My code looks for artist by GID. It keeps finding 2 artists for one track.
2004-01-09 00906, 2004
melange
number 5
2004-01-09 00920, 2004
melange
it really doesn't like id'ing this track
2004-01-09 00907, 2004
yalaforge
night!
2004-01-09 00912, 2004
melange
'night
2004-01-09 00918, 2004
yalaforge has quit
2004-01-09 00921, 2004
rjmunro
I can't believe that the same MP3 file should not return the same TRM every time. I think we may have a fairly important tagger bug here.
2004-01-09 00942, 2004
ruaok
rjmunro: its not a tagger bug.
2004-01-09 00952, 2004
ruaok
Its an unfortunate sideeffect of some TRMs.
2004-01-09 00957, 2004
ruaok
TRM == black magic
2004-01-09 00904, 2004
rjmunro
But the linux tagger gives the same result every time?
2004-01-09 00906, 2004
ruaok
and black magic is not always deterministic.
2004-01-09 00912, 2004
rjmunro
Err, TRM tool, not tagger...
2004-01-09 00917, 2004
melange
rjmunro: yes, every time
2004-01-09 00932, 2004
Hashar has quit
2004-01-09 00933, 2004
ruaok
Hmmm. I didn't realize that.
2004-01-09 00945, 2004
ruaok
Ok, can you please mail me the track and I'll have a look?
2004-01-09 00953, 2004
melange
the tagger seems to have latched on to one trm now, but it's different from the linux one, and different from the one I submitted to the database earlier
2004-01-09 00901, 2004
djce
another silly quiz for people with nothing better to do: when looking at everyone's passwords, and seeing which ones are most often used, how many of the top ten passwords can you guess? :-)
2004-01-09 00914, 2004
melange
qwerty
2004-01-09 00916, 2004
melange
asdfgh
2004-01-09 00919, 2004
djce
melange: #3
2004-01-09 00920, 2004
melange
12345678
2004-01-09 00924, 2004
sward_3
password
2004-01-09 00926, 2004
djce
#1!
2004-01-09 00927, 2004
rjmunro
password
2004-01-09 00931, 2004
ruaok
musicbrainz
2004-01-09 00934, 2004
djce
#4
2004-01-09 00934, 2004
sward_3
blah
2004-01-09 00938, 2004
djce
nope
2004-01-09 00944, 2004
melange
god
2004-01-09 00945, 2004
melange
sex
2004-01-09 00952, 2004
djce
no, no
2004-01-09 00959, 2004
melange
whatever
2004-01-09 00904, 2004
djce
oh, sorry. melange: 12345678 = #8
2004-01-09 00905, 2004
ruaok
what is #2 and #5 ?
2004-01-09 00917, 2004
djce
ok, here are the top 10, in order.
2004-01-09 00924, 2004
djce
password. 123456. qwerty.
2004-01-09 00929, 2004
djce
musicbrainz. 12345.
2004-01-09 00933, 2004
sward_3
regarding different guids for the same mp3, the only reason that should happen is that somebody else tagged something between the times you queried, and it matches the new track "better"
2004-01-09 00936, 2004
djce
1234. music. 12345678.
2004-01-09 00947, 2004
djce
and rounding it out: trustno1. letmein.
2004-01-09 00902, 2004
melange
heh "trustno1"
2004-01-09 00917, 2004
sward_3
if you get that without a time space between queries, that _would_ be a bug =)
2004-01-09 00928, 2004
rjmunro
How many people use each one?
2004-01-09 00946, 2004
djce
actually, less than 1% for all of those.
2004-01-09 00948, 2004
melange
at least now I know what my next password will be
2004-01-09 00958, 2004
djce
" password" = 207 users, of 26161 total.
2004-01-09 00915, 2004
djce
I'm pretty surprised by that, actually.
2004-01-09 00923, 2004
djce
I thought it would have been nearer 10%
2004-01-09 00938, 2004
rjmunro
I think you need to add a "your password has expired, please change it" feature for everyone whose password is in the top 10.
2004-01-09 00948, 2004
djce
:-) not a bad idea!
2004-01-09 00904, 2004
melange
and give them a drop down list of the other nine ;)
2004-01-09 00928, 2004
melange
number 6
2004-01-09 00946, 2004
melange
I get a new trm if I quit the tagger and open it up fresh
2004-01-09 00952, 2004
melange
then load this track
2004-01-09 00910, 2004
melange
ruaok: priv me your email address ..
2004-01-09 00958, 2004
melange
on it's way
2004-01-09 00922, 2004
ruaok
thanks.
2004-01-09 00947, 2004
melange
wtf?
2004-01-09 00903, 2004
ruaok
# 7?
2004-01-09 00924, 2004
melange
linux tagger suddenly decided to give me another trm
2004-01-09 00926, 2004
melange
go figure
2004-01-09 00939, 2004
ruaok
that makes me less worried now. :-)
2004-01-09 00941, 2004
melange
after giving me the same one 20 times in a row
2004-01-09 00919, 2004
melange
does it depend on which way the wind is blowing or something?
2004-01-09 00923, 2004
melange
:)
2004-01-09 00930, 2004
melange
bloody voodoo
2004-01-09 00938, 2004
ruaok
You got it!
2004-01-09 00945, 2004
ruaok
bloody voodoo!
2004-01-09 00916, 2004
melange
it's consistently giving me this other trm now
2004-01-09 00932, 2004
ruaok
melange: you better stop. it will drive you crazy. :-)
2004-01-09 00935, 2004
melange
all I did when it changed was check another track's trm
2004-01-09 00943, 2004
melange
argh!
2004-01-09 00901, 2004
melange
maybe we need a way to flag some songs as "generates useless trms"
2004-01-09 00948, 2004
djce
and I thought it was deterministic... shows what I know.
2004-01-09 00940, 2004
melange
oooh
2004-01-09 00953, 2004
melange
the linux and the windows tag match up
2004-01-09 00909, 2004
melange
quick, export it to the database!
2004-01-09 00910, 2004
melange
;)
2004-01-09 00903, 2004
ruaok
melange: do you know anything about multidimensional pattern matching algorithms?
2004-01-09 00919, 2004
ruaok
In this case 133 dimensions.
2004-01-09 00923, 2004
melange
other than being a tongue twister, not the foggiest
2004-01-09 00935, 2004
ruaok guesses. sward_3 won't tell him.
2004-01-09 00959, 2004
ruaok
I could attempt to explain what is going on and what you're doing, but...
2004-01-09 00959, 2004
melange
I'm wondering .. it's pretty quiet for the first 25 seconds or so, and it's vbr, so it would be sampling at low bitrate ..
2004-01-09 00917, 2004
melange
and as the trm only cares about the first 30 seconds ..
2004-01-09 00921, 2004
ruaok
ARG!
2004-01-09 00931, 2004
ruaok
You're giving it nothing to work with.
2004-01-09 00951, 2004
melange
still, it managed to deal with a 9 second silent track consistently
2004-01-09 00954, 2004
ruaok
The less signficant the data in the first 30 seconds, the less reliable the TRM will be.
2004-01-09 00912, 2004
rjmunro
But it should be constant with the same data, shouldn't it?
2004-01-09 00920, 2004
melange
yes
2004-01-09 00946, 2004
ruaok
yes and no.
2004-01-09 00955, 2004
rjmunro
Can you add the feature to the tagger that tells the server that the TRM is wrong?
2004-01-09 00905, 2004
ruaok
Since there is so little significant data in this track the data values are likely to be small.
2004-01-09 00910, 2004
rjmunro
It would be good to see some stats on that.
2004-01-09 00914, 2004
ruaok
And windows and linux have different rounding errors.
2004-01-09 00935, 2004
ruaok
and the TRM algorithm will probably scale the small range of value to a large domain.
2004-01-09 00940, 2004
ruaok
er range.
2004-01-09 00914, 2004
ruaok
and thus the roundoff errors get exaggerated, which can account for different TRMs on different machines.
2004-01-09 00935, 2004
melange
at different times of the day
2004-01-09 00900, 2004
djce guesses CPU core temperature
2004-01-09 00905, 2004
ruaok
even that I could explain
2004-01-09 00922, 2004
melange
cosmic rays
2004-01-09 00927, 2004
ruaok
the lookup 'dataset' probably gets influenced by other sketchy tracks being looked up.
2004-01-09 00942, 2004
ruaok
and being inserted into the TRM system.
2004-01-09 00956, 2004
ruaok
So if you were the only person using it, you'd get more consistent results.
2004-01-09 00913, 2004
melange
hmm
2004-01-09 00927, 2004
melange
I think "voodoo" was probably a good enough explanation
2004-01-09 00929, 2004
melange
:D
2004-01-09 00954, 2004
ruaok
I'm going to disappear for a bit. Back soon.
2004-01-09 00901, 2004
ruaok has left the channel
2004-01-09 00923, 2004
nobody431 has quit
2004-01-09 00931, 2004
melange
I wonder how you could determine if a track will produce TRMs like this that aren't useful?
2004-01-09 00946, 2004
melange
without running the TRM generator over it many times, that is
2004-01-09 00940, 2004
djce plays with more random SQL, and actually finds something interesting this time.
2004-01-09 00951, 2004
djce
I ran the "TRM collisions" report...
2004-01-09 00910, 2004
djce
the top TRM had 48 tracks and no lookups
2004-01-09 00921, 2004
djce
and of those 48, that's only over three artists
2004-01-09 00926, 2004
djce
and lots of different songs.
2004-01-09 00938, 2004
djce
So, I looked up which moderations those TRMs came from.
