well, given user a submits time of 1:20, and user b submits time of 1:22, and user c submits time of 1:21, we end up with an average time of 1:21 - I don’t see the issue....
2007-10-01 27412, 2007
dholmes
brianfreud: In that case it would make sense, but say all rips of all versions of an album result in either 1:20 or 1:22. Then either of those could be considered "correct", but 1:21 would be meaningless
2007-10-01 27422, 2007
ruaok
what if one user submits 1:20 one 2:15 and one 6:30?
2007-10-01 27423, 2007
dholmes
I don't think that's too uncommon a scenario
2007-10-01 27444, 2007
ruaok nods at dholmes
2007-10-01 27456, 2007
brianfreud
yes, but isn’t 1:21 then also the most “sensiblle”?
2007-10-01 27425, 2007
brianfreud
even muisicip uses a +/- 5s filter for puid generation...
2007-10-01 27459, 2007
slaad has quit
2007-10-01 27417, 2007
brianfreud
but given a choice between ?:?? and the avg of, say, 3 users’ local track times, I’d much rather the latter than the former
2007-10-01 27420, 2007
FauxFaux looks at /lastlog, thanks ruaok. :)
2007-10-01 27435, 2007
FauxFaux
brianfreud: Pink who?
2007-10-01 27421, 2007
brianfreud
lo;
2007-10-01 27401, 2007
brianfreud
MBChatLogger: off
2007-10-01 27401, 2007
MBChatLogger
is not logging
2007-10-01 27424, 2007
MBChatLogger
is logging
2007-10-01 27408, 2007
brianfreud has quit
2007-10-01 27426, 2007
brianfreud joined the channel
2007-10-01 27446, 2007
FauxFaux snickers at http://www.last.fm/user/Faux/charts/?charttype=weekly&subtype=artist, guess what I was listening to last week.
that is one project that is rising higher on my personal want list.
2007-10-01 27423, 2007
ruaok
better artist page.
2007-10-01 27435, 2007
warp
ah well. one thing at a time.
2007-10-01 27436, 2007
ruaok prods brianfreud
2007-10-01 27458, 2007
ruaok
yeah, I think redesigning that page would also have a drastic impact on server load.
2007-10-01 27413, 2007
ruaok
not that things are bad right now, but we could extend the life of our current db server.
2007-10-01 27414, 2007
warp
good or worse?
2007-10-01 27424, 2007
ruaok
good.
2007-10-01 27433, 2007
warp
ah :)
2007-10-01 27434, 2007
ruaok
because we would be serving less data overall.
2007-10-01 27447, 2007
ruaok
one page that is fatter, but gets cached.
2007-10-01 27456, 2007
ruaok
then data via JSON webservice as needed.
2007-10-01 27409, 2007
Jugdish joined the channel
2007-10-01 27412, 2007
ruaok
pull only what you need.
2007-10-01 27417, 2007
ruaok
works well for the xml ws.
2007-10-01 27420, 2007
ruaok
right Jugdish ?
2007-10-01 27427, 2007
Jugdish
exactly :P
2007-10-01 27430, 2007
ruaok
:-)
2007-10-01 27432, 2007
warp
yay, cdjapan is friendly.
2007-10-01 27406, 2007
Jugdish
ruaok: speaking of caching, what do you think about a HEAD feature for the xml ws? that is, a timestamp that tells you when the particular data that you want to retrieve was last updated
2007-10-01 27425, 2007
Jugdish
so like in picard, before pulling in an entire release with all its ARs, tracks, track ARs, etc. first you just ask when that data was last updated, if your tags are newer than that, no need for picard to retrieve any of the data
2007-10-01 27409, 2007
ruaok
it needs to happen.
2007-10-01 27421, 2007
Jugdish
heh
2007-10-01 27422, 2007
ruaok
we need to put timestamps on ALL the entities.
2007-10-01 27428, 2007
Jugdish
cool
2007-10-01 27447, 2007
ruaok
then we can also do RSS feeds and Changed-Since (or whatever) HTTP headers.
2007-10-01 27403, 2007
ruaok
once we do that, we can have google cache our pages. :)
2007-10-01 27421, 2007
Jugdish
and a way to propagate the timestamps of entities up to their descendants. like the last-updated time of an entire artist is the last-updated time of any of its albums, and so on
2007-10-01 27435, 2007
ruaok puts apr, timestamps on his todo list
2007-10-01 27449, 2007
ruaok
ohhh, cute idea.
2007-10-01 27417, 2007
ruaok
will you please remind me of that in an enhancement request? either that, or send us a patch. :)
2007-10-01 27422, 2007
Jugdish
sure sure
2007-10-01 27427, 2007
Jugdish
lol
2007-10-01 27439, 2007
ruaok
thanks for all those patches, btw
2007-10-01 27451, 2007
Jugdish
well i'll see if i can find some time to actually implement it, but at the very least i'll express my ideas on the matter in a ticket
2007-10-01 27458, 2007
Jugdish
sure! thanks for applying them :)
2007-10-01 27433, 2007
ruaok
while you're here, can I bug you to sanity check my tagging ui improvements?
There are a bazillion recordings of that work by different artists, all with the same track listings
2007-10-01 27417, 2007
dholmes
And that one, from FreeDB, has no performer associated with it, so there's no way whatsoever to know which it is, unless somebody has a disc with that disc id
2007-10-01 27443, 2007
dholmes
But somebody's associated PUIDs to it from Angela Hewitt's recording, which is already in MB elsewhere
2007-10-01 27442, 2007
dholmes
Can I just delete it? I can't think of any sane way to merge it with any other release