Smerdyakov has quit [Read error: 110 (Connection timed out)]
Smerdyakov has joined #ocaml
bk_ has quit ["Leaving IRC - dircproxy 1.1.0"]
fab__ has joined #ocaml
_fab has quit [Read error: 110 (Connection timed out)]
monochrom has joined #ocaml
cjohnson has quit ["Leaving"]
mlh has joined #ocaml
vezenchio has quit ["Hydrocon technology produces two waste products: steam and green goo. We dispose of the steam through various vents, strategi]
m3ga has joined #ocaml
m3ga has quit ["Leaving"]
pattern has joined #ocaml
grirgz has quit [Read error: 110 (Connection timed out)]
mlh has quit ["who are you people and what are you doing in my computer!?"]
Walnutssh has quit ["Client exiting"]
withersoever has quit ["Leaving"]
gl has quit ["My damn controlling terminal disappeared!"]
gl has joined #ocaml
grirgz has joined #ocaml
withersoever has joined #ocaml
kuz has quit [Read error: 110 (Connection timed out)]
gzl has quit [Remote closed the connection]
gzl has joined #ocaml
durnew has joined #ocaml
cjohnson has joined #ocaml
kuz has joined #ocaml
m3ga has joined #ocaml
m3ga has quit [Client Quit]
<pnou>
gl !
<pnou>
avec un peu de retard
duality has joined #ocaml
monochrom has quit [Read error: 104 (Connection reset by peer)]
<gl>
pnou !
<pnou>
gl !
<gl>
ah :)
<pnou>
:)
<gl>
wassup ?
<pnou>
ben là je m'occupe de mon mémoire de DEA
<gl>
mitou
<pnou>
enfin pas très activement comme tu le vois :)
<pnou>
tu soutiens quand ?
<gl>
le 10, rendu le 3
<gl>
toi ? ta chef de stage est devenue moins con avec le temps ?
<pnou>
heu ben, en fait je l'ai pas vu depuis presque deux mois
<pnou>
elle est en vacances
<pnou>
elle rentre le 15
<gl>
sympa, dis donc
<pnou>
ouais
<pnou>
t'es dans les temps ?
<gl>
non :)
<gl>
enfin, le bon point c'est que je dois faire 30 pages au plus
<pnou>
je me sens moins seul :)
<gl>
j'ai deja 20 pages, sans les resultats ...
<gl>
(parce que j'ai pas encore de resultats ...)
<pnou>
c'est sur quoi déjà ?
<gl>
mais ca avance, petit a petit l'oiseau fait son nid
<pnou>
au moins en maths c'est pas compliqué, t'obtiens pas de résultat et tu recopies un bouquin
duality has quit [Read error: 110 (Connection timed out)]
<gl>
a la base c'etait "translation binaire optimisee par simulation", mais le co-encadrant s'etant barre en cours de route en effacant les sources du projet, ca a change en "faciliter l'optimisation de code pour processeur VLIW"
<gl>
en gros, faire un compilateur/simulateur generique pour processeur VLIW
<gl>
reciblable et tout
<gl>
impossible a faire dans les temps, mais bon, ma chef pense que si
<gl>
elle me met a fond la pression parce qu'elle veut que je reussisse, sinon les gens vont se demander pourquoi ca chie, et la ils apprendront que le mec qui s'est fait la malle est son thesard de 3eme annee
duality has joined #ocaml
<gl>
ce qui la discrediterait legerement, n'est-ce pas.
<pnou>
mouais, ils ont souvent à penser que tout est faisable, surtout quand c'est pas par eux
<pnou>
arf
<gl>
d'autant plus que ce meme thesard a ete soutenu par elle au moment des allocations de bourses MENRT parce qu'il a eu une pneumonie le jour des soutenances (comme par hasard) et n'a pas pu presenter son taf
<pnou>
c'est pire que dallas le milieu de la recherche
<gl>
le dindon de la farce c'est necessairement moi, dans cette histoire
<gl>
et l'an prochain, tu fais quoi finalement ?
<pnou>
je sais toujours pas
<gl>
uh
<pnou>
y a de grande chance que j'abandonne la théorie des modèles
<pnou>
mais pour faire quoi... je sais pas
<gl>
du java :)
<pnou>
arg !
<pnou>
j'espère que non :)
<pnou>
c'est chiant de pas pouvoir continuer ses études éternellement
<gl>
c'est sur
<pnou>
et toi, tu vas faire quoi ?
<gl>
une these
<gl>
mais je change de labo, je change meme de discipline :)
<pnou>
dans quel labo ?
<gl>
je me rapproche des maths et de la physique, je sens que je vais souffrir, pour me mettre a la page
<gl>
le labo communication et electronique de l'enst
<pnou>
ça commence à me saouler les sujets de maths super abstraits complèmement déconnectés de tout
<gl>
a la fois je flippe un max parce qu'a mes entretiens avec les mecs je pigeais pas le 1/10 de ce qu'ils disaient, et en meme temps je sens que ej vais apprendre des milliards de trucs qui recoupent plusieurs disciplines
gim has joined #ocaml
<gl>
oui c'est sur, faut etre motive dans ton cas
<pnou>
bah c'est normal que tu y piges rien, pas de quoi se faire du soucis
<gl>
non mais c'est pas trop mes domaines quoi
<pnou>
ouais et faut avoir la foi
<gl>
j'ai recupere le demazure, la, pour me mettre a niveau en algebre, mais bon
<pnou>
et je commence à le perdre
<gl>
par contre les gens la-bas ont l'air super sympas
<gl>
allons allons
<gl>
tu veux pas faire de l'info? :)
<pnou>
ha t'as besoin d'algèbre un peu évolué ? cool ça
<pnou>
gl: ben justement j'y pense
<gl>
ben je sais pas si c'est tres evolue, mais y'a une grosse composante crypto
<gl>
<flood>
<gl>
Pour te donner un aperçu, il s'agit d'utiliser les
<gl>
idées de la "conservative logique" pour tirer des conclusions quand à
<gl>
nature et à la réalisabilité de fonctions combinatoires d'une "certaine"
<gl>
non-linéarité. Ce paradigme est utilisé (sans que les auteurs ne s'en
<gl>
rendent compte) dans de nouveaux paradigmes de chiffrement (les codes de
<gl>
scrambling), tels que DeKaRT. La même conclusion s'applique aussi à la
<gl>
condensation (SHA-1, SHA-256, etc.). Mais on reparlera de cela quand tu
<gl>
commenceras à proprement parler ta thèse.
<gl>
</flood>
<pnou>
si t'as des questions d'algèbre, n'hésite pas, je me sentirai utile :)
<gl>
ah merci :)
<pnou>
hum je crois que ce genre de thèse me plairait bien
<gl>
ca serait plus ton domaine que le mien
<pnou>
heu, tu veux que je montre mon mémoire voir si ça a un rapport ? :)
<gl>
j'ai vu que 3 personnes pour le moment, dans l'equipe: un cryptologue, un physicien, et un matheux qui a fait de la physique quantique
<pnou>
le pire c'est que je suis sérieux, j'y comprends pas grand chose à tout ça, mais apparemment ma directrive non plus, je me sens moins seul
<gl>
tu dois rendre le tien a quelle date ?
<pnou>
je sais pas trop
<pnou>
quelque chose du genre dernière semaine de septembre j'imagine
<gl>
ah ben ca va, encore
<gl>
moi j'ai plus que quatre jours
<pnou>
à vrai dire j'aimerais bien qu'il me reste plus beaucoup de temps
<pnou>
bon, quatre jours c'est pas assez
<pnou>
mais une dizaine de jours ça serait bien
<pnou>
histoire d'avoir enfin des vacances
<pnou>
c'est dur d'avoir un poste de maitre de conf en info ?
<gl>
je sais pas trop
<gl>
a mon avis c'est un peu moins la deche qu'en maths, quand meme
<gl>
et visiblement en maths le localisme joue a fond, pour l'obtention de poste
<pnou>
sauf que la plupart des labos refusent de prendre quelqu'un qui a fait sa thèse dans le labo
<gl>
ah ?
<pnou>
ça doit pas être propre aux maths d'ailleurs
<gl>
ca depend des labos
<pnou>
c'est ce que ma directrice m'a expliqué
duality has quit ["Don't talk to those who talk to themselves."]
<gl>
t'es a paris7 ?
<pnou>
d'après ce qu'elle me disait c'est très répandu
<pnou>
ouais
<gl>
ah bon, ca doit pas etre la meme chose avec les postes d'ATER/demi-ATER/Post-Doc alors, mon frere, sa copine et la mienne ont fait des demandes cet ete, et ils ont tous ete acceptes dans leurs propre labo, meme s'ils n'y vont pas forcement
<pnou>
peut-être
<pnou>
ils étaient dans quelle fac ?
<gl>
nanterre, orsay et orsay
<gl>
en droit, maths et physique
<gl>
mh en fait en physique le labo etait a orsay en etant paris-7
<pnou>
il me semble qu'elle m'avait justement dit qu'Orsay était très strict à ce sujet
<pnou>
bon, je vais m'y mettre...
gim_ has joined #ocaml
gim has quit [Read error: 110 (Connection timed out)]
Lemmih_ has joined #ocaml
jyrix has joined #ocaml
<jyrix>
hi again
<jyrix>
what's the difference between sml and ocaml?
z|away has left #ocaml []
<mflux>
does there exists operator < similar to ==?
<mflux>
say, if I wanted to build a Map that is based on physical equality, not structural..
<jyrix>
what's operator ==?
<mflux>
it tests for physical equality
<jyrix>
what's = then??
<mflux>
structural equality ;)
<ronwalf>
structural equality
<ronwalf>
How would a < work in a physical way?
<mflux>
it would use some secret information, maybe their physical addresses
<jyrix>
could i get an example?
<jyrix>
i still dont get it
<mflux>
(ref 4) == (ref 4) is false
<jyrix>
ooh
<mflux>
(ref 4) = (ref 4) is true
<jyrix>
okay
<jyrix>
what's the difference between sml and ocaml?
<ronwalf>
There is !=
<mflux>
yes there is, but you can't build a tree with !=
<mflux>
or atleast it would be very dull tree
<ronwalf>
Well, you can, it's just very boring
<ronwalf>
Indeed
<jyrix>
why would you want to build a tree based on physical equality?
<jyrix>
is sml > ocaml or not
<mflux>
physical equality is sufficient for me, no need to go through the data structure to check further
<grirgz>
jyrix: ocaml > *
<ronwalf>
ocaml > ocaml ?
<jyrix>
okay now im confused
<mflux>
* = everything but ocaml?
<grirgz>
ocaml > [^(ocaml)]
<jyrix>
really confused
<jyrix>
not anymore!
<jyrix>
good ol' regex
<jyrix>
but how are they different
<jyrix>
sml and ocaml i mean
<jyrix>
there are lots more people here than in #sml
<jyrix>
although none of you is very active or helpful
<mellum>
We only lurk here to occasionally bash newbies
<jyrix>
well could you atleast bash me then?
<jyrix>
how's ocaml != sml
<mellum>
No standard. Only one implementation. OO cruft.
<jyrix>
and that's it?
<mellum>
That's about it.
<jyrix>
does anyone use that oo stuff
<mellum>
Yes, for some things it's useful.
<mellum>
I don't, though.
<jyrix>
i still dont quite get how you can do anything without oo
<jyrix>
if a signature is an interface for structures why couldnt they call it an "interface"?
<Smerdyakov>
If an interface is a signature for structures, why couldn't they call it a "signature"?
<jyrix>
oh i get it
<Smerdyakov>
ML is older than Java by a few decades.
<jyrix>
sig doesnt collide with int
<Smerdyakov>
Java should be using "signature." :P
mimosa has joined #ocaml
<jyrix>
what's a "function" from structures to structures?
<jyrix>
and dont tell me its a functor
<Riastradh>
...why shouldn't he tell you that?
<jyrix>
cause im dumb
Nutssh has joined #ocaml
Nutssh has left #ocaml []
<jyrix>
could some please explain this to me
<jyrix>
Functors are ``functions'' from structures to structures. They are used to express parameterized structures: a structure A parameterized by a structure B is simply a functor F with a formal parameter B (along with the expected signature for B) which returns the actual structure A itself. The functor F can then be applied to one or several implementations B1 ...Bn of B, yielding the corresponding structures A1 ...An.
<jyrix>
waitwait!
<jyrix>
nah.. i still dont get it
<ronwalf>
Function takes a structure, returns a different structure based on it?
<jyrix>
and that's it?
<jyrix>
that cant be all
<ronwalf>
In a functional way, of course
<jyrix>
there are much more words in that than there are in your answer
<ronwalf>
Well, if you're providing a mathematical basis for your language features, then yes
<jyrix>
omg!
<jyrix>
it's teh manual
<jyrix>
you could have told me its the manual
<jyrix>
so ocaml is sml with oo stuff?
<jyrix>
and therefore ocaml > sml
<jyrix>
this is much harder than brainfuck
<Riastradh>
jyrix, please define what you mean by 'OO.'
<jyrix>
that "class" keyword
<Riastradh>
So is Haskell OO, because it has a 'class' keyword?
<jyrix>
dunno
<Riastradh>
(Hint: no.)
<jyrix>
alright it has inherit aswell
<Riastradh>
(I don't think anyone would agree that Haskell is 'OO.')
<Riastradh>
So do Haskell type classes.
<Riastradh>
My point is that particular keywords of a language don't matter; that's just syntax. I'm asking you about a semantic definition.
<jyrix>
<mellum> No standard. Only one implementation. OO cruft.
<jyrix>
that's what i mean
<jyrix>
are those all the differences
<Riastradh>
Pretty much.
<jyrix>
but ocaml seems to be more popular
<Riastradh>
How do you measure popularity?
<jyrix>
more people using it?
<Riastradh>
How do you measure this use?
<jyrix>
do i really have to define everything i say
<Riastradh>
Everything that is, without definition, otherwise meaningless, yes.
<jyrix>
well umm.. you ask people?
<jyrix>
lemme check the dictionary
<Riastradh>
You have conducted a survey of programmers to see who uses SML and who uses OCaml?
<jyrix>
:d
<jyrix>
:D
<jyrix>
no
<jyrix>
when i did that id find that they're both dead
<jyrix>
lol
<Riastradh>
So how are you measuring the use of SML vs OCaml?
<jyrix>
argh!
<jyrix>
is this "bashing newbies"?
<jyrix>
do you think ocaml is more popular than sml or not?
<jyrix>
yes/no, no defining anythin
<Riastradh>
No. I'm just somewhat indirectly trying to show you that your statements about popularity are rather empty.
<Riastradh>
I honestly don't know.
<jyrix>
well that's it! all i wanted to know
<jyrix>
thank you
<Smerdyakov>
OCaml is full of undesirable hacks.
<Smerdyakov>
That is why I prefer to use SML when possible.
<jyrix>
what kind of hacks
<mellum>
Smerdyakov: which implementation do you use?
<Smerdyakov>
==
<Smerdyakov>
mellum, MLton and SML/NJ
<jyrix>
== is a "hack" ?
<jyrix>
wow
<mellum>
Hey, Scheme has three kinds of equality :)
<Smerdyakov>
Yup. It would be even worse if anyone had any notion that there could some day be multiple OCaml implementations.
<Smerdyakov>
== would probably allow a program to figure out which implementation it's running in.
<Smerdyakov>
Using very underhanded means.
<jyrix>
well if it werent made by people like you, the compiler would just say who it is
<Smerdyakov>
It does.
<Riastradh>
mellum, Scheme also doesn't have a fascistly strict static type system for them to break it.
<Smerdyakov>
But you can probably use == to tell the difference between different garbage collection algorithms, for instance.
<Smerdyakov>
And that, of course, means your program might work with one algorithm and not work with another.
<jyrix>
well okay thats kinda hackish
<Smerdyakov>
Of course, you're used to this from C, Java, et al., I'm sure, but that's no reason to settle for it in ML.
<jyrix>
but optimizing and taking advantage of new cool features is okay with me
<Smerdyakov>
Pointer equality is not a "cool feature."
<mellum>
Smerdyakov: so don't use it, then.
<mellum>
Doesn't strike me as a large problem.
* jyrix
agrees
<Smerdyakov>
mellum, it makes program analysis systems for OCaml significantly more complicated.
<jyrix>
a new super-fast garbage collector is a "cool feature"
<Smerdyakov>
OCaml has never had a super-fast garbage collector.
<mellum>
Smerdyakov: Does anybody really do that? Doesn't seem like the ideal target
<Smerdyakov>
mellum, yes. French people do.
<mellum>
Smerdyakov: Okay, except them of course :)
<Smerdyakov>
mellum, why would I want to use a library when I don't know if its functions will respect referential transparency? I prefer the language to _force_ them to do this.
<jyrix>
well i dont like languagues that force me to do stuff
<jyrix>
id like them to recommend, but not force
mimosa has quit [Read error: 104 (Connection reset by peer)]
<Smerdyakov>
jyrix, no offense, but most everyone who thinks like that turns out to be a mediocre programmer at best.
<Smerdyakov>
jyrix, programming is HARD. Someone who understands the issues wants as much automated help as possible in avoiding mistakes.
<jyrix>
but its not always helping
<Smerdyakov>
It's not always helping _for_your_current_style_.
<jyrix>
lol
<Smerdyakov>
You can change your style, and in doing so optimize your productivity.
<mellum>
I agree with Smerdyakov. It's really helpful if you can absolutely rely on something not being possible instead of just hoping it won't happen because it's bad style.
<jyrix>
what's caml light?
<Smerdyakov>
A toy
<jyrix>
so sml doesnt have == or any other (")hacks(")?
<Riastradh>
SML has hacks that are just as bad: the built-in operator overloading cruft.
<jyrix>
ooh operator overloading!
<jyrix>
so thats umm.. bad?
<mflux>
jyrix, don't get exited. with ocaml you can define operators of your own!
<jyrix>
yay!
<jyrix>
always wanted that
<jyrix>
what about scala?
<Riastradh>
The concept alone without principle like Haskell is bad enough, but by 'built-in' I meant that it's totally inextensibly encrusted into the system and you can't define your own.
<mellum>
You can have great operators, like <<<-<<$!@!!!
<Riastradh>
...er.
<mflux>
mellum, funny, I use that operator for everything.
<Riastradh>
By 'without principle like Haskell,' I referred to Haskell's principles of 'overloading,' not its lack thereof.
<Smerdyakov>
Riastradh, OCaml has _many_ more hacks.
<Riastradh>
Smerdyakov, I didn't claim otherwise.
<jyrix>
its ocaml for me then, id really like operators like that
<mellum>
Maybe we should just go and write our own ML clone.
<Smerdyakov>
jyrix, actually, SML is even more flexible about defining your own operators.
<Smerdyakov>
jyrix, you can use words, not just symbols.
<jyrix>
ocaml doesnt do words?
ADD-101 has joined #ocaml
<mflux>
yeah, that's a stupid limitation, considering ocaml has 'mod'-operator
<Smerdyakov>
I think infix operators have to be symbolic in OCaml.
<Smerdyakov>
mflux, no, it's a good one. SML fixities are a big pain with separate compilation.
<jyrix>
still, what do you think about scala?
<mflux>
smerdyakov, oh, so ocaml learned something ;)
<mflux>
but it's still inconsistent to have buil-in operators built that way