hibiscuskazenek1: yeah, I get search errors in beta too - probably index building stuff and should be temporary but I've asked in case
2015-07-29 21023, 2015
reosarevok
search seems back, is stuff better for you too?
2015-07-29 21001, 2015
mr_maxis has left the channel
2015-07-29 21059, 2015
hibiscuskazenek1
I'm still getting errors, lemme try again
2015-07-29 21029, 2015
hibiscuskazenek1
nope, still not working
2015-07-29 21043, 2015
hibiscuskazenek1
(I'm running beta in the tab I'm using)
2015-07-29 21029, 2015
pankkake
http://coverlib.com/stats/ they manage to have a lot more kb per image, but I'm not surprised here (very high quality usually)
2015-07-29 21047, 2015
pankkake
2 TB of cover art, still quite manageable for a single user
2015-07-29 21025, 2015
hibiscuskazenek1
I don't think I have that much HD space
2015-07-29 21025, 2015
pankkake
these days drives <2TB are boring to me :)
2015-07-29 21059, 2015
gioele
pankkake: do you know what is the legal status of those cover arts? Could they be uploaded to MB/Archive.org?
2015-07-29 21028, 2015
pankkake
the coverlib ones? I don't think they have a better (or any) legal status
2015-07-29 21003, 2015
pankkake
apparently incorportated in Germany
2015-07-29 21031, 2015
gioele
pankkake: somebody should get in contact. For example, the very latest cover uploaded to coverlib is a 2000x2000px scan + scans of the booklet. The cover for the same release in MB is a 1000x1000px scan with no booklet images
2015-07-29 21033, 2015
pankkake
I use coverlib to find artwork, but I don't really see how more could be done
2015-07-29 21052, 2015
pankkake
because they do not really identify individual releases, you have to actually look at the cover art details
2015-07-29 21037, 2015
pankkake
perhaps could be in guide for finding good images, next to "don't use google images" ;)
2015-07-29 21020, 2015
gioele
pankkake: I was thinking of a report. There was a report that "guessed" relations between a wikipedia page and a release. You had to check manually, but if it was correct it would create the relation for you in two clicks. Something similar could be done here. You get a list of guesses, if you find an exact match you click somewhere and the cover art edit page is prefilled with the covers found on coverlib
2015-07-29 21056, 2015
pankkake
ah, yes, that could work
2015-07-29 21051, 2015
xps2 joined the channel
2015-07-29 21028, 2015
reosarevok
gioele: remember that needs to be done without the covers ever touching MB servers
2015-07-29 21044, 2015
reosarevok
(if that's doable though, it might be interesting)
2015-07-29 21004, 2015
DarkAudit joined the channel
2015-07-29 21005, 2015
gibbon_
I guess the covers could be loaded by the client from the appropriate page at coverlib... If its OK for the metadata to be processed, I guess that should be enough to suggest cover art for a release, shouldn't it?
2015-07-29 21045, 2015
gibbon_
MB Servers don't have anything to compare the coverart to anyway, so processing them would not be necessary
2015-07-29 21054, 2015
hibiscuskazeneko joined the channel
2015-07-29 21007, 2015
reosarevok doesn't really know what's enough or not to be honest, I'm not sure how much or how little we're supposed to do with the stuff
2015-07-29 21009, 2015
gibbon_
but thats a person talking who knows nothing about MB infrastructure and should keep quiet on that matter
2015-07-29 21044, 2015
reosarevok
I would suspect that we wouldn't want to add such a thing to MB directly but it could work as an external tool
2015-07-29 21049, 2015
gibbon_
My general opinion though is that these relations would be dangerously easy to get wrong
2015-07-29 21016, 2015
reosarevok
(although if the other site gave specific permission then it might work? dunno)
2015-07-29 21021, 2015
gibbon_
but thats probably true for a lot more
2015-07-29 21011, 2015
gibbon_
yes, I guess the right people should get in touch and talk about if something like that would be ok and possible
2015-07-29 21041, 2015
pankkake joined the channel
2015-07-29 21052, 2015
Afterster joined the channel
2015-07-29 21011, 2015
obXiDeJSFQ joined the channel
2015-07-29 21016, 2015
gioele
a greasemonkey script like those that exist for bandcamp or discogs would be a nice way to implementi it
2015-07-29 21039, 2015
gibbon_
yep
2015-07-29 21030, 2015
reosarevok
Yeah, a script would definitely be fine - it would still have to download the images first though since right now we don't have any code to get the images from a URL
2015-07-29 21059, 2015
reosarevok
(it works in Chrome in Windows for some reason because of how it deals with pasting URLs in the "open file" dialog, but it's not an intentional feature)
2015-07-29 21021, 2015
reosarevok
There's a ticket for that, but given the focus is in community things at the moment I doubt it'll be done soon unless a volunteer does it
2015-07-29 21022, 2015
reosarevok
(also I dunno if it would need changes on the IA side to work, probably would...)
