| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9
Comments:
<0> you?
<1> I dunno, afraid to claim fluency in anythng, but I'd say C++ and python
<1> lisp soon hopefully
<1> but, my point was that I rarely find people fluent in C++ and also in another language
<1> makes me sad
<0> I like many parts of lisp
<1> don't you feel C++ is somewhat, I dunno, old-fashioned?
<1> it's somewhat constraining, perhaps?
<0> it's unfortunate (re: C++ and not another language), but I find C++ very liberating
<0> it supports so many different paradigms
<0> you could even claim fluency in c++ templates as a separate language
<0> or with the c preprocessor
<1> yea, I suppose
<0> I don't find c++ old fashioned at all
<1> you don't feel that coding in C++ doesn't "flow" as easily as it should?
<0> that I do feel, but no worse than many other general purpose languages
<0> but I'm working on my own toy language, and improving that is one of the primary goals
<0> maybe I'll be done in 10 years :P
<1> oh sweet
<1> haha, hurry up
<1> I dunno, since learning Python I'm finding myself coding in Python when I can, C++ when I have to
<1> it's this flow thing
<0> python has things going for it, but I can't/haven't gotten into it
<1> oh?
<1> hm
<1> how will your toy language "fix" C++?
<0> it's not attempting to fix c++, though it is borrowing heavily
<0> I haven't worked on it in a couple weeks, let me find some of the theoretical examples
<1> well surely there must be some things that you can consider better done your way?
<1> otherwise you wouldn't work on it would you? :)
<0> http://dump.qxxy.com/temp/examples/
<0> first cl*** functions and arrays (including literals), for one, a little more relaxed variable declaration
<1> first cl*** functions and more relaxed variable declaration
<1> I'd say those thigns fix C++
<0> most of it is just syntactic sugar, I don't know enough about the higher order things yet
<0> delegated constructors
<1> what do you mean delegated constructors?
<0> struct S : B { S( int ); S() : S(3) {} };
<0> along the lines of "if a ctor has the ctor for the same cl*** in its init list, it cannot have anything else in its init list"
<0> s/has the/has another/
<1> ah
<0> so the specified ctor is used to initialize the object, localize code
<1> although what do you mean by the struct S : B notation?
<0> without having to use hacks that put member variables into intermediate states and a separate init function (which iostreams use)
<0> that example code is practically c++, B is a base cl*** of S that S() doesn't have to initialize (because S(int) does)
<1> ohoh, ok
<0> I probably won't be using the type name for ctors and dtor
<0> and I'm thinking about a separate dtor mechanism, possibly for any exception, but geared towards the "recovery is the only startup" idea
<1> mm
<0> that is, an application/code is never 'started from scratch', you're always recovering from the last run
<1> right
<0> and hence, dump and recover should be as cheap as possible
<0> (but no cheaper :P)
<1> aye
<0> learning a lot from reading d&e
<0> the long plan is to build my own system once I have the language and tool chain
<1> "system" meaning..?
<0> OS, tools, applications
<1> ah
<0> it'd be nice to do the physical stuff too, but that's on an completely different level :P
<1> haha
<1> mmm ok, well 530am, I'm off to bed, good luck on your stuff
<2> anyone here used IBM Rational Rose?
<2> coz i can't install the bastard
<2> my system meet the requirements but it says i did not
<2> i tried MS Visual Modeler and it really sux
<2> i can't find the sticky figure, the actor
<3> _m_
<3> here?
<3> lol
<4> Nach dem Tod der Mutter ?bersiedlung nach Wien, wo er sich zweimal vergeblich an der Kunstakademie bewirbt
<4> ja
<4> ?
<5> hy
<5> I have a array a[10] , and I want to ***ign to every a[i] a string from an enum , how should I do that ?
<6> Back.
<7> std::vector<string> v; for (int i = 0; i < 10; ++i) { v.push_back(boost::lexical_cast<string>(someEnum)); }
<7> std::string
<5> thanks, but I don`t really understand :)
<5> I have:
<5> enum culoare {alb,negru,gri,verde};
<5> and I want to ***ign to a[i] , culoare[2] for example
<7> oh you want the actual string representation? you're going to have to create a function that converts the enum's value to a string representation, as there is no ***ociation of an enum and it's identifier at runtime
<7> std::string culoare_to_string(culoare value) const { switch (value) { case alb: return "alb"; case negru: return "negru"; ... default: throw whatever(); } }
<5> ok, thanks a lot
<7> (the const qualifier ***umes it's a member, btw, but it probably makes more sense as a global)
<7> man, i feel like crap
<8> you know...
<8> i **** at c++ :P
<7> was up too late trying to coax vista into not screwing up my tv tuner
<7> it's not really vistas fault, as i saw the exact same behavior in the x64 nvidia driver
<7> (winxp x64)
<7> and it's also likely the crappy tv tuner software's fault for not doing something right with the drawing surface; basically, i get the video, but the colors are all screwy
<9> NineVolt's request is probably one place where the stringize macro is useful.
<8> i have a request?
<9> I mean, VulturU
<9> Stringize can't help you ****ing at C++
<8> indeed :P
<8> though i guess it's more a design problem than a c++ one.
<10> :)
<8> i have a pointer to an array of pointers. would i declare it as myType **list; ?
<11> yes
<7> stringize, pfft
<8> and by array i don't mean array.
<11> you still mucking w/ arrays?
<8> heh
<8> yea i'm building my own list cl*** :P
<11> what DO you mean?
<11> oh god
<12> uh-oh. how do I kill a service run as LocalSystem?
<8> i just need a block of memory that can be resized.
<8> for my neural net, i'm recoding it from scratch.
<11> NineVolt why are you doing this?
<11> so use the standard library
<8> the one i had before used a fixed-size table to represent the digraph.
<11> writing all you own stuff is folly
<8> it was max_neurons * max_neurons big.
<8> the bigger the network, the more memory it wasted.
<8> i'm looking to move to lists to try to save some memory while not slowing it down too much.
<8> i think i've got the memory allocating now at least.
<13> ninevoly: how bix is one item in the max_neurons*max_neurons map?
<13> *big
<8> hm
<8> floats.
<8> they're all floats.
<8> using lists, they're gonna be addresses -and- floats.
<11> NineVolt what are you trying to represent with this coollection
<8> so while each interconnection will take up more memory, there'll be a -lot- less recorded interconnections.
<8> i'm representing the presence (or lack thereof) of interconnections between neurons.
<8> and the weight (float) of the interconnection, if present.
<11> that sounds like a graph
<8> right.
<8> that's what i said before :P
<11> maybe you should use one of the graph frameworks (BGL for example)
<8> no :P
<11> why not?
<7> wtf
<7> the tv tuner just started getting the correct colors
<8> i need to do something with the data while i traverse the graph in a very particular order.
<7> yay now i can keep vista
<7> no problems
<11> NineVolt I don't see your point, but whatever
<8> i'm not sure how well a third party graph framework will get along with neural net backpropagation training.
<8> training is sorta part of the traversal.
<11> well what you got doesn't
<8> o_O
<8> what i got now doesn't do much of anything. but that's because i just started yesterday.
<8> my original one does, yes.
<11> "doesn't work", "don't know if it works" ... which shall I pick
Return to
#c++
or
Go to some related
logs:
phpize possibly undefined macro: AC_MSG_ERROR
#squid
#squid
wine _dl_out_of_memory
#linuxhelp
#linux
#javascript
moon rise EST
#chatzone
#chatzone
|
|