-01 06134, 2012
amino_
and also, how do you indicate composition/lyrics credits for the remaster recording, if you don't link it to a work?
2012-03-01 06107, 2012
jacobbrett
amino_: As long as the behaviour is consistent :D ..to answer your question..by following the remaster relationship to the original master and checking its related work.
2012-03-01 06140, 2012
reosarevok
Hmm
2012-03-01 06153, 2012
reosarevok
I might expect a bot to add those links to works maybe
2012-03-01 06155, 2012
amino_
jacobbrett: sure following could work, but it becomes far more complex to retrieve basic information ;)
2012-03-01 06159, 2012
reosarevok
But I'd expect them there
2012-03-01 06148, 2012
jacobbrett
amino_: I suppose a database like this is bound to have complex lookups...
2012-03-01 06128, 2012
reosarevok
jacobbrett: sure, but why should we make them intentionally harder?
2012-03-01 06133, 2012
jacobbrett
reosarevok: Well, in that case, perhaps the database should store these relationships transparently, to cut down on (what I think is) semantically redundant edits.
2012-03-01 06146, 2012
rswarbrick
jacobbrett: That's not the right solution. The UI code should be inferring them automatically: I know that it doesn't at the moment, but putting workarounds in the DB code to add redundant edits is a bit silly.
2012-03-01 06133, 2012
jacobbrett
rswarbrick: That's basically the same effect I meant, though with a slightly different user interface implementation...when a remaster is linked to a master, you could just have the DB automatically add the rec-work relationship that already exists with the master. Editing the relationships at either the master or remaster would affect both relationships.
2012-03-01 06128, 2012
rswarbrick
Ick. I think I'd prefer to have a distinction between inferred relationships and manually added ones.
2012-03-01 06136, 2012
rswarbrick
It makes overriding more sane, I think.
2012-03-01 06152, 2012
jacobbrett
Why would a different work exist for a remaster?
2012-03-01 06131, 2012
rswarbrick
D'oh. I was being silly.
2012-03-01 06148, 2012
rswarbrick
I still prefer the inferred option, but I agree that the end result should be identical.
Split or keep together? On one hand it seems to be seen as one "band" in itself, but on other, it feels weird...
2012-03-01 06119, 2012
v6lur
keep
2012-03-01 06134, 2012
reosarevok
Ok
2012-03-01 06110, 2012
v6lur
if anything, it could be seen as an artist credit for vennaskond
2012-03-01 06141, 2012
nikki_ joined the channel
2012-03-01 06144, 2012
Shepard joined the channel
2012-03-01 06145, 2012
v6lur
but i wouldn't merge it into them, either
2012-03-01 06112, 2012
reosarevok
Fine
2012-03-01 06124, 2012
reosarevok
It just looked strange, but I guessed there was a reason it was still there :)
2012-03-01 06100, 2012
v6lur
it's basically members of j.m.k.e., with trubetsky on vocals, playing songs of vennaskond
2012-03-01 06156, 2012
nikki joined the channel
2012-03-01 06133, 2012
reosarevok
Oh, by the way, v6lur: dunno if you've seen, but I've started adding ester links to stuff with the "discography page" rel (which might not fit it exactly, but then again, it fits exactly almost nothing)
2012-03-01 06105, 2012
reosarevok
It'd be cool if you did the same when checking ester for stuff, so we can slowly build a mapping (if for nothing else, for quicker reference later)
2012-03-01 06121, 2012
v6lur
hm, okay, i've thought about doing it myself a few times, too, but it didn't seem to fit :P
2012-03-01 06132, 2012
v6lur
but ok
2012-03-01 06125, 2012
reosarevok
Well, I thought that too
2012-03-01 06158, 2012
reosarevok
But then I saw the pages for it mentioned all kind of random stuff from Allmusic to Amazon, so...
2012-03-01 06150, 2012
reosarevok
(and if someone were to RFC for a relationship for libraries later, it's not as if it would be hard to move them with a script, being fairly predictably found with a search for ester.ee :) )
2012-03-01 06144, 2012
Vorpal joined the channel
2012-03-01 06124, 2012
v6lur
reosarevok: you use the "püsilink", right?
2012-03-01 06130, 2012
ruaok joined the channel
2012-03-01 06130, 2012
reosarevok
Yeah
2012-03-01 06128, 2012
v6lur
k
2012-03-01 06117, 2012
v6lur
would it matter if i were to use et or en as interface language?
2012-03-01 06108, 2012
reosarevok
Probably not
2012-03-01 06111, 2012
reosarevok
We use et
2012-03-01 06117, 2012
reosarevok
(at least I do)
2012-03-01 06146, 2012
reosarevok
But I don't think it matters too much, the only important part of the URL is the b1429022 thing after all
ianmcorvidae, one row per fileid with columns fileid, mb_artist_id, mb_album_id
2012-03-01 06107, 2012
Munger
The media.tags table contains a bunch of tags extracted from files with one row per tag
2012-03-01 06142, 2012
Munger
Just trying to pivot relevant tag rows into columns
2012-03-01 06111, 2012
ruaok joined the channel
2012-03-01 06127, 2012
Munger
I have tried to decipher that page but can't quite get my head around it
2012-03-01 06102, 2012
ianmcorvidae
yeah
2012-03-01 06110, 2012
ianmcorvidae
I'm having trouble getting it myself, heh
2012-03-01 06120, 2012
SultS
reosarevok, do you know is there direct link to MARC data in ESTER #tab3, that doesn’t contain any other data?
2012-03-01 06100, 2012
ianmcorvidae
my guess would be something like SELECT * from crosstab('select fileid, tagsubtype, tagdata from media.tags where [...] order by 1,2') AS ct(fileid text, albumid text, artistid text); or so
I think you can do that with the AS ct(fileid text, albumit text, artistid text); thing
2012-03-01 06117, 2012
ianmcorvidae
but I'm not completely sure
2012-03-01 06140, 2012
ianmcorvidae
aside: we really need to figure out how to make voting happen, we haven't had >200 people voting in a given week since the end of last May (just after NGS release)
I guess it should be possible to extract user-specific vote counts by time, I haven't worked on it though
2012-03-01 06111, 2012
nikki
I could probably extract it myself, but I bet it's in the edit dump
2012-03-01 06134, 2012
nikki
and I'm not interested enough to download and extract a 1.46 gig dump
2012-03-01 06148, 2012
ianmcorvidae
you could use rika
2012-03-01 06156, 2012
nikki
edits are on there?
2012-03-01 06157, 2012
ianmcorvidae
that has not one but two copies of the edit DB :P
2012-03-01 06105, 2012
nikki
ooh
2012-03-01 06106, 2012
ianmcorvidae
yeah, why would I have bothered excluding them?
2012-03-01 06106, 2012
Munger
Bum! Just remembered that some tags have multiple artist IDs --- dfe9a7c4-8cf2-47f4-9dcb-d233c2b86ec3/cfbc0924-0035-4d6c-8197-f024653af823
2012-03-01 06117, 2012
nikki
I always do, I forget they exist :P
2012-03-01 06124, 2012
nikki
because they're too big for me to care
2012-03-01 06128, 2012
ianmcorvidae
haha
2012-03-01 06134, 2012
ianmcorvidae
well, rika has that nice 1TB array :P
2012-03-01 06143, 2012
nikki
unlike my poor server
2012-03-01 06142, 2012
nikki
oh ffs
2012-03-01 06148, 2012
nikki
"Bad: new password is too simple"
2012-03-01 06159, 2012
derwin
is it something like
2012-03-01 06106, 2012
reosarevok
"password"
2012-03-01 06106, 2012
derwin
"purple monkey dishwasher ice cream sundae"
2012-03-01 06108, 2012
derwin
?
2012-03-01 06114, 2012
derwin
which is in fact a great password?
2012-03-01 06114, 2012
nikki
it's a random string of letters and numbers :/
2012-03-01 06128, 2012
ruaok
derwin: it is.
2012-03-01 06132, 2012
ruaok
xkcd says so.
2012-03-01 06135, 2012
derwin
ruaok: exactly.
2012-03-01 06150, 2012
ruaok needs to adopt that methodology
2012-03-01 06147, 2012
warp
if they're all dictionary words, is it still a good methodology (in X years when many people have adopted it? obviously right now it is)
2012-03-01 06149, 2012
ianmcorvidae
warp: depends if it becomes a useful attack vector as a result
2012-03-01 06129, 2012
ianmcorvidae
it's still pretty resilient against bruteforcing (because it's very long) and most dictionary attacks (since those don't usually account for multiple words + space, a special character)
2012-03-01 06138, 2012
derwin
warp: also, if you slightly punctuate that sentence of english words
2012-03-01 06146, 2012
nikki
ianmcorvidae: what's the db called?
2012-03-01 06157, 2012
derwin
purple monkey? 1 drink away from the edge
2012-03-01 06103, 2012
ianmcorvidae
nikki: musicbrainz_db_static is a feb. 2 dump, musicbrainz_db_slave is the replicated one
