#musicbrainz-devel

/

      • hawke_ joined the channel
      • 2010-11-30 33419, 2010

      • ocharles
        luks, warp - wouldn't mind feedback on https://gist.github.com/722090
      • 2010-11-30 33427, 2010

      • ocharles
        just messing about with ideas for a new edit schema
      • 2010-11-30 33421, 2010

      • warp
        yikes, the commented parts on gist.github are unreadable.
      • 2010-11-30 33453, 2010

      • ocharles
        heh
      • 2010-11-30 33411, 2010

      • ijabz joined the channel
      • 2010-11-30 33444, 2010

      • djce joined the channel
      • 2010-11-30 33414, 2010

      • ijabz joined the channel
      • 2010-11-30 33413, 2010

      • luks
        ocharles: well, it seems like you are designing something from a technical point of view
      • 2010-11-30 33425, 2010

      • luks
        there are high-level issues with the edit system
      • 2010-11-30 33411, 2010

      • ocharles
        yes, i've tried to solve the high level issues that I'm aware of
      • 2010-11-30 33414, 2010

      • ocharles
        but i'm not a heavy editor
      • 2010-11-30 33458, 2010

      • luks
        then you should take the most complex use case, changing a release (including AR credits)
      • 2010-11-30 33424, 2010

      • luks
        if you fix how ARs are voted on, you fix at least 80% of all problems
      • 2010-11-30 33458, 2010

      • ocharles
        well i guess in this each AR change/addition/deletion is an edit, and they are all grouped together. a group is atomic - it either passes or fails
      • 2010-11-30 33427, 2010

      • ocharles
        if an editor doesn't like a certain change, they can let the original editor know, and they can revision certain edits. that would probably invalidate all votes and require a revote or something
      • 2010-11-30 33455, 2010

      • ocharles
        or, voting is done per edit so some of the group can pass and some can fail
      • 2010-11-30 33403, 2010

      • ocharles
        i dunno which works better
      • 2010-11-30 33419, 2010

      • luks
        I don't know
      • 2010-11-30 33424, 2010

      • luks
        I recently started to think that having generic ARs in MB was a mistake (from user perspective)
      • 2010-11-30 33400, 2010

      • ocharles
        and it would be better to have something more freeform?
      • 2010-11-30 33403, 2010

      • ocharles
        or less freeform?
      • 2010-11-30 33407, 2010

      • luks
        so if I were to redesign the edit system, I'd probably try to hide them (show them as "credits", "members", "urls", "...")
      • 2010-11-30 33413, 2010

      • ocharles
        yea, we're moving towards that
      • 2010-11-30 33437, 2010

      • ocharles
        the new relationship stuff will get rid of the "has/had" terminology and lend itself better to being integrated in other places
      • 2010-11-30 33416, 2010

      • ocharles
        showing band members in the side bar for example - going a step further editing an artist would show a set of "members" fields, if the type is a group. I guess this is the kind of thing you're thinking of
      • 2010-11-30 33433, 2010

      • luks
        yeah, but now we are talking about editing
      • 2010-11-30 33404, 2010

      • luks
        I'm not sure if makes sense to expose them as "ARs" while editing something
      • 2010-11-30 33439, 2010

      • ocharles
        I'm not exposing them as ARs during editing - see above, they would just show up as a set of text fields (autocomplete) to choose the member artists. maybe with founding checkboxes and begin/end dates
      • 2010-11-30 33452, 2010

      • ocharles
        but never mentioning them as ARs really, or letting the user choose the type here
      • 2010-11-30 33412, 2010

      • ocharles
        i dunno if you consider that still too much like ars=
      • 2010-11-30 33445, 2010

      • luks
        well, in the mockup you talk about object graphs
      • 2010-11-30 33451, 2010

      • luks
        I think that's one of the main problems
      • 2010-11-30 33459, 2010

      • luks
        and one of the main places where ARs show up
      • 2010-11-30 33420, 2010

      • luks
        I'd try to flatten everything in the edit system
      • 2010-11-30 33449, 2010

      • ocharles
        i'm not sure how that backend representation matters
      • 2010-11-30 33452, 2010

      • luks
        e.g. release would consist of all the mediums, tracklists, tracks, credits
      • 2010-11-30 33400, 2010

      • luks
        it does
      • 2010-11-30 33409, 2010

      • luks
        editing a graph is HARD
      • 2010-11-30 33420, 2010

      • ocharles
        no argument there, but I'm not suggesting graph editing really
      • 2010-11-30 33438, 2010

      • luks
        yeah, but that's what we do right now
      • 2010-11-30 33450, 2010

      • luks
        and the mockup doesn't seem to be changing that
      • 2010-11-30 33459, 2010

      • ocharles
        hm
      • 2010-11-30 33427, 2010

      • luks
        I'd take something as stupid as the discogs edit system as a reference
      • 2010-11-30 33431, 2010

      • ocharles
        Well no, I guess it doesn't - this is just simplifying how we store the data, and making sure everything is valid for historical
      • 2010-11-30 33437, 2010

      • ocharles
        purposes*
      • 2010-11-30 33438, 2010

      • luks
        you have edit artist, edit release, edit label
      • 2010-11-30 33440, 2010

      • luks
        nothing else
      • 2010-11-30 33443, 2010

      • ocharles
        right
      • 2010-11-30 33423, 2010

      • luks
        I believe it can be done in MB as well
      • 2010-11-30 33431, 2010

      • luks
        it will remove some flexibility
      • 2010-11-30 33439, 2010

      • luks
        but it would be much more usable
      • 2010-11-30 33446, 2010

      • hawke_ joined the channel
      • 2010-11-30 33450, 2010

      • ocharles
        So what do you mean by flattening? You mean having less edits?
      • 2010-11-30 33430, 2010

      • ocharles
        The way I see this is: a group is like a transaction, an edit is like an operation.
      • 2010-11-30 33459, 2010

      • ocharles
        so you only enter what is effectively edit artist, create artist, edit release etc. but the implementation splits them apart, because that's easier for me to work with
      • 2010-11-30 33403, 2010

      • ocharles
        and operations are composable
      • 2010-11-30 33432, 2010

      • luks
        I don't see an easy way to combine them to see the end result
      • 2010-11-30 33439, 2010

      • ocharles
        the edit artist "group" is a composition of an edit artist edit, and maybe some changes to ar links
      • 2010-11-30 33410, 2010

      • luks
        but if I have *everything* about a release in a single edit, I can preview it, revert it, etc.
      • 2010-11-30 33430, 2010

      • luks
        I'd probably even consider the entities to be "wiki pages"
      • 2010-11-30 33408, 2010

      • luks
        it might look weird, but it seems to work more than fine for wikipedia
      • 2010-11-30 33422, 2010

      • ocharles
        ok, I see the argument here (and some of it is doable with split edits, some is not) - but now how are you going to handle voting?
      • 2010-11-30 33423, 2010

      • luks
        they have all kinds of structured metadata on a plain-text page
      • 2010-11-30 33433, 2010

      • ocharles
        because "edit release" is going to be huge
      • 2010-11-30 33449, 2010

      • luks
        is that a problem?
      • 2010-11-30 33453, 2010

      • ocharles
        wikipedia works because it doesn't have voting
      • 2010-11-30 33455, 2010

      • luks
        at least it's revertable
      • 2010-11-30 33440, 2010

      • ocharles
        The idea of edit revisions and large edits could work though
      • 2010-11-30 33443, 2010

      • luks
        and I think that if we have 2k edits instead of 30k edits, more edits would get voted on
      • 2010-11-30 33448, 2010

      • luks
        so they would pass faster
      • 2010-11-30 33404, 2010

      • ocharles
        So I insert one big edit release edit. someone has a problem with 1 tiny bit, I create a new revision of my edit and address that, and then the edit can pass
      • 2010-11-30 33441, 2010

      • ocharles
        does that sound sensible?
      • 2010-11-30 33425, 2010

      • hawke_
        Isn’t this a similar problem to content management’s “current approved version” vs. “draft version” things like what is done for http://musicbrainz.org/doc/ ?
      • 2010-11-30 33430, 2010

      • luks
        yes, probably, maybe, I don't know :)
      • 2010-11-30 33442, 2010

      • luks
        I've been thinking about MB editing a lot
      • 2010-11-30 33443, 2010

      • ocharles
        :)
      • 2010-11-30 33450, 2010

      • luks
        and I'm still not sure what's the best approach
      • 2010-11-30 33456, 2010

      • luks
        the problem is that there *is* an object graph
      • 2010-11-30 33400, 2010

      • luks
        there is no way around that
      • 2010-11-30 33414, 2010

      • luks
        it's a matter of drawing the line between entities
      • 2010-11-30 33422, 2010

      • luks
        I'm not sure where that line is
      • 2010-11-30 33427, 2010

      • hawke_
        If we’re just trying to get the outstanding edits count down, I think the first thing would be to shorten the time for edit approval on things of “normal” data qaulity.
      • 2010-11-30 33431, 2010

      • hawke_
        *quality. :-D
      • 2010-11-30 33456, 2010

      • luks
        it's more about "useful" edit system
      • 2010-11-30 33409, 2010

      • luks
        if you have 100 AR edits, it's not easy to vote on
      • 2010-11-30 33420, 2010

      • luks
        you either don't vote, or blindly vote yes on all of them
      • 2010-11-30 33402, 2010

      • hawke_
        Mostly “don’t vote”.
      • 2010-11-30 33449, 2010

      • hawke_
        That’s the other reason to shorten the vote time: so few people are likely to know or care about any given edit enough to vote on it.
      • 2010-11-30 33437, 2010

      • hawke_
        [that it makes the voting system almost useless as is]
      • 2010-11-30 33444, 2010

      • ocharles
        luks: large edits look interesting. http://gist.github.com/722499
      • 2010-11-30 33458, 2010

      • ijabz joined the channel
      • 2010-11-30 33429, 2010

      • ijabz joined the channel
      • 2010-11-30 33436, 2010

      • murdos
        ijabz: ping
      • 2010-12-01 33516, 2010

      • automator joined the channel
      • 2010-12-01 33550, 2010

      • automator
        luks: I created a JIRA about retrieving pre-NGS collaboration artists @ jira.musicbrainz.org/browse/MBS-1123
      • 2010-12-01 33523, 2010

      • dinog1 joined the channel
      • 2010-12-01 33508, 2010

      • ijabz_ joined the channel
      • 2010-12-01 33548, 2010

      • djce joined the channel
      • 2010-12-01 33509, 2010

      • ijabz joined the channel
      • 2010-12-01 33525, 2010

      • nikki
        luks: ping
      • 2010-12-01 33527, 2010

      • ijabz
        murdos: ( a rather delayed) pong
      • 2010-12-01 33551, 2010

      • murdos_ joined the channel
      • 2010-12-01 33518, 2010

      • murdos_
        ijabz: could you remind me why we have some much artist field for release/recording/release-group/work?
      • 2010-12-01 33520, 2010

      • murdos_
      • 2010-12-01 33542, 2010

      • murdos_
        ARTIST_NAME, ARTIST, ARTIST_NAMECREDIT
      • 2010-12-01 33519, 2010

      • murdos_
        it's really confusing and not really needed IMO
      • 2010-12-01 33549, 2010

      • ijabz
        Iit was added as a response to a gripe you had on this page http://wiki.musicbrainz.org/?title=User:Murdos/NG…, (that youve now deleted) to do with artist credits
      • 2010-12-01 33518, 2010

      • ijabz
        Some fields are required for searching, and some for returning the data
      • 2010-12-01 33507, 2010

      • ijabz
        this page explains the search fields http://wiki.musicbrainz.org/Next_Generation_Schem…
      • 2010-12-01 33500, 2010

      • murdos_
        we just need 3 fields: ARTIST (including all different names) + ARTIST_ID for searching, and storing ARTIST_CREDIT for output
      • 2010-12-01 33501, 2010

      • ijabz
        If you don't have a separate field for individual artists, and the complete artist credit for the album you cannot do exact matches properly because you dont know what the artist is for the album
      • 2010-12-01 33548, 2010

      • ijabz
        They are necessary thats why we put them in, and I don't see any advantage in removing size, (index creation time and index time will be affect will be neglible)
      • 2010-12-01 33544, 2010

      • murdos_
        I just extremely confusing... Although I'm one of the dev on the search server, I just says WTF when I got not results from http://test.musicbrainz.org/search?query=artist%3…
      • 2010-12-01 33521, 2010

      • murdos_
      • 2010-12-01 33531, 2010

      • murdos_
        and "artist", "artistname" and "creditname" are not really clear
      • 2010-12-01 33552, 2010

      • murdos_
        there's a point where, if you want exact matches properly, you'd better use a database
      • 2010-12-01 33554, 2010

      • ijabz
        artist credits ARE confusing :)
      • 2010-12-01 33536, 2010

      • ijabz
        but the thoase releases a re credited to renaud not renaud3 so why should renaud3 return matches for the default artist field
      • 2010-12-01 33509, 2010

      • murdos_
        but we can perfectly hide this complexity in the search server
      • 2010-12-01 33519, 2010

      • murdos_
        ijabz: renaud == renaud3
      • 2010-12-01 33527, 2010

      • ijabz
        Yes but one is an alias or somthing
      • 2010-12-01 33504, 2010

      • murdos_
        no, "renaud3" is the actual artist name, "renaud" is the specific credit name used for recordings
      • 2010-12-01 33558, 2010

      • ijabz
        You can hide the complexity in the user interface, if you want to search both you just search both fields, once you merge into one field youve lost that information
      • 2010-12-01 33559, 2010

      • murdos_
        (because I renamed the artist "Renaud" to "Renaud3": http://test.musicbrainz.org/edit/13577883)
      • 2010-12-01 33545, 2010

      • murdos_
        sorry, I've to go, but I hope we can clear this issue later
      • 2010-12-01 33504, 2010

      • ijabz
        phew, I also need to get on with someting else
      • 2010-12-01 33518, 2010

      • dinog joined the channel
      • 2010-12-01 33552, 2010

      • kurtjx joined the channel
      • 2010-12-01 33501, 2010

      • ocharles
        warp: I might have to lump a bit of JavaScript work on you
      • 2010-12-01 33518, 2010

      • ocharles
        I've done the work to create new artists and labels, and for the most part it works - but the display is kinda messed up :)
      • 2010-12-01 33513, 2010

      • ocharles
        At the moment it creates artists if you leave the artist field blank and add a artist credit name. we should probably have an option in the drop down called "add new artist" which fills in the artist credit name and does something special to the artist selection box (maybe italic text saying create new artist or something)
      • 2010-12-01 33551, 2010

      • kurtjx joined the channel
      • 2010-12-01 33557, 2010

      • murdos_ joined the channel
      • 2010-12-01 33545, 2010

      • ijabz
        murdos_I thought about your example some more, it is very contrived, I think in reality if you were to change the artist name you might expect the credit names to also change and that is the problem.
      • 2010-12-01 33507, 2010

      • ijabz
        Where the artist name and credit name differ it will be important in some cirumstances to search for these individually