[mb-style] CSG compromise?

[Brian Schweitzer]

On Sat, Mar 1, 2008 at 7:44 PM, Lauri Watts <krazykiwi at gmail.com> wrote:
>
> On Sun, Mar 2, 2008 at 12:53 AM, Aaron Cooper <cooperaa at gmail.com> wrote:
>  > On 1-Mar-08, at 2:16 PM, Brian Schweitzer wrote:
>  >
>  >  <big snip>
>  >
>  >
>  >  > So until that happy day of NGS support, how do we compromise?  How can
>  >  > we ensure that the data allows the BBC to identify the work from the
>  >  > track, allows Aaron and me to tag and not find the data "quite messy",
>  >  > allows Leivhe and David to tag and have complete but not "overly"
>  >  > complete data, and still allows Lauri and her mother-in-law to tag and
>  >  > have what she wants?  (Just using us as examples of general user-base
>  >  > opinions here.)
>  >
>  >  <br>
>  >
>  >
>  >  > If we're aiming for "as on the liner", which liner,
>  >  > which translation, and what still gets fixed - caps, typos, mislisted
>  >  > works, etc?  If we're aiming for CSG, how complete, which languages,
>  >  > and are we essentially admitting that classical track titles per
>  >  > liners are descriptions and not "titles" per say?  Where can we all
>  >  > compromise such that the (I think majority) who wants at least some
>  >  > structure in the titles isn't left waiting 1-2 years for NGS to
>  >  > replace CSG, plus all the quite likely years of editing work to
>  >  > actually implement it across our classical listings?
>  >
>  >  Proposition:
>  >
>  >  a) For "chillout classical" releases where tracks are titled "Andante"
>  >  or "Allegro from Sonata No. 14" etc, enter the track titles in MB as
>  >  they appear on the CD and follow regular MB guidelines (http://wiki.musicbrainz.org/PartNumberStyle
>  >  , http://wiki.musicbrainz.org/SubTitleStyle, etc)
>  >
>  >  b) For "real classical" releases, use the CSGS lists for ease of
>  >  entering the releases and consistency.
>  >
>
>  Works for me, for the meantime at least.
>
>  Perhaps with a rider to say:
>
>  If you can't tell which is which, ask in the edit note (because
>  honestly, at this point, we don't
>  need a big discussion on which is which, when it can be easily solved
>  by asking case by case,
>  if common sense alone doesn't work.)
>
>  Or err in the direction of 'copy the cover', because it's easier to
>  fix 'incomplete -> CSGS' if necessary
>  than deal with completely misidentified tracks from people being
>  stunned like deer in headlights by
>  CSGS list and blindly copying anything that looks like it might be
>  close (if you know what
>  I mean, kind of the way PUID's end up assigned all over the darn place.)
>
>
>  Oh and someone wanted some 'ridiculous covers' as examples.  I present:
>
>  http://people.fruitsalad.org/lauri/stuff/the_best_classical_album_in_the_worldever_2004_retail_cd-back.jpg
>
>  http://people.fruitsalad.org/lauri/stuff/the_classical_album_2007_2006_retail_cd-back.jpg
>
>  http://people.fruitsalad.org/lauri/stuff/the_only_classical_album_youll_ever_need_2003_retail_cd-back.jpg
>
>  Of which my personal favourite for idiocy is the last.  I bet you
>  didn't know Pachelbel's Canon is famous because
>  of "Wool advertising"
>
>  Now if I may, I suggest everyone head back over to Olivier's thread
>  and continue dissecting
>  the current CSG paragraph by paragraph, it really needs it.
>
>  So do the rest of our guidelines though,  and if you don't know what I
>  mean, think about
>  "Some Famous DJ & DJ Someone Else Famous pres. DJ Nobody vs. Someone More
>  feat Someone Entirely Else - Some track name - Some Group Name mix (radio edit)"
>
>  ... bearing in mind, Some Famous DJ and DJ Someone Else are both
>  members of the 'Some Group Name',
>  and 'Someone More' could well actually be a performance name for that
>  group (or a group with the
>  exact same members, if you will).  This is an only slightly contrived
>  example though.  And you don't
>  really need to tell me how you'd enter it, it's just to point out, you
>  _have_ to be familiar with the music and
>  the artist to even figure out how to split it into an artist name and
>  a track title, guess case will likely mangle
>  it beyond recogition, and even then, our guidelines on the whole deal
>  are eyebleedingly confusing.
>
>  So though some of us rail a little at the CSG, it's not 'the CSG' that
>  is the problem, really, it's that we actually
>  expect all our users to have some familiarity with the music they own,
>  when they probably bought the
>  classical chillout album on a whim at the gas station, and they
>  probably got that Ibiza club album
>  on the airport on the way home from two weeks on the Costa del Sol,
>  and they know nothing about either
>  and us being impatient with them getting the details wrong when they
>  enter them in MB is not
>  really helping that.

Well, not to throw a kink in the discussion, but if we're talking
about perhaps moving to pseudo master lists now, and the more I think
about that idea, the more attractive it seems to me, then why mess
with CSG or the tracktitles (such as they are) at all?

Yes, it means we get the messiness of some of the types of titles
Lauri provided as examples.  Yes it means all that messiness of
multi-language titles, or releases re-released with different titles
on different liners.  But we knew that - that's exactly what a large
part of this discussion has been about, that CSG avoids.  But I'll be
honest - give me some way to identify / tag to a works list, and I
suddenly care much less about the track titles on the release; let
them be just as ugly as we want.  So we get "Canon in D (made famous
by this and that)" or The Well-Tempered Clavier Book I / Das
Wohltemperierte Clavier < Teil I / Le Clavier bien tempéré > Livre I /
Il clavicenbalo ben temperato Volume I" as a track titles.  But that
is, after all, what is actually on the liner.  If it's important
enough that we're willing to jump the gun on NGS, as it were, and go
straight to the best we can currently pull off in terms of work lists,
then let's stick to that principle, not adding, subtracting,
correcting, etc anything, except perhaps capitalization per the
language guidelines.  Chaotic?  Sure.  Messy?  Definately.  But that's
exactly why CSG came into being in the first place, as an attempt to
clean it all up.  Now we have ARs, and we can implement work lists
(though not quite as flexible of ones as we'll eventually have under
NGS).   Right?  :)

Now, I still care as far as not duplicating needlessly, so I still
think having 3, 4, 6, 10 listings for the same release, just in
different languages, is truly wasteful.  And assuming we do agree to
do something like a real work list and real linking ARs, regarding
track titles, I'm happy.  But I would hope, for what's left after we
then pull CSG out from releases, we can still find some way to avoid
this needless duplication, with all the additional maintenance and AR
work it implies.

As for going back to bit-by-bit rewriting the docs, I think it makes a
huge difference what we decide here.  If we do really decide to put
CSG titles into a works list, separated from the releases, then that
doc should be much more written for those doing those lists, and
cat-corner-coverage is then very important, while any
emphasis/reasoning behind making it "casual classical editor friendly"
is essentially also completely removed.  A separate doc, then, apart
from "work list CSG" would be needed to guide the person entering the
actual titles for a release, addressing the types of issues I
mentioned above.

Brian