the issue with playdar is, of course, "/api/?method=resolve&artist=Joe%20Satriani&album=&track=Surfing%20with%20the%20Alien"
2012-05-18 13903, 2012
reosarevok
I don't know exactly how to do it
2012-05-18 13906, 2012
hawke_1
reosarevok: Yeah, but then we have a range of ranges
2012-05-18 13907, 2012
reosarevok
If not, I'd have asked for it
2012-05-18 13911, 2012
hawke_1
which would be *ugly*
2012-05-18 13916, 2012
kepstin-work
only supports doing artist + album + track name text matches :/
2012-05-18 13934, 2012
reosarevok
hawke_1: would it?
2012-05-18 13941, 2012
jcazevedo joined the channel
2012-05-18 13943, 2012
reosarevok
Sounds like a display problem
2012-05-18 13947, 2012
reosarevok
Conceptually, it's not ugly
2012-05-18 13900, 2012
hawke_1
reosarevok: You’d need to have an end date for the beginning and a begin date for the beginning
2012-05-18 13920, 2012
hawke_1
and then the same for the ending
2012-05-18 13923, 2012
reosarevok
Sure
2012-05-18 13940, 2012
reosarevok
Hide them by default, a checkbox "approximate dates" opens that
2012-05-18 13946, 2012
reosarevok
Dunno
2012-05-18 13948, 2012
hawke_1
I suppose that would work
2012-05-18 13910, 2012
hawke_1
How about allowing substitution of ? or X for part of the date?
2012-05-18 13912, 2012
hawke_1
e.g. 195X
2012-05-18 13917, 2012
hawke_1
or 195?
2012-05-18 13921, 2012
reosarevok
Not too good
2012-05-18 13923, 2012
hawke_1
ick, not the ?
2012-05-18 13941, 2012
reosarevok
If you know it's 1949 to 1951, you need to say it's the whole 20th century ;)
2012-05-18 13903, 2012
hawke_1
This is why I haven’t proposed circa. :-)
2012-05-18 13931, 2012
reosarevok
I still think ranges inside ranges are the most sensible option
2012-05-18 13934, 2012
reosarevok
Even if not a pretty one
2012-05-18 13955, 2012
brianfreud
hawke_ what about the other part of that which always has bothered me to have it missing - "fall 1959", "early 1922", "spring 1930", "spring or early summer of 1785"
2012-05-18 13907, 2012
kepstin-work
1949-1951–1979
2012-05-18 13914, 2012
reosarevok
brianfreud: ranges too?
2012-05-18 13931, 2012
reosarevok
"fall 1959" - "1959 9 - 1959 12"
2012-05-18 13933, 2012
reosarevok
Or whatever
2012-05-18 13944, 2012
kepstin-work notes that you'll only be able to understand that if your irc client font uses variable-width font with good dash characters.
2012-05-18 13945, 2012
reosarevok isn't even sure when fall starts and ends
2012-05-18 13955, 2012
brianfreud
'he started composing it some time between the fall of 1899 and the spring of 1901, and completed it in late spring of 1905"
2012-05-18 13935, 2012
reosarevok
brianfreud: you enter approximated date ranges, and then you don't even need to understand the concept of "spring" for it to work :)
2012-05-18 13907, 2012
reosarevok is not sure if "spring" translates well or not - while a calendar is just math
2012-05-18 13918, 2012
kepstin-work suspects that spring/summer/fall/winter is pretty universal for cultures in temperate climates, but other places may be split more along the lines of "dry season/rainy season" :)
2012-05-18 13929, 2012
reosarevok
Yeah
2012-05-18 13944, 2012
reosarevok
And in 100 years, the concepts might have changed
2012-05-18 13954, 2012
reosarevok
But numbers probably won't
2012-05-18 13904, 2012
noobie
style question
2012-05-18 13915, 2012
reosarevok
Wear the tie. Bowties look silly
2012-05-18 13924, 2012
brianfreud
'he started composing it some time between the fall of 1899 and the spring of 1901, and completed it in late 1905"
2012-05-18 13926, 2012
noobie
Phantom Vibrations, Parts 1-3
2012-05-18 13935, 2012
hawke_1
including seasons or multiple ranges in one AR feels like trying to do too much.
2012-05-18 13937, 2012
noobie
or Phantom Vibrations (Part 1-3)
2012-05-18 13954, 2012
hawke_1
I’d be happy with a basic “circa” checkbox for starters
2012-05-18 13959, 2012
reosarevok
I'd say ", Parts 1-3"
2012-05-18 13905, 2012
reosarevok
hawke_1: I don't see why
2012-05-18 13909, 2012
noobie
that's what guess case suggests^^
2012-05-18 13910, 2012
kepstin-work
hawke_1: I'd call the checkbox "approximate"
2012-05-18 13912, 2012
brianfreud
Hell, I'd be happy with BC support :P
2012-05-18 13944, 2012
kepstin-work
brianfreud: we just need to allow negative years, right? ;)
2012-05-18 13950, 2012
brianfreud
lol
2012-05-18 13906, 2012
hawke_1
brianfreud: For one thing, with composition you can have multiple ARs with different years to the same work…so that part should be doable without change
2012-05-18 13913, 2012
brianfreud
hehe... Julian calender anyone?
2012-05-18 13927, 2012
reosarevok
kepstin-work: or store the dates in... dunno, what's the Kelvin degree equivalent for dates?
2012-05-18 13936, 2012
kepstin-work
yeah, the fact that there are multiple calendar systems is really annoying.
2012-05-18 13953, 2012
brianfreud
hawke, sure, but I'm talking about imprecise date ranges, not multiple date ranges
2012-05-18 13959, 2012
kepstin-work
e.g. native display of classic chinese or japanese calendars would be cool.
2012-05-18 13904, 2012
reosarevok
There must be some calendar starting at some kind of "absolute zero", even if the absolute zero is "arbitrary, but old enough that nothing happened before it"
2012-05-18 13911, 2012
hawke_1
brianfreud: the “…and completed it” confuses the issue. :-)
2012-05-18 13915, 2012
brianfreud
reosarevok, and don't forget that there is no year zero :D
2012-05-18 13923, 2012
kepstin-work
reosarevok: you could use unix timestamps and allow negative numbers
2012-05-18 13931, 2012
kepstin-work
those give you second accuracy :)
2012-05-18 13932, 2012
hawke_1
64-bit unix timestamps.
2012-05-18 13944, 2012
reosarevok
brianfreud: Of course there is. It's there, on Nine Inch Nail's overview
2012-05-18 13946, 2012
kepstin-work
yep, a 64-bit signed unix timestamp should do it.
2012-05-18 13953, 2012
reosarevok
Well, now it might be hidden by bootlegs, lol
2012-05-18 13943, 2012
kepstin-work
(converting classic calendars to a number of seconds prior to the unix epoch might be a little approximate, though :)
2012-05-18 13911, 2012
brianfreud
kepstin-work, one drawback I've heard about, if you use that for this type of thing, is that it becomes interpreted by javascript in browsers as that time stamp in the current user's time zone, so if you set a unix date 100000, and send it to a user in a sufficiently different time zone, it becomes 100000 for them, but because of the date line, it should be 100001 for them :P
2012-05-18 13941, 2012
kepstin-work
brianfreud: we'd just say that the timestamps are in UTC, and have people convert them if needed
2012-05-18 13911, 2012
reosarevok
kepstin-work: it can't be more approximate than our 0-day bands
hey guys, want to form a band this weekend? we can record one song, publish it, then disband :)
2012-05-18 13910, 2012
kepstin-work
(all in the same day)
2012-05-18 13918, 2012
brianfreud
let me throw just one last crazy possible wierdness that I ran into with Mozart...
2012-05-18 13910, 2012
brianfreud
the start date is approximately known, but inconclusive. There's sufficient evidence to support two different approx dates, but not enough to say which one is correct.
2012-05-18 13928, 2012
reosarevok
Yay!
2012-05-18 13936, 2012
reosarevok
So we need ranges inside ranges inside ranges?
2012-05-18 13942, 2012
brianfreud
So... 'he started composing it either in the fall of 1899, or the spring of 1901. He finished it in late 1905"
2012-05-18 13952, 2012
reosarevok
This is starting to look like my syntax tree exercises
2012-05-18 13957, 2012
brianfreud
it's now not a range, it's two possible dates.
2012-05-18 13914, 2012
reosarevok
Two possible ranges
2012-05-18 13915, 2012
reosarevok
But yes
2012-05-18 13919, 2012
hawke_1
“fall of 1899” is not a date :-p
2012-05-18 13941, 2012
ehrgeiz
uh. for vinyl/casette tracknumbers A1, A2, B3, B4 or A1, A2, B1, B2 ?
2012-05-18 13950, 2012
hawke_1
Anyway, thanks for reminding me why I shouldn’t propose “circa”
2012-05-18 13955, 2012
hawke_1
ehrgeiz: Depends on the release
2012-05-18 13959, 2012
kepstin-work
i'd just make an approximate date fall1899-spring1901 and mention the details in an annotation.
2012-05-18 13927, 2012
hawke_1
ehrgeiz: If not specified at all: A1, A2, B1, B2. (IMO of course, there’s no guideline)
2012-05-18 13941, 2012
hawke_1
kepstin-work: +1
2012-05-18 13943, 2012
reosarevok
kepstin-work: booo, that's the easy way out :p
2012-05-18 13946, 2012
brianfreud
remember, Mozart's catalogue is date-sorted. But a lot of it is relatively dated, not absolutely dated. "B was composed after A, and before C. We know it wasn't composed in July, but it could have been composed in June, or in late fall."
2012-05-18 13919, 2012
reosarevok
brianfreud (and others): speaking of catalogues - how could we store them?
2012-05-18 13937, 2012
kepstin-work loves pulseaudio's support for moving individual programs between audio devices, and having different programs on each audio device.
2012-05-18 13948, 2012
brianfreud
well, if you just want to pick 1 catalog, it's easy - add a cat # field to works.
2012-05-18 13922, 2012
kepstin-work
we could add a catalog entity, with a work-catalog ar that has a cat# field.
2012-05-18 13923, 2012
reosarevok
brianfreud: no, I want a catalogue *page*
2012-05-18 13941, 2012
reosarevok
So I can browse a catalogue, even if it contains works by different artists
2012-05-18 13947, 2012
kepstin-work
or maybe make it an actual relationship and not an ar.
2012-05-18 13949, 2012
brianfreud
But the better way would be to allow for editors to define catalogs for artists, to tie that catalog to the cataloger (via the existing cataloger AR), and allow artists to have multiple catalogs.
2012-05-18 13904, 2012
brianfreud
reosarevok, me too. Mozart's works page is useless, sorted as it is now.
2012-05-18 13918, 2012
brianfreud
But sorted by which catalog, if you support only one?
2012-05-18 13921, 2012
reosarevok
brianfreud: not for artists
2012-05-18 13934, 2012
brianfreud is confused what you mean for "catalog" then
2012-05-18 13942, 2012
reosarevok
Well, most of them *are* for artists
2012-05-18 13955, 2012
reosarevok
But there's no reason we wouldn't want to support ones that have works by different artists
which indicates that a song was not composed by a particular artist.
2012-05-18 13956, 2012
brianfreud
(expand out the annotation
2012-05-18 13925, 2012
reosarevok
heh
2012-05-18 13940, 2012
kepstin-work could add this AR to say, e.g., that Back did not compose "Sexy and I Know It"
2012-05-18 13943, 2012
kepstin-work
Bach*
2012-05-18 13956, 2012
reosarevok bets it was Pachelbel
2012-05-18 13944, 2012
brianfreud
lol, in that list, you can prob find every possible type of funky work credit case. "Never existed, but attr to him", "Attr to him incorrectly", "Attr to him incorrectly and by the wrong title", "Covered, not written, by him"
2012-05-18 13954, 2012
kepstin-work
hmm. rhythmbox lacks a 'repeat one track' option.
2012-05-18 13959, 2012
reosarevok
"Never existed but attributed to him"?
2012-05-18 13902, 2012
kepstin-work makes a note to add that to riker.
2012-05-18 13911, 2012
reosarevok
Like, a legendary "lost" song?
2012-05-18 13915, 2012
brianfreud
famous Rolling Stones spoof issue
2012-05-18 13954, 2012
reosarevok is not too familiar with... well, any "classic" rock / folk
2012-05-18 13903, 2012
reosarevok
(much less with its weird issues :) )
2012-05-18 13928, 2012
reosarevok
"I was at a Neil Young concert one and it was cool" is about the whole of my experience with all of it
2012-05-18 13936, 2012
reosarevok
*once
2012-05-18 13917, 2012
dubwai
reosarevok: look may be instead of artist group just replace in relationships "legal name:" and "performs as:" to something like this "another perform name" Because now we got "legal name" in aliases
2012-05-18 13944, 2012
brianfreud
This is too long for IRC, but take a glance; there's 3 famous "non-existant, but credited to him" lists for Dylan: http://pastebin.com/4B2Lw9HM
2012-05-18 13929, 2012
kepstin-work
dubwai: but then if an artist has multiple performance names, you have to link all of them together separately, which is very annoying to do.
