2007/10/12
IRC [01:09] *** J__ joined the chat.
IRC [01:10] <unniiii> anyone here
IRC [01:12] *** unniiii parted the chat.
IRC [09:12] <Dossy> Heh. People's odds of getting a response increases the longer they stay. :-)
IRC [10:06] *** knifehat_ joined the chat.
IRC [10:16] *** knifehat parted the chat.
IRC [12:06] *** holycow joined the chat.
IRC [12:25] *** uniiiii joined the chat.
IRC [12:27] *** uniiiii parted the chat.
IRC [12:30] *** uniiiii joined the chat.
IRC [12:33] <uniiiii> ???
IRC [12:33] *** uniiiii parted the chat.
IRC [12:35] *** uniiiii joined the chat.
IRC [12:38] *** uniiiii parted the chat.
IRC [14:32] <Spark> Dossy: would you be interested in reading our paper?
IRC [14:32] <Spark> deadline is tomorrow :)
IRC [14:32] <Spark> http://www.doc.ic.ac.uk/~dc04/other/misc/cc_cunningham.pdf
IRC [14:33] <Spark> it would be nice if we had more confidence that we're not making absurd claims about aolserver and the function of tclvar.c :)
IRC [15:21] <Dossy> I'll take a quick look at it now
IRC [15:23] <Spark> the aolserver stuff is near the end
IRC [15:23] <Spark> in a section titled "Experiment" or such
IRC [15:25] <Dossy> you use the phrase "atomic section" -- different from "critical section"?
IRC [15:25] <Spark> not really
IRC [15:25] <Spark> it's a bit more well-defined
IRC [15:25] <Spark> first paragraph of the introduction defines it
IRC [15:25] <Spark> ish
IRC [15:25] <Spark> basically it means "executes as if sequentially"
IRC [15:25] <Spark> and allowing other threads in parallel is a transparent optimisation
IRC [15:27] <Dossy> well, how would do you define "critical section" then?
IRC [15:27] <Spark> hmm the way it's laid out, it might look like Fig.4 and Fig.5 are part of Experiment whereas they're actually left over from the previous section
IRC [15:27] <Dossy> and how does that definition differ from your definition of "atomic section"?
IRC [15:27] <Spark> i'm not sure i've ever seen it defined
IRC [15:28] <Spark> i think the MSDN docs use it with the same meaning as "synchronised block"
IRC [15:28] <Dossy> lol, the notion of critsec's in concurrent programming is old. very old ...
IRC [15:28] <Spark> so is atomic
IRC [15:28] <Spark> 1976 i think :)
IRC [15:28] <Dossy> http://en.wikipedia.org/wiki/Critical_section: "In concurrent programming a critical section is a piece of code that accesses a shared resource (data structure or device) that must not be concurrently accessed by more than one thread of execution."
IRC [15:29] <Spark> right, that's different
IRC [15:29] <Dossy> Okay, how?
IRC [15:29] <Dossy> IMHO, it's dangerous to "invent" new nomenclature unless there's a really good reason to do it.
IRC [15:29] <Spark> ok well it doesn't define "concurrently accessed"
IRC [15:29] <Spark> but that may allow interleaving
IRC [15:30] <Dossy> What I'm getting at is the paper seems to essentially make claims around a technique for automated critical section inference.
IRC [15:31] <Spark> ok well we don't infer sections, we just fill them in with appropriate locks
IRC [15:32] <Spark> but
IRC [15:32] <Spark> i see what you mean
IRC [15:32] <Spark> you're asking why not call it a critical section instead of an atomic section, and have the syntax be critical { ... }
IRC [15:34] <Spark> ok i found a paper from 1974 that uses critical section to mean mutual exclusion
IRC [15:35] <Dossy> Anyhow, section 5 looks okay.
IRC [15:35] <Spark> cool thanks
IRC [15:35] <Spark> do you mind being added to the acknowledged? :)
IRC [15:36] <Dossy> reading the conclusion now -- IMHO, that's the part that really matters :-)
IRC [15:36] <Dossy> mind? not at all
IRC [15:36] <Dossy> in references, fix "Aolserver" to "AOLserver" please? :-)
IRC [15:36] <Spark> ah right yes
IRC [15:36] <Spark> bibtex has a mind of its own when it comes to capital letters
IRC [15:36] <Spark> but it can be overridden
IRC [15:37] <Spark> how shall i refer to you
IRC [15:37] <Spark> i think it would be funny to just put "``Dossy'' from freenode #aolserver", but it would look better to have a real name :)
IRC [15:37] <Dossy> Dossy Shiobara and the AOLserver project?
IRC [15:38] <Spark> ah it is your real name
IRC [15:38] <Dossy> nod
IRC [15:38] * Dossy nods
IRC [15:38] <Spark> cool
IRC [15:40] <Dossy> I forget--is this your PhD thesis or something?
IRC [15:41] <Spark> yeah, my favourite chunk of it
IRC [15:41] <Spark> if this gets published then it will become half my thesis
IRC [15:41] <Dossy> _half_? ouch. :-)
IRC [15:41] <Spark> whereas if it doesn't, we'll add some proofs and submit it to a theory conference
IRC [15:41] <Spark> the other half is already done :)
IRC [15:41] <Dossy> Ah! What's the other half?
IRC [15:41] <Spark> http://www.doc.ic.ac.uk/~dc04/universes_vamp.pdf
IRC [15:42] <Spark> that one didn't go down too well
IRC [15:43] <Spark> basically all the interesting stuff got done 5 years earlier, and we didn't have an implementation so couldn't quantify our alleged improvements
IRC [15:43] <Spark> it's a big chunk of theory though, took bloody ages, so it ought to go into my thesis
IRC [15:46] <Dossy> yeah, I'm on p. 18 and damn, I'm glad I didn't have to learn all this in order to be a decent concurrent programmer :)
IRC [15:46] <Dossy> too much greek in there. you were a math undergrad, weren't you
IRC [15:47] <Spark> nah
IRC [15:47] <Spark> i did a BSc in CS at warwick (UK)
IRC [15:47] <Spark> then the MSc at Imperial College
IRC [15:47] <Spark> also in CS
IRC [15:47] <Spark> which is where i learnt how to do maths :)
IRC [15:48] <Spark> i've always had a soft spot for maths though, it's less frustrating than programming
IRC [15:50] <Spark> although it's not as fun to do as a hobby
IRC [15:51] *** holycow parted the chat.
IRC [15:52] *** holycow joined the chat.
IRC [16:02] <Dossy> Grr. Why is there no decent free Windows software for video compositing/editing? Sgigh.
IRC [16:02] <Dossy> er, sigh.
IRC [16:10] <Spark> right, it's now in the hands of the reviewers
IRC [16:10] <Spark> thanks for your help, i'm off to the pub :)
IRC [18:18] *** holycow parted the chat.
IRC [18:18] *** IEF parted the chat.
IRC [18:18] *** knifehat_ parted the chat.
IRC [18:18] *** frankie parted the chat.
IRC [18:18] *** tekbasse parted the chat.
IRC [18:18] *** CIA-24 parted the chat.
IRC [18:18] *** jhavard parted the chat.
IRC [18:18] *** Dossy parted the chat.
IRC [18:18] *** Spark parted the chat.
IRC [18:19] *** Spark joined the chat.
IRC [18:19] *** Dossy joined the chat.
IRC [18:19] *** jhavard joined the chat.
IRC [18:19] *** knifehat_ joined the chat.
IRC [18:19] *** holycow joined the chat.
IRC [18:19] *** IEF joined the chat.
IRC [18:19] *** CIA-24 joined the chat.
IRC [18:20] *** frankie joined the chat.
IRC [18:20] *** tekbasse joined the chat.
IRC [19:10] *** holycow parted the chat.
IRC [22:21] *** uniiiii joined the chat.
IRC [22:24] <uniiiii> any 1 here
IRC [22:26] <uniiiii> ????
IRC [22:27] <uniiiii> any 1 here
IRC [22:28] *** uniiiii parted the chat.