pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | <stevan> Moose... it's the new Camel ":P | .pmc == PPI source filters!
Set by Alias_ on 16 March 2006.
00:02 KingDiamond joined
TimToady the only think I can see in atoms.t that might have changed recently is qq>hello<; 00:04
I wonder if the parser went all backtracky on me... 00:05
another good reason not to backtrack in a parser...
00:12 KingDiamond joined 00:14 KingDiamond joined 00:15 Daveman joined 00:30 arcady joined 01:07 hexmode joined 01:12 lichtkind joined
lichtkind ?eval 1; 01:12
01:17 f0rth joined 01:24 geoffb joined 01:41 david__ joined 01:55 azuroth joined
azuroth hmm. anyone know of any programs that can read visio files? 02:07
PerlJam azuroth: you're using perl6 to read visio files? 02:08
azuroth no :-)
azuroth runs away
svnbot6 r9930 | lwall++ | atoms.t took a gigabyte or so of memory to parse qq>hello<. 02:21
wolverian heh 02:22
02:22 mjk joined
svnbot6 r9931 | lwall++ | 1. must now be 1.0 02:28
02:29 xern joined 02:30 xern joined
arcady hooray for parseability! 02:32
FurnaceBoy_ cheers 02:35
at MS that would be a "Pass! RTM!" 02:36
FurnaceBoy_ ducks
02:37 justatheory joined, amnesiac joined 02:43 nothingmuch joined 02:44 r0nny joined 02:46 mako132_ joined 03:14 scw joined 03:24 dduncan joined
dduncan well, we're drawing near commit 10,000 ... so any guesses who's going to commit that one? 03:28
03:43 Khisanth joined
svnbot6 r9932 | Darren_Duncan++ | ext/Test/ : restored my name to the authors list of lib/Test.pm 03:55
FurnaceBoy_ shhh! that's audreyt's birthday present 04:03
04:04 tek9 joined 04:08 weinig is now known as weinig|sleep
dduncan whenserbirthday? 04:19
arcady "soon" 04:20
like, a week or so?
look it up in the IRC logs
04:25 coumbes joined 04:32 pdcawley joined 04:39 kakos joined
PerlJam Does audrey celebrate 2 birthdays? 05:03
svnbot6 r9933 | lwall++ | eval {} should be try {} 05:20
05:24 justatheory joined 05:27 justatheory joined
svnbot6 r9934 | lwall++ | bit more eval {...} to try {...} cleanup 05:29
05:30 azuroth joined
gaal morning 05:40
audreyt: if you're around, wanna gobby? 05:41
TimToady haven't seen audrey all day, prolly $jobbing 05:43
svnbot6 r9935 | lwall++ | Added .= as assignop, so it at least parses (then doesn't find infix:<.> yet.)
gaal TimToady: oh. I think I'll start by writing the Q part of the Capture FAQ then. 05:44
heh, that's one advantage of having more than one calendar, you can have more than one birthday. 05:57
TimToady: got gobby? want to take a look at this? 06:01
TimToady what's gobby? 06:05
gaal TimToady: a collaborative editor
TimToady don't know if my brain is ready for that yet...
gaal # darcs.0x539.de/trac/obby/cgi-bin/trac.cgi
I've not used it myself yet, audreyt has with stevan I think, when they were moosing around the MM stuff. 06:07
okay, I'll wait for her to show up (she said she might have some free time, but probably she's useing it to, like, rest)
s/sei/si/ 06:09
TimToady I hope so, she really exhausted herself the last few weeks.
but will gobby let me type with vim instincts? doesn't say... 06:10
so I presume not... 06:11
gaal no, I'm afraid it's a gnomy thing
wolverian uses vim-gnome 06:12
gaal they're doing it right, with the editor just being a layer over a library that provides the collaboration thing, so I suppose sometime this'll get vim integration
TimToady that'd be cool. 06:13
dduncan I was just looking over the #perl6 log for yesterday 06:16
do you prefer the term 'byte' or 'octet' to refer to something that's unambiguously 8 bits? 06:17
TimToady I always use byte.
dduncan I prefer 'octet' myself
TimToady these days it always means 8 bits. 06:18
dduncan bytes can be other lengths, I think
TimToady that's the party line, but it's old news
byte hasn't meant anything but 8 bits for years.
arcady but what about all those 36 bit machines?
TimToady what about 'em?
arcady I'm sure there's still one or two running, somewhere
gaal arcady: I think 32 != 8 also 06:19
TimToady I'm sure there must be. But I'm not going to fight the language, and the language has chosen "byte" to mean 8 bits.
eventually octet will just become a shibboleth for standards writers. 06:20
dduncan more to the present, I think something we may have to deal with is that the term 'tuple' has different meanings depending on context
TimToady Yes, that's more of a problem.
dduncan currently, that is listed as one of Perl's standard types, and its meaning is different than what I think of
arcady what's the perl definition? 06:21
dduncan I think of a tuple as a set of name-type-value triples, where the name part has to be unique for each triple in the set
the uniqueness aspect is like that of a set 06:22
arcady python, of course, disagrees
TimToady haskell too
dduncan my meaning comes from the relational data model, which is based on math or logic
TimToady but dduncan's definition is right out of the D literature.
dduncan although I understand that math also has other meanings 06:23
TimToady and of course Haskell has nothing to do with math. :P
arcady the relational data model's definition seems like the reasonable one
dduncan Haskell is about functional purity, is it not
as is math
TimToady yes, but it knows about positionals, something that is vehemently denied in the 3rd Manifesto stuff I was reading. 06:24
dduncan yes
a tuple in TTM is like a set, but that each element is typed
gaal dduncan: what you described is called a "record" in haskell etc.
dduncan TTM is based on sets
yes, the Tuple of TTM is like a record or struct or class 06:25
gaal TaPL-style tuples have fixed size and anonymous members
but their positions are of course fixed.
arcady those are the kind one sees around programming languages
TimToady in essence, their position in the tuple is their name.
dduncan whereas, TTM opposes things being addressed by order, but rather says things have to be addressed by name only 06:26
TimToady and of course, they can be well typed.
the name of this element is "3rd". :)
dduncan actually, I found out where the name "3rd manifesto came from" ...
gaal dduncan: when you gave your definition, you said a tuple is "a set of ... triples"
a triple is one kind of tuple :)
dduncan essentially, TTM was a response 2 two other documents that people wrote about databases 06:27
so it was noted 3rd
I can pull the names of the first 2 if people want
TimToady gee, it's an iterative process. must be bogus...
arcady of course, they don't even give their manifesto a proper name... only a number 06:28
gaal gee, it's a manifesto, it must be bogus :-P
dduncan that's just a name
probably just a situational name that stuck
TimToady the named-arg part of a Capture is like a D tuple, while the positional part is like a Haskell tuple. 06:29
dduncan what it actually claims to be is an abstract blueprint for the design of a DBMS and the language interface to such a DBMS 06:30
that's what TTM is
a rationale for tuples being wholly named based in a relational model is that it makes doing joins simple 06:31
TimToady Interestingly, P6 has been pushing more towards the named parameters than the positionals.
dduncan you don't have duplicate names after joins, causing ambiguity problems 06:32
and you don't have fields that lack names entirely
TimToady but all the parameters actually have a name--the positionals are just a convenient shorthand.
arcady named parameters are less confusing
dduncan both of which can happen in SQL because SQL lets rows be defined in order
TimToady actually, we can have nameless parameters too, but they're pretty useless... 06:33
but I can't think of anything that's quite like a join offhand. probably just lack of imagination on my part, though. 06:34
what does D do about name collisions on join? didn't see anything about that in the Chapter 3 I read... 06:35
dduncan anyway, I bring up this particular issue because the word 'tuple' is strongly associated with the relational data model and set mathematics before that
TimToady, I'll answer that ...
in D, there are 2 central kinds of joins, which are 'natural' joins and 'cross' joins 06:36
06:36 pmichaud joined
dduncan the latter is a cross-product, where inputs of N and M items produce an output with N*M items 06:37
the former will match up each name in both inputs, lining them up where they are equal, and the output contains that column just once
TimToady also known as "inner" and "outer" joins? 06:38
dduncan so, say you have 2 relations with attributes { A, B } and { B, C } respectively ...
a 'join' between those would produce 1 relation with attributes { A, B, C } 06:39
it is essentially an inner join with same-named columns implicitly matched pu
s/pu/up/
TimToady sure
dduncan so in D, you never specify 'join conditions' like in SQL's inner joins 06:40
they just implicitly join on the same names, sort of like a SQL natural join
partly as a result, it is very terse syntax to join N items
TimToady okay, basically you have some kind of a URI-ish rule that says the name has to mean the same thing universally.
dduncan or rather, the signature of join() can take N relations as input 06:41
no
you can still join on dis-similar names ...
but you first transform one or both of the input relations so that they have same names where they should match up, and different names where they should not 06:42
TimToady why does this remind me of ! addresses on Usenet...
dduncan in D, all relational operators are pure functions, that take one or more relations as input, and produce one relation as output
the join() function connects them 06:43
the rename() function is 1:1, and lets you rename attributes
of course, if you design your database well in the first place, you would just happen to tend towards having same and different/names where the meanings are meant to be same/different 06:44
so you have relatively few renames to do later
TimToady okay, so basically you have an aliasing mechanism for papering over the lack of universal names.
dduncan so you don't need universal names at all, it just helps
yes, rename() is an aliasing mechanism of sorts
SQL's analogue is the 'as' syntax 06:45
TimToady I can see a D Best Practices coming out about naming conventions... :)
dduncan I will note that TTM intentionally focuses on certain core issues related to databases, and largely leaves other issues up to third parties, though they make a point of thinking through that everything they propose is actually implementable 06:46
eg, they don't talk too much about an architecture for users and permissions and such
one reason that I think this stuff or something similar can be built into Perl 6 is that, really, a relation is just a data type, and its operators are just functions 06:47
it could be included the same as you have arrays or objects or junctions
TimToady I can see this falling into the functional programming fallacy of hiding a lot of context in the "current execution state", but what happens when conflicting naming authorities are trying to run their own set of functions over the same data. it gets "webby".
dduncan there is no need for a naming authority, as all the names are locally defined 06:48
TimToady what the heck does "local" mean?
local is the enemy of webby
dduncan perhaps I take that back
what do you mean by "authority"
?
TimToady I mean whatever it is that TTM is sweeping under the carpet. :) 06:49
dduncan TTM is a set of suggestions about a specific problem domain ... what it largely doesn't specify is some tangential things
TimToady If I send out a relation in an RSS feed, what do I *really* call my columns? 06:50
dduncan the 'columns' of relations are 'attributes'
the 'fields' of tuples are also 'attributes'
a relation is a set of distinct tuples where all the tuples have the same set of attributes 06:51
TimToady yeah, yeah, cheap pun on RSS and 5th column and all that newsy stuff.
06:51 Aankhen`` joined
dduncan in loose implementation terms, we could say that a relation is like a class, and a tuple is an object of that class 06:51
but a relation is actually just data too 06:52
TimToady except for minor considerations of tuples being immutable, while the whole *point* of objects is to store mutable state...
dduncan strictly in data terms, a tuple is a set of attributes, and a relation is a set of tuples
yes
I was just using the term class/object here to provide certain analogies to other languages ... perhaps wrongly 06:53
its probably simpler to say that a relation is a set of sets
TimToady so attributes is another term that actually means different things in various places.
dduncan but that every element of the relation is the same size
yes, 'attribute' has a variety of meanings depending on context 06:54
an attribute in TTM speak is a value of any data type
TimToady then why don't they just call them values? 06:55
dduncan we could
I just said attribute there because thats the terminology that TTM used ...
but in a Perl implementation we can just call them values
or rather, an attribute is a slot that represents or stores a value 06:56
...
TTM distinguishes between values and variables ... as does Perl
a value is something that goes in a variable
and an expression has a value / return-value
TimToady I'd like to know more about how TTM treats time.
dduncan a value that isn't in a variable is immutable 06:57
TimToady since time is what makes things mutable, almost by definition...
dduncan when you assign to a variable, you are replacing its content value with a different value
the term 'value' in TTM is something that is immutable 06:58
eg, the number 4 is a value
you can store a different value, such as 3, in a number variable that used to hold 4
TimToady um, I've heard of variables... 06:59
dduncan that changes the variable, but the 4 itself didn't change
of course
what's your question?
TimToady I'd like to know more about how TTM treats time.
dduncan for example? 07:00
TimToady how it, um, emulates the mutability of objects over time.
dduncan if this helps, what TTM describes is meant to be a portion of a larger language, not a complete language to its own 07:01
in TTM speak, a mutable object is a variable
in TTM speak, we have data types ...
every value is of a type, and every variable is of a type
a type definition can be arbitrarily complex 07:02
a number or a character string is a type
a tuple is a type and a relation is a type
conceptually speaking, if a variable is of a complex type ... 07:03
if you were to assign to a member of that type in the variable; eg, if the variable was what we consider an 'object' and we assign to one of its attributes ...
07:04 pmurias joined
dduncan conceptually we are replacing the entire larger value with another one that is the same but for the changed attribute value 07:04
pmurias hi
TimToady right, but I sincerely doubt you make a copy of the entire database every time an attribute is modified.
dduncan not in actuality
conceptually we are 07:05
but the language can still provide syntax for appearing to update just part of it ...
and behind the scenes, the implementation will do likewise
a consequence of this is that if we want to, say, copy a whole database, like to make a backup, ... 07:06
the language syntax is essentially: $db2 = $db1;
you don't have to manually specify all the parts to copy
TimToady or we can just run the whole thing on a quantum computer and really change the whole database in constant time. :)
dduncan yes 07:07
wolverian that should be fine for databases of 6 bits or so ..
TimToady but then I always did believe in the transactional interpretion of QM...
dduncan now, I know it may take time to explain, but I think that any concerns you may not actually have corresponding problems
eg, anything that you think you should be able to do, you probably can and it will be fast 07:08
TimToady no, I'm not worried about that. I'm looking for the things I'm not allowed to think.
for instance, if there's never any ordering anywhere, how do I serialize? 07:09
dduncan serialization is external to the relational algebra
...
you essentially use an operator that changes a relation into something that isn't a relation, like an array 07:10
since relations are inherently orderless
it can be done, no problem
TimToady so I saw the title of some paper that talked about the relationship of time to all this. My question is whether that is "outside the relational algebra".
dduncan also keep in mind that some tasks people think they need to have serialized actually don't 07:11
think like your junctions concept for example
when transforming relations, it doesn't matter what order the tuples are processed, so they can be done in any order or parallelized; it will come together in the end
serialization only really is useful when moving data between the relational and non-relational realms, such as for display to a user 07:12
TimToady I don't mind that. Again, I'm just trying to see what the relational algebra *doesn't* express.
07:12 XaXXon joined
dduncan the scope is probably analagous to what kinds of algebra you can have for a 'set' 07:13
XaXXon ?eval 1+1
hrm
TimToady In any given linguistic system, there are the things you have to say vs the things that are optional.
07:13 XaXXon left
dduncan eg, with a set, you can add or remove elements, do union, intersection, difference ... 07:13
of multiple sets
TimToady In a programming context, there are things you have to promise to the computer, and the things that you can promise optionally.
dduncan stuff like that
TimToady junctions are an optional promise of non-interference between the alternatives. 07:14
dduncan since relations are sort-of 2-dimensional sets, you get the adding/removing unioning, differencing, intersecting in multiple dimensions
TimToady one could write the same thing without making the promise.
07:15 XaXXon joined
dduncan every element in a relation is, by definition, unique, and addressable by name rather than position 07:15
like with a set
TimToady well, I need to go to bed, and I think we're just talking vaguely in each other's direction at this point. 07:16
wolverian it's like two conversations..
good night :)
dduncan of course, if I actually made that implementation, it should be easier to discuss :)
TimToady nite
dduncan good night
also, for now I'll just name my impression of a tuple as "Relation::Tuple" 07:17
no ambiguity
ultimately, the difficulty in using a relation shouldn't be much more than in using an array or a hash or a set 07:18
or a simple object
and that's it for me tonight too 07:19
12:19am and all
pmurias good night 07:22
XaXXon configure: error: GHC is required unless bootstrapping from .hc files. Haskell support anyone? :) 07:26
the "building guide" on the web page is a 404
rgs install ghc ?
XaXXon oh I thought that was what I was doing 07:27
I downloaded ghc-6.4.2.xxxxxxxx, untarred it and read the README and it said to run ./configure
and it gave that error.. 07:28
I see.. I'm supposed to start from a binary release..
rgs yes
XaXXon of course there's no universal os x build
I guess I rosetta bootstrap 07:29
07:30 azuroth joined
XaXXon sigh.. too hard :( 07:34
rgs I don't use OS X :/ 07:36
pmurias XaXXon: maybe try compiling from .hc's 07:38
07:38 drrho joined
XaXXon no, actually I just grabbed the wrong package.. 07:38
I grabbed the "pretty os x package" instead of the "unixy os x package"
so there was no "ghc", there was "GHC Compiler Kit.app" 07:39
doing the stuff from .hc files is waaay more than I'm prepared for.. I read a tad about that..
pmurias i use gentoo, so i never had any problems, just a quick emerge ghc-bin 07:40
XaXXon yeah.. darwin is a pita sometimes
pmurias getting the graphic card working is the biggest pita on gentoo, everything else just works (except true type fonts in xterms) 07:43
XaXXon yeah, I've run gentoo
pmurias why did you switch?
ayrnieu everybody was doing it. 07:44
the computers were so pretty.
XaXXon switch? oh I haven't
ayrnieu Mac people sneered at me for twisting to adapt to inhumane interfaces.
XaXXon I have a windows box, a linux box, and a mac (actually 2 counting my powerbook) in my office
the new mac minis are so sweet, I had to grab one 07:45
ayrnieu the initial state, where you're just learning and willing to adapt because you want to get a water-runs-downhill view of life in OSX, is very very pleasant, because Apple loves those who do what Apple says.
XaXXon yeah, a lot of it pisses me off, I'll go with that 07:46
it's not for power users.. but if you're sitting in an app doing non-coding type work, it's nice
I use iMovie and iDVD a lot
ayrnieu life past that probably depends on how 'twisted' you are.
XaXXon but mostly I use it as a tiny, quiet dual-core unix-y system 07:47
if it didn't have so many damn wires coming out of it, I'd probably lose it
pmurias i just use xterm's and firefox and the gimp if i do masochistic script-fu scripting
XaXXon GHC was thinking about building from source.. but oh well.. 07:48
I should be able to install pugs now?
or not 07:49
pmurias is there a terminal emulator which can display images(no as backgrounds) but as part of the text flow?
XaXXon would it wrap vertically? 07:50
I mean.. err.. huh?
ayrnieu pmurias - cleverness along w3m's lines?
someone wanted such a thing a while ago, and I thought them a bit crazy, but that suddenly makes sense for some new-type IF games. 07:51
pmurias it makes sense for doing a lot of stuff 07:52
ayrnieu I don't see those yet, though.
pmurias you could do: icat image.png | filter1 | filter2 and the image would display filtered twice 07:53
ayrnieu that has nothing to do with your terminal emulator.
pmurias i spoils the fun if it displays in a new window 07:54
07:55 Santana joined
ayrnieu the view-jpg at the end of that stream is only part that matters. 07:55
it could talk to X or it could talk to your terminal; your terminal can't possibly care about the filtering that happens up to then. 07:56
pmurias visiting a friend& 07:58
07:59 Santana left
integral has done pipelines like that for video... (ffplay/ffmpeg work fine on pipes) 08:01
08:15 KingDillyDilly joined
svnbot6 r9936 | Darren_Duncan++ | trunk/Rosetta/ : a few minor updates; this now corresponds to the Perl 5 version's 0.724.0 CPAN release 08:23
dduncan er, NOW I'm really going to be gone 08:24
Larry, if you backlog and see this, and you're looking at the TTM stuff online, make sure to see the Chapter 4 of the 3rd edition, and not the Chapter 3 of the 2nd edition, which may be what you had seen online before a few days ago ... significant improvements 08:25
or it may not change your impression at all, who knows
regardless, what I propose for Perl 6 is only a small subset of what is mentioned there, and so the Relation for Perl 6 is essentially just an object, and is as mutable as a Perl 'Array' or 'Set' or 'Hash' etc 08:26
I'll defer further discussions until I have an implementation ... any week now 08:27
so to avoid us accidentally speaking past each other
er, make that, TimToady, if you see this ... in case your client highlights your nick 08:28
good night, anyway 08:29
wolverian good night :)
08:33 dduncan left
KingDillyDilly The future of Perl depends on TimToady reading the above message? :-/ 08:33
Someone should invent email or something. 08:35
XaXXon or usenet 08:36
now I know why perl6 is taking so long
buu Pssh. Email sucks.
08:39 integral is now known as ntgrl 08:43 KingDiamond joined
KingDillyDilly Diamonds can't cut Dillies. 08:44
08:44 integral joined 08:58 chris2 joined 09:22 nothingmuch joined
nothingmuch pmurias: ping 09:24
09:25 Khisanth joined 09:54 KingDillyDilly joined 10:07 david__ joined 10:09 _bernhard joined 10:11 iblechbot joined 10:12 PolettiX joined
pmurias nothingmuch: pong 10:15
i'll be back in 1h
b, raining 10:28
10:30 KingDiamond joined
ingy hola 10:33
10:36 elmex joined
ingy elk 11:01
pmurias ? 11:02
_bernhard pmurias: Just cleanup up some Parrot_init() calls in Parrot 11:10
nothingmuch hola ingy 11:19
pmurias _berngard: which one should i clean up? 11:23
_bernhard Sorry, I meant that I just cleaned up Parrot_init() calls. Parrot SVN r12217 11:30
nothingmuch pmurias: got /msg? 11:35
pmurias no 11:42
__bernhard: i'm recompiling parrot to see if it still works 11:46
nothingmuch pmurias: mail me an ssh key 11:50
pmurias: or maybe /msg on irc.perl.org? 11:51
pmurias i'll msg on irc.perl.org
11:52 pmurias joined
nothingmuch pmurias: urf, got nothing from yo on perl.org 11:57
11:57 nnunley joined
pmurias pmurias: sent now 11:58
nothingmuch pmurias: are you messaging in the correct network? 12:01
i got nothing except the "here i am"
freenode--
12:03 KingDillyDilly left 12:13 turrepurre joined 12:20 kolibrie joined 12:27 Limbic_Region joined
pmurias nothingmuch: msg the key to you 12:34
s/msg/msg'ed/
is it correct? 12:35
nothingmuch looks OK
12:38 nnunley_ joined 12:55 KingDiamond joined 12:58 cognominal joined 13:00 dakkar joined
gaal interesting lead: "packrat parsing" 13:09
# pdos.csail.mit.edu/~baford/packrat/
audreyt gaal: how's your time look like in the next... er, 96 hours?
13:09 SamB joined
gaal audreyt: have to leave in fifteen minutes till... next morning, basically, IDT 13:10
audreyt idt?
gaal GMT+3
uh, so next 19 hours or so are out, unfortunately 13:11
audreyt that's fine... I'll work rebooting the journal instead then 13:12
fascinating news that pmichaud's tree starts to blossom
gaal ok let's try gobby just for tech test, please connect to sike.forum2.org 13:13
audreyt woot 13:16
I mean, moose
(it works)
gaal whee, astounding, it works
you must acquit
audreyt :D
I must blog, is more like it
PerlJam good morning #perl6 people
audreyt several friends and acquaintances threatened dire actions if I continue to skip journaling 13:17
gaal
.oO( audreyt: I must ACCKKEKKEEK!!! QUIT!!! )
continuation style skip journalling
audreyt ok, the subject of this entry should be "reset from the shift" 13:20
gaal++ # providing obscure CPS reference
theorbtwo You know, it took me a moment to realize that the Airport Express had nothing to do with 802.11 13:21
audreyt which airport was that? :) 13:22
audreyt had that moment in Narita @ tokyo
gaal audreyt: don't you mean obscure capture?
i'll miss references. they are the longest word I know with only one vowel.
audreyt no, I mean CPS as in Capture / Pair / Signature 13:23
the fundamental confusion-generating traid over perl6 calling convention
triad, even
gaal see, it's so obscure even I missed it
good thing we have you around :)
theorbtwo Gaal: Mississippi is tied, I think.
audreyt that's okay, I missed it the first time too
backronym is fair play 13:24
theorbtwo Longer, even.
gaal "what's got four eyes and can't see?"
azuroth hmm. c# uses {} in some strings, but there's no way to escape them? lame..
gaal (joke from Mississippi Burning)
okay, must moose. elk y'all later! & 13:25
audreyt g'alces, gaal 13:26
rafl Juerd: That's fine with me. 13:33
Juerd: How about the stuff of mine you still have? :-)
Juerd rafl: Yea, I should probably send that some time :) 13:43
rafl: You were still in the process of making up your mind about ultrabay and such 13:44
rafl Juerd: No money for that, currently. Sorry.
Juerd: What things do you still have?
Juerd rafl: Eh, I forgot. At least PBP. 13:45
I have it in a bag in a corner of my room
rafl Juerd: That's also the only think I'm aware of. 13:46
Juerd: And the thing I'd really like to have :-)
Juerd I can imagine
Limbic_Region Juerd - I had a dream I was teaching someone to dypte dvorak last night 13:50
PerlJam I once had a dream that we'd have a parrot-based perl6 compiler and lo one exists today! ;) 13:51
Limbic_Region PerlJam - should I backlog for an announcment or something? 13:52
and did we skip over pugs 6.28.0 ?
rafl Juerd: You forgot your s/.*/perl/; shirt. 13:53
PerlJam Limbic_Region: no, I'm referring to parrot/languages/perl6
Limbic_Region oh - that's been there for a looooong time PerlJam (albeit not in its current form)
PerlJam Limbic_Region: but now it works! 13:54
PGE+TGE+pmichaud == perl6 compiler based on parrot
Juerd rafl: Oh... 13:55
Limbic_Region: I actually did it for real two nights ago.
Limbic_Region PerlJam - and it does what as an object model?
PerlJam nopaste.snit.ch:8001/6938
Juerd Limbic_Region: In a dvorak workshop. There were 10 attendees.
PerlJam Limbic_Region: it's not *quite* that far along in working :)
Limbic_Region cool Juerd - I really need to find more time to practice. There are just so many interesting distractions for me 13:56
Juerd Maybe I should publish my slides for that
Limbic_Region Juerd - please do
Juerd The slides aren't finished yet; I forgot some important stuff, I discovered during the workshop
rafl forgot the gcc flag which skips linking.. could someone please remind me?
Juerd (Prepared them in one hour)
Limbic_Region: juerd.nl/files/slides/dvorak/dvorak.html 13:59
Limbic_Region: juerd.nl/files/slides/dvorak/blue_h...y_bold.ttf ;) 14:00
integral rafl: -c to compile a single .c to a .o 14:01
rafl integral: Yes! Thank you. 14:02
Limbic_Region thanks Juerd 14:03
dakkar Juerd: dvorak.nl is the same server as juerd.nl ? 14:04
Juerd dakkar: Yes, why?
dakkar it keeps sending me just the first few lines of every image ;-)
dakkar uses a dvorak keyboard 14:05
Juerd Lines of images?
And do you have this problem on both juerd.nl and dvorak.nl?
dakkar yes, like the pictures in the slides, or the photos on the "buy keyboard" pages
Juerd Weird...
dakkar i get only part of each image...
Juerd What's your IP? 14:06
dakkar now I'll try to see if it's something on my side
14:06 prefiks joined
dakkar Juerd: might be 85.45.142.2 (I'm behind a couple of NATs) 14:06
Juerd dakkar: tnx.nl/ip
dakkar I confirm my ip 14:07
Juerd ok
dakkar: Huge latency on your line... 14:08
between hops 8 and 9: 14:09
8 r-rm199-vl3.opb.interbusiness.it (151.99.29.152) 72 ms 90 ms 81 ms
9 r-pi31-rm199.opb.interbusiness.it (151.99.101.22) [MPLS: Label 481 Exp 0] 1410 ms 1266 ms 1213 ms
dakkar quite probable... I'm at work
Juerd My server has very strict timeouts
Limbic_Region Juerd - it might be helpful to show how to change keyboard layout on a couple of the most common OSes
dakkar it's the backbone trunk between Rome and Pisa...
nice to know
Juerd Limbic_Region: dvzine.org
Limbic_Region: I want something that can create Flash movies out of screenshots 14:10
dakkar mencoder!
Juerd dakkar: Under Windows
And *including* the cursor
dakkar mencoder+cygwin ;-)
ouch
Juerd As in: demo the entire thing
People used to GUIs get lost easily if you show them only intermediate screenshots. 14:11
I'm off now; going to reinstall my laptop
stevan audreyt: ping 14:16
audreyt stevan: pong 14:18
stevan audreyt: I was wondering about the YAPC::NA hackathon
the wiki says it is post-conference
any details on that?
audreyt stevan: June 29 to July 2 (plus or minus pi days) 14:19
stevan cool
accomidations? or are we on our own?
audreyt ingy: accomodations? or are they on their own?
nothingmuch yay! birthday 14:20
stevan wonders if ingy has a cabin somewhere in the woods
audreyt (it's possible that we'll come to ingy's parents' place in Antioch)
stevan ingy has parents?
he wasnts just spawned from the internet?
audreyt surprising discovery, no?
stevan re-evaluates his entire perception of reality
audreyt to quote from ingy's blog: 14:21
Like my Dad always told me, "Son, you can't always get your namesake domain, but you can always make your domain your namesake".
see? he has a Dad.
stevan :D
class Ingy { has @parents; }
class Ingy::d?t::Net { has @parents handles <hackathon_accomidations>; } 14:24
audreyt it's scary that when I wget from ..../moose
wget saves it as moose.5
evidently I've been brainwashed 14:25
theorbtwo You already have moose and moose.[1234]?
audreyt yeah, those are my favourite temp file names on other machines
(and then I wget them back)
stevan thinks we should talk to o-reilly about changing the animal for Perl 6
14:25 kane_ joined
rgs $ meta --ws antlers 5 14:26
hirvi velvet_antler caribou deer moose
favorite metasyntactic variable names :p
kolibrie Juerd: www.macromedia.com/software/captivate/ 14:30
14:36 ghenry joined 14:41 q[uri] joined
Juerd kolibrie: Thanks. I'll have a look at it soon. 14:45
14:45 kane__ joined
Juerd Hoi kane__ 14:45
14:52 KingDiamond joined
spinclad [aa|a|b] could compile into [a:[a:|]|b:] 1<2 > 15:05
15:13 xinming1983 joined, vel joined
audreyt 1<2> ? 15:13
audreyt decides to sleep instead of expanding more from the freshly-posted journal entry 15:14
Juerd Sleep well, audreyt 15:16
15:17 avarab joined
Juerd Wow, long post. Thanks :) 15:17
svnbot6 r9937 | pmurias++ | works now for simple things 15:21
r9937 | pmurias++ | expressions are yet to be done
r9936 | Darren_Duncan++ | trunk/Rosetta/ : a few minor updates; this now corresponds to the Perl 5 version's 0.724.0 CPAN release
nothingmuch i need a volunteer to finish off the smoke server
i can'ty
i just don't have design-fu
Juerd What kind of design? 15:23
pmurias nothingmuch: you are not the only one who dosn't have much 15:24
design-fu is IMHO the part of programming that requires most experience 15:26
15:26 FurnaceBoy joined
pmurias mine suck completly, 15:27
nothingmuch oh 15:28
that kind of design fu i think i have =)
i meant html-design-fu
pmurias: wrt the mail address: the only thing you should know is that occasionally (~once per month) i have cable-carrier downtime, that is the connection drops for ~10-20 mins 15:30
and in addition my ISP is filled with assholes
but soon i will probably have a decent ISP in the states
15:32 justatheory joined
kolibrie nothingmuch: how long are you going to be in the states? long-term? 15:32
pmurias nothingmuch: we can make a deal, you help me design the emitter i'll do the html design for you :)
nothingmuch okay =) 15:33
kolibrie: 6 months, it seems
kolibrie nothingmuch: cool
pmurias i learned html/css the reverse-ajax way so don't expect art-design-fu 15:35
nothingmuch pmurias: this is simple layout stuff
i'll have a brain farting session with you later tonight, kay? 15:36
pmurias ok
where can i find the stuff you want layout? 15:39
ingy hi nothingmuch 15:40
xinming1983 Round-half-even, what does round half even mean? 15:41
bbl
15:41 xinming joined
xinming back 15:42
what does round-half-even mean? hmm, It's from www.pldesignline.com/howto/showArti...=175801189
Limbic_Region xinming - perhaps to the next even number? 15:43
xinming Limbic_Region: hmm, I don't know. :-/
theorbtwo xinming: round half to the nearest even number.
xinming so, 4.2 will be rounded to 5? 15:44
Juerd Wha, fuck, my backup ends at half of my music archive :(
xinming and 5.2 will be rounded to 5.0?
Juerd xinming: 5 is not an even number.
Limbic_Region it is defined further on xinming
dakkar xinming: no, it only kicks in if you round at helf
xinming oops...
Juerd xinming: 2, 4, 6, 8, ... are even, 1, 3, 5, 7, ... are odd
theorbtwo It only applies if you're at the halfway point. 4.2 rounds to 4.
Limbic_Region search for Round-half-even: If half-way values are always
xinming yea, I just confused these 2 words. :-)
dakkar 4.5 rounds at 4, 3.5 rounds at 4
Juerd KingDiamond: So 4.5 is rounded to 4, 5.5 is rounded to 6
theorbtwo 4.5 rounds to 4, because it's round and 5 isn't.
Er, s/round and/even and/ 15:45
Juerd even
xinming how about 4.7?
5.0?
dakkar yes
nearest
Juerd 4.7 is always rounded to 5, because it's not .5
.5 is special because it's *exactly* halfway.
If it's .499999999999999999 or .50000000000000000001, there's no longer any question.
xinming so, no matter what the number after the point... It will always reach the even number? 15:46
Juerd No...
Only if the to-be-truncated part is exactly half (that is, 5), it rounds toward the nearest even number
xinming 4.2 --(round-half-even)--> 5.0?
oops 15:47
Juerd No, 0.2 is less than 0.5, so it's always rounded down.
Limbic_Region no xinming - did you read the explanation yet
Juerd 0.6 is more than 0.5, so it's always rounded up.
0.5 is exactly 0.5, so it rounds to the nearest even.
xinming Limbic_Region: yes, but I don't understand as there isn't many example out there...
Limbic_Region basically it is a way to remove the bias from a system that always rounds .5 up
imagine you have hundreds of floating point numbers that you wanted to do some statistics on 15:48
first you decide to round the numbers to make things more digestible
by rounding all the numbers at exactly .5 up - you could possibly bias the resulting statistics
if instead - you rounded numbers that were exactly #.5 to the nearest even number 15:49
you have a much better chance of averaging things out
Juerd perl -e'for ($_ = 0; $_ <= 10; $_ += 0.1) { printf "%.1f --> %.0f", $_, $_; print " <--" if $_ =~ /\.5$/; print "\n"; }'
Limbic_Region of course for datasets produced by a function that leads to more odd numbers than even - this is flawed thinking 15:50
make more sense xinming
Juerd If it's at .5 without any <--, then see perlfaq for the explanation :)
xinming Juerd: perl-fect one-liner. :-P
PerlJam Limbic_Region: you should write that up in a little more expanded form and put it on perlmonks as a tutorial or something (Assuming it isn't already there somewhere)
Limbic_Region PerlJam - first I had encountered it. I am adding it to my list of things to contemplate 15:51
xinming Round-half-up
Juerd: from your example, it seems it is the same as round-half-up
15:52 sahadev joined
Juerd xinming: No, look at 2.5, 4.5, 6.5 15:52
Limbic_Region that's actually a really good article 15:55
Limbic_Region bookmarks
xinming `head TASKS` :-) 15:56
dakkar Juerd: what perl are you running? here it rounds up... 15:57
Juerd Oh, it rounds 2.5 up, 4.5 down, 6.5 down
Heh, add %.20f 15:58
And another , $_
Then you see why :)
dakkar I hate limited-precision math... 15:59
me ne vado a casa 16:02
(wrong window) 16:03
Limbic_Region was going to ask dakkar what print 4 * atan2 1, 1; should print on his limitless-precision math computer 16:06
16:24 hexmode joined
obra any of the javascript gang around? 16:28
miyagawa audrey++ # pugs.blogs.com/audrey/2006/04/post.html 16:32
otori-tan is very lovely
16:34 Daveman joined
theorbtwo Is 'the Japanese romanization' really correct there? 16:35
Juerd rafl: I lost all the geektour photos. Sorry.
rafl: My backup is broken - the tar.bz2 is truncated at 4.0 GB :( 16:36
miyagawa theorbtwo: yeah, it's correct. To be more accurate, the only difference is T vs. D in the middle of o and ri, but other than that, it's very naturally same
16:37 frederico joined
rafl Juerd: Ouch! 16:37
Juerd rafl: Yes.
I lost my music archive and photo archive.
rafl Juerd: Oh well.. I only took quite bad photos. They are at gallery.perldition.org.
Juerd: *OUCH*~ 16:38
Juerd So... people with good music, donate some tgz's please :)
rafl only has CC-licensed music on his hard disk, currently.
rgs that "otori-tan" naming story is awesome 16:39
theorbtwo Oh! 16:40
You take "Audrey Tang", convert it into hirigana using the Chinese rules, then into Latin with the Japanese rules, and try to pronunce that. 16:41
xinming I worked with Larry to remove the concept of references from Perl 6, along with confusing concepts such as auto-enreferncing and auto-dereferencing. Now a Scalar holding an Array object is... well, just that.
16:41 nnunley_ is now known as nnunley
xinming hmm, what will the "reference" be called in the future? :-) 16:41
miyagawa Well, when we try to "parse" Audrey Tang as Japanese, it sounds like Audrey + Tan. And "-tan" means a postfix for lovely girls. Audrey can be kanjized using "otori" which magically is the same kanji with her Chinese name for Phoenix 16:48
Juerd Complex. 16:50
I'm glad my name is ASCII :)
xinming is happy with several names. :-P 16:53
obra miyagawa: are your plane reservations set? 16:55
miyagawa obra: Not yet. Need to sort out my work schedule ...
I'll fix it on Monday 16:56
obra ok 16:57
miyagawa worst case is I stay in .jp for this golden week
obra heh. *nod* 16:58
miyagawa but anyhow I'm going to US end of May, and I'll fly to East Coast during the stay until August
obra nice 16:59
pmurias Juerd: what do you call good music? 17:08
17:19 zgh joined
Juerd pmurias: Almost anything that has a normal rhythm :) 17:21
And doesn't sound like people whining
I just discovered that I still have my Metallica, Weird Al, and SoaD 17:22
FurnaceBoy so, no Alanis? 17:23
pmurias we have a *lot* cds arranged mostly in piles in the house but it's mostly jazz,classical and King Crimson 17:24
Juerd Classical's great. 17:25
cognominal juerd : "I'm glad my name is ASCII :)" and you prononce it "weird"? 17:34
Juerd cognominal: ASCII is not my name, it's a property of it :)
And no, Juerd is not pronounced "weird" 17:35
kolibrie Juerd: is there a diphthong in Juerd? 17:38
Juerd kolibrie: If you want 17:39
ue can be either a diphthong or uë 17:40
In case of ue, it's pronounced as the dutch uu
In case of üe, as uu-e
eh
I meant uë
Though if you are/know German, feel free to imagine an umlaut on the u. 17:41
pmurias Juerd: i could post you a list of mp3 on my fathers hd
Juerd pmurias: And then? :)
kolibrie Juerd: I probably say it mostly like dutch uë
Juerd kolibrie: When have you mentioned my name irl then? :)
kolibrie Juerd: perhaps never 17:42
Juerd Hmm :)
Confusing
kolibrie Juerd: I just pronounce things in my mind while I read 17:43
Limbic_Region pronounces Juerd as Gerard in his head for some reason
Juerd kolibrie: Hm, everything?
Limbic_Region: HUH?
pmurias Juerd: downloading via http would be ok, but i don't think you'll like any of that (unless you like Dylan (not the language))
ingy a really drunk guy at a bar asked me "Are you a rock star?" 17:44
Juerd pmurias: Doesn't ring any bell. I wasn't very serious, by the way, I don't think it's a good idea to ask other people to do illegal things for me :)
ingy I said, "Um, yes actually"
"What's your band?"
Limbic_Region Juerd - gerard pronounced jer rod
Juerd ingy: Why else the ö
ingy "Perl"
"What instrument do you play?" 17:45
FurnaceBoy "keyboard"
ingy "hmmm, um, keyboard"
yeah
FurnaceBoy :)
miyagawa ingy++
kolibrie is still puzzling over whether he pronounces everything in his mind while he reads
ingy then he said he was Butch Vig's cousin
Juerd kolibrie: ...
kolibrie Juerd: I give up
Juerd kolibrie: ...
(?)
17:47 jserv-- joined
kolibrie Juerd: I think I can read faster than I speak, so I must not say everything in my mind, but when I slow down, I think I mentally pronounce what I'm reading 17:47
Juerd I have to copy things to force myself to slow down. 17:49
kolibrie finds he does most things very slowly 17:51
18:12 frederico joined 18:15 Quell joined 18:31 dduncan joined 18:34 KingDiamond joined
xinming gaal: could you please tell me what do you use to make this slide? perlcabal.org/~gaal/peek/slide7a.html 18:52
obra that looks like spork 18:57
xinming obra: home page please... googled for it... but lost in so many entries 19:01
obra um. it's something of ingy's. try cpan
Juerd xinming: <meta name="generator" content="Spork v0.20" /> 19:02
xinming :-) sorry, I missed that... 19:03
spinclad re round-half-even: for exact arithmetic choose an exact binary fraction: perl -e'for ($_ = 0; $_ <= 10; $_ += 0.125) { print $_; printf " --> %.0f", $_; print " <--" if $_ =~ /\.5$/; print "\n"; }' 19:13
19:19 rindolf joined 19:20 rindolf left
Juerd spinclad: Clever! 19:25
19:29 xinming1983 joined 19:31 xinming joined 19:38 ahdahm joined 19:44 weinig|sleep is now known as weinig 19:47 avarab joined 20:02 FurnaceBoy_ joined 20:28 FurnaceBoy_ is now known as FurnaceBoy
gaal xinming: spork, yes, and a plugin for the syntax coloring 20:34
Kwiki::VimMode
cognominal scrapbook++ 20:44
firefox++
Juerd What's scrapbook? 20:52
svnbot6 r9938 | Darren_Duncan++ | added beginnings of ext/Relation/ - a logical Relation data type for Perl 6 (a superset of the Set data type)
cognominal oops, wrong window. Anyway that is amb.vis.ne.jp/mozilla/scrapbook/ a nice tool to archive pages 20:55
Juerd Thanks
cognominal recursive save is too agressive so.
Juerd Looks useful
cognominal you can save page weeded from a lot of junk 20:56
while we are at it, I recommand aluminum.sourmilk.net/reveal/ too, it allow to search material in all the tabs of a firefox wiindow 20:59
Juerd Neat 21:00
I don't use a lot of extensions 21:01
noscript / adblock / adblock filterset.g updater / downthemall / web developer / fasterfox is my standard set nowadays 21:02
I'll see if reveal and scrapbook will be added. Probably :)
cognominal I need an all encompassing extension for dealing with bookmarks, there are many incompatible half baked :( 21:03
Juerd Holy cow. reveal is like expose for the only thing on earth I've ever used expose for! :)) 21:04
Juerd hugs cognominal
You're my hero for today.
s/expose/exposé/ # I keep forgetting I have utf8 working now, and wanted to pretend I'm a unicode fanboy 21:05
cognominal I am using my kbd as qwerty so I am at a loss for diacritics outside emacs.
Juerd I use standard X11 Multi_key 21:06
Which lets me compose things.
cognominal downthemall sounds nice
Juerd for é I hit multi,',e 21:07
On this box, my right ctrl key is mapped to multi
Do you run X?
cognominal yes, I am on linux.
Juerd xmodmap -e 'keycode 117 = Mode_switch' 21:08
ehh
xmodmap -e 'keycode 117 = Multi_key'
cognominal I have been to lazy to search but because you are feeding me with a spoon...
Juerd That binds your windows menu key to multi
gnome and kde have ways of setting this up gui-ly - if you use anything else, just script it so that xmodmap is started at startup 21:09
cognominal I must say I sincerely hate X
Juerd Everyone does
But it's the best thing we have
cognominal long ago, I worked peripherally for NeWS. that was awesome but never done properly 21:10
conceptually X is crap compared to NeWS
gaal Otori-tan: ping 21:15
21:42 Kattana joined 22:01 lichtkind joined
lichtkind ?eval (2&8) == (2|3|5|7); 22:04
?eval 1; 22:05
can someone please start evalbit? 22:06
bot
Juerd lichtkind: Do you not have any pugs installation to test these things with?
lichtkind: Even if evalbot would be around, this stuff would not be the intended purpose. 22:07
lichtkind: evalbot is for showing people something, not for testing.
lichtkind: You can get access to feather if you don't have a machine that can compile pugs
(And even if you have such a machine, you can get access :))
lichtkind your right 22:12
but im still writing this thing an wantet test only 2 3 cases
Juerd You can run pugs -e for that
pugs -e'say (2&8) == ...' 22:13
Just like perl -e
TimToady someone remind me why pugs always recompiles Syck? it does it when you do a make install as root, which then breaks your build as non-root.
22:17 hlen_ joined
TimToady audreyt: see above 22:18
22:19 hlen_ is now known as hlen 22:20 nnunley_ joined
lichtkind * % / ** impose num context too? 22:25
22:27 xX[ReP]Xx left
TimToady Let's say the standard definitions are biased towards num interpretation. You could define any * % / ** MMD method you like and give it your own semantics, but stock Perl will treat Str as Num there. 22:32
whether that's by coercion or by just defining Str MMDs with num semantics, I don't know.l
coercion is just a way of autogenerating missing methods, is one way to look at it. 22:33
LeTo why coercion, when MMD exists? 22:36
dduncan TimToady, regarding that relation stuff, I will try to be brief and concise today ... 22:37
first of all, I draw a clear separation between the data type I propose for perl 6 and anything to do with databases
what I propose for perl 6 is just a set-like class, which is mutable like a set or array or hash data type 22:38
and a few operators for it
that's all
lichtkind Thanks Tim (still enjoying your last talk)
dduncan so TTM et al can basically be ignored
I have also started to put an implementation in ext/Relation/ today 22:39
Juerd thinks it's funny that lichtkind calls TimToady Tim.
svnbot6 r9939 | lwall++ | Added * to %ENV and @ARGS that seem to be missing it (and aren't historical).
r9939 | lwall++ | Also added tests for importation of %ENV from GLOBAL.
dduncan which should be done over the next few days
Juerd LeTo: Because otherwise you have to define a method for each existing type? 22:40
dduncan emphasizing that clear separation, I see that this Relation proposal is completely separate from my Rosetta project, so none of the Rosetta/TTM stuff needs to be read into it
Juerd Nevermind. There is of course a single catchall type.
LeTo MMD on 'Any' can still DTRT
dduncan just instead think of Relation as ext/Set plus more, or something
22:49 larsen joined
dduncan question ... is anyone else finding that the make-test for Pugs modules is reaaaaalllllllyyyyyy sllllooooowwww? 22:53
eg, running 'make test' for just ext/Test/ ...
takes several minutes to run
can this be affected by heap size?
or does a larger heap just prevent crashing? 22:54
the figures I see today specifically are: Files=3, Tests=83, 219 wallclock secs (210.59 cusr + 4.88 csys = 215.47 CPU)
whareas, back around when 6.2.11 was new, that completed in just a few seconds 22:55
(you can test individual modules after 'make' completes on the main pugs by first cd-ing to the ext module dir, like into ext/Test/ and running 'make test') 22:56
note that launching Pugs directly only takes about 3 seconds to load the prelude, and normal statements seem to run right away 22:57
so the prelude is pre-compiled
this is all standard Haskell runcore, what is used by default 22:59
23:02 fglock joined
dduncan note that the speed problem here has been occurring for weeks, and I suspect it isn't happening to others as then it would have been fixed 23:02
will try another clean build ...
23:05 Limbic_Region joined
dduncan maybe explicitly specify a larger heap too ... 23:08
svnbot6 r9940 | fglock++ | PCR - small fixes 23:12
r9941 | fglock++ | PG-P6 - 03-statement-control.t passes all tests
23:14 nothingmuch joined 23:16 stclare joined
svnbot6 r9945 | fglock++ | PG-P6 - removed List.pm 23:33
23:36 kolibrie joined
fglock pmurias++ - emitter for Pugs-Grammar 23:39
audreyt: Pugs::Grammar main engine looks ready - I'll start adding stuff to try to parse Pugs tests 23:41
later & 23:42
23:42 fglock left

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!