#musicbrainz

most of the info accrues incrementally anyways

I think the idea is to use it as a disambiguator, no?

like people fill in more deatils as time goes on

details*

hawke_: yes. So it'd work more or less like the way we do it now. Trust editors not to make overly long "feat." parentheticals

most people making classical entries are familiar with the minimal closure rule

i.e, just enough info on release titl to avoid ambiguity

title*

anyway, if search worked worth a dam "feat." wouldnt be necessary

you do a google search with enough terms and right recording comes to the top

mchou: yeah, I agree with the "just enough" approach.

the disjoint tables of MBZ when it comes to search is really aggravating

I (still) disagree with the idea that "feat." was ever meant to overcome limitations in the search. but maybe that's a distinction without a difference.

caller_6: I dunno what you mean

caller_6: you mean it doesn't help search or something else?

From everything I've read, "feat." is meant as a /human readable/ disambiguation. Not a /machine readable/ one. I admit, it does help the search function. I just don't think it was /intended/ to.

"was meant" I should have said.

caller_6: well, I disagree with you. I think "feat. was intended to to disambiguate and simplify serch

search*

I’m pretty sure it was just there to make an artist page be all in one place

if MBZ didnt force disjoint search then feat. wouldn't have been necessary

hawke_: what?? That makes no sense what so ever

So you wouldn’t have to look at 10 different artist pages to see everything by an artist.

unless you want to see it in your track name

mchou: I didn’t say I agreed with it.

hawke_: how's that any different from disabiguation and ease of search?

hawke_: that's a disnction w/o a difference

distinction*

mchou: It’s a different purpose which happens to have the same effect.

I dunno how you search but I certainly dont do a primary search by artist

I do a search like "Beethoven fifth symphony Kleiber" and the right recording comes up

that's why "feat" is useful

mchou: I tend to look at artist pages, so…

otherwise you spend all day wading through "Beethoven fifth symphony"

mchou: I think the current thinking is that we'll move(feat. [performer]) from the Title field to the Release/RG Comment field. When that comes up on the mailing list, I'll ping (if you feel strongly about it)

caller_6: i'm saying fix MBZ search is HIGHER priority

caller_6: Does Release/RG make sense? I would think recordings make more sense…

mchou: Perhaps if there are problems with the search, you should file tickets against that?

hawke_: the problem is nobody thinks search is high priority enough to fix

mchou: that's news to me

so this is the workaround

ocharles: when MBZ search becomes as good as google you let me know :)

mchou: would a friendlier "advanced search syntax" be enough?

no

there is no need for advanced search syntax

mchou: “make search be as good as google” is not a very good starting point

google doesnt use it

why should MBZ?

hawke_: why not?

server load I assume...

mchou: Because it doesn’t give a useful place to work from.

it doesn't have much to do with server load

If you say “I search for this and I don’t find that, and I feel that I should”

hawke_: I just gave you an example

“Beethoven fifth symphony kleiber”?

"Beethoven fifth symphony Kleiber"

yup

Right, and that’s arguably a good example, and maybe a ticket should be entered about that

come on man

how are you going to structure the ticket?

hawke_: umm, that's the title on the album cover

so why would the search be so "foreign?"

mchou: Cover of what?

of that recording

mchou: …which release

Beethoven: Symphony No. 5 Carlos Kleiber

mchou: I see http://musicbrainz.org/release/9e097bfc-e792-42... and http://musicbrainz.org/release/e1bb14ae-cb3d-45...

that's the same recording session

just because it's on separate discs hardly matters

hop, french team has now 3 members ! and fully translated :)

the point is it narrows down t 2 where otherwise you'd wade through pages of "beethoven Symphony no. 5"

s/t/to/

mchou: So that’s how I’d structure the ticket: “I search recordings for ‘beethoven fifth symphony kleiber’ and the results should include [recording link] at the top, but they don’t”

or releases, or whatever it is you’re searching

hawke_: It came on top because of (featured)

not because MBZ is good

MBZ search*

mchou: it doesn’t for me.

I was comparing size of similar responses from ws1 and ws2, and in some cases the latter is 5 times bigger

I think json would help here

mchou: Furthermore, if those recordings are the same they should presumably be merged.

And have related works, performers, conductors, all that good relationship stuff.

hawke_: they probably were merged at one point

mb search cannot search for what isn’t there.

mchou: http://musicbrainz.org/recording/2f9f7d83-c49c-... says not

NGS has a funny habit of separating re-releases

hawke_: what?

mchou: re-releases *should* be separated

I'm reading that link

But the recordings should be the same, probably

(unless they're remasters)

I dont see why re-releases should be separated

before ngs re-releases made sense

now re-releases are all over the place

Because they're a different release

mchou: If they had “is a later release of” relations pre-NGS, then the recordings are the same in NGS.

let's say disc A first released in 1960, and another release in 1965

There were a lot of missing “is a later release of” relations though.

And if they were together as one release with several release events, then they're the same too

in NGS they are two separate entities

pre NGS it was same entity (with different release date)

so what happened?

mchou: release events went away, and became releases instead.

hawke_: well, do you think that makes sense?

mchou: yes.

hawke_: why?

pre-NGS it was arbitrary

if the re-release happenend to have an extra track, you had to add a new release

mchou: First, because now disc IDs can be associated with a release.

if there wasn't a bonus track, you could add a release event

luks: I'm talking instances where disc A is exactly the same (down to disc id)

mchou: (if it helps) there was some talk on the mailing list about giving an NGS Release multiple country/date fields to combine otherwise identical Releases (i.e. barcode, cat# the same)

just different release dates

mchou: but there are many re-releases where that's not the case

so they had to be treated differently

luks: right, I'm not talking about those.

I'm talking (exact same) re-release

I'm talking about rereleases in general, making things consistent is one of MB's main features

the situation was anything but consistent

mchou: and second, because release groups now actually show all releases, instead of just showing *some* releases only if they happen to have different track lists

hi

hi Shisma

I have an internal server error

Internal Server Error

Oops, something went wrong!

Error message: (No details about this error are available)

We're terribly sorry for this problem. Please wait a few minutes and repeat your request — the problem may go away.

If the problem persists, please report a bug and include any error message that is shown above.

hawke_: I don't think you understand my point

mchou: fwiw, http://musicbrainz-mailing-lists.2986109.n2.nab...

Shisma: and what are you trying to do when you get it?

add an release

caller_6: I'd say that's a different situation

mchou: perhaps not.

hawke_: pre-ngs, I can tell you for same exact re-release, they belonged in the same release group (just different release dates)

now they are all separate entities

mchou: and that’s still true.

they still belong in the same release group

totally confusing and messed up

reosarevok: i guess the problem is on the "Add Missing Entitis"-steo

hawke_: that's not true

mchou: It’s really simple: each release gets its own Release.

I'll upload a screenshot

(unless the only difference is the label/cat no, like is done for stereo/mono releases at times)

hawke_: in classical that makes no sense