Yesterday we found a bug that prevents the import of a post schema change update data set. We’ve pushed out a fix for this and tagged it with:
v-2012-05-15-import-fix
If you’re planning on importing a new data set, make sure to check out this tag, rather than the tag mentioned in this entry.
MetaBrainz Blog 
What is exactly the meaning of a « SCHEMA CHANGE » ?
I thought it was some technical database language something stuff like that.
But it seems that it meant some no-RFC undiscussed major ARTIST PAGE CHANGE.
Instead of SCHEMA CHANGE I think the blog should have said ARTIST PAGE CHANGE so that it would have probably be noticed before the change is done than after with all its drawbacks.
The artist page now shows stuff like pirate compilation among official albums.
While including irrelevant stuff, it excludes relevant stuff like official (mini) albums from the timeline.
Other people complain about OST or live being in there too.
I don’t care but it means more discussion before rather than after would have been betters.
GUI change’s importance is lowered but it’s the main stuff anyone can see when browsing MB.
I found a moreseriousbug than this that should require a fast revert.
David Bowie, for instance, now has 140 albums instead of 24.
Don’t try to find those 24 albums in MB any more since they are lost in the junk flood of hundreds of non albums.
+1 that.
This is a total mess guys. Heck, you might as well chuck singles, eps and other in there as well.
Can someone explain how this is a good idea?
The change was released because the underlying database changes needed to be released in this schema change, or in October, while the display can be changed at any time (and there was no time to work on it 😦 ). The current display is indeed a mess, and a better one should be released soon (hopefully before the end of May).
As of why the underlying change is a good idea, it both allows marking stuff as being several things at the same time (like a compilation of remixes, or a soundtrack EP) and easily introducing new ones (like “DJ Mix” which will be added soon and probably some more after it).
The biggest problem here, I’d say, is that unofficial releases are listed indistinctly with the official ones. Which is kinda stupid, but not new – we did it all the time. It just became more obvious now.
Is nothing planned to revert this ?
I’ve seen many shocked people but the official stand of MB seems to not care at all and being fine with the new mumbo jumbo.
Why is there no comment on this ?
It’s a pity the forum is down at this bad moment…
« The current display is indeed a mess, and a better one should be released soon (hopefully before the end of May). »
Where can we see the discussions about this ↑ ?
Annoyed people should be able to participate.
Once again, the forum is still down, maybe we have to wait for it to be online again (MBH-229) ?
This change apparently broke Sound Juicer–all track titles are coming up as blank.
This client had been broken for almost a year by a previous MB change. The client maintainers put a lot of work into getting it working again, and the fixed version just got into most users hands at the end of last month in a new Ubuntu release. So it worked for two weeks, and now is broken again.
I don’t understand why you don’t consider compatibility with existing clients when you make these changes.
I for one am starting to think I was a sucker to put a lot of work into entering releases and editing existing data over the past 2+ years. It’s clear you feel like you can do anything with MB that you want, with no regard for the people using it.
John, I’m sorry to hear that you feel tricked by MusicBrainz but I can assure you we do put an awful lot of effort in trying to maintain backwards compatibility. The bug that hit SoundJuicer (https://bugs.launchpad.net/ubuntu/+source/libmusicbrainz/+bug/1005075) was caused by it linking to a version of libmusicbrainz that couldn’t handle unknown elements in the XML. This is a fault of libmusicbrainz, and should not have been a problem for anything with slightly saner XML parsing. As noted in that ticket, a new version has been released which addresses this problem.
We have limited resources and try and consider clients and do communicate with authors of some apps, but we can’t reach everyone. Our lives would be made easier if we could get developers to test new features before we roll them out, and I hope we will have a better testing process in the next schema change release (in October).
John, I can only apologise for the issues you found with Soundjuicer. As acid2 says above, they were caused by some broken parsing in libmusicbrainz, as issue that has been present but unnoticed since its initial release some months ago.
Going forward, the issue should now be resolved, and from now on libmusicbrainz should just (correctly) ignore any unrecognised tags in any responses it sees.
Andy