http://www.valgrind.org

I believe your assessment is incorrect. My coworker sitting next to me happens to be quite an excellent programmer with a degree in mathematics. :)

Perhaps the biggest connection between maths and programming that was bestowed upon me was the notion that when a mathematician talks about abstraction, he is indeed talking about the exact same thing as a computer programmer or scientist: a fixed interface which works for a large field of objects.

Ever since I understood this, it has become much easier to grok maths (especially abstract ones.) Not because this gives me insight into the actual equations, but because it gives me insight into the structure and reasoning behind such mathematical abstractions.

http://fuse.sf.net

And since everybody has screenshots, I'll show one of my XMonad layout, although this isn't my work desktop; it's actually my PS3 running Ubuntu. I just made XMonad start in my ~/.xsession and made GDM run my init script on login rather than the default one which starts GNOME. I also installed a recent kernel on it since this picture to fix some PS3-specific linux issues in the kernel (I think now it's outdated though, it's running 2.6.31-rc2 or something, and 2.6.31 has already been released.)

http://www.nijoruj.org/~as/images/xmonad -ps3.jpg

At work I write code on Linux (Debian 5.0) AND Windows (with keyboard/mouse shared between screens with synergy - I recommend it.) I also write code on Mac OS X at home, which is fairly unix-y.

First: I do not believe your options for programming language choice is more restricted on Linux/OSX. If you want your .NET languages, you can have them. If you want your scripting languages, you can use them (AND you can use them on Windows too, you know.) C/C++ is fine. Java is not hard to install on either. Plenty of languages run on the .NET and JVM, which are available for both platforms.

I do not generally find the tools more restrictive than Windows either. I think that the effort invested in learning to use your command line and the tools it has effectively is very much worth it in the long run. If you are used to the command line (cmd.exe) on windows, know it is essentially the most inferior command line shell on earth. Shells like ZSH have great features that make life easy. I think that doing things the 'unix' way of just having commands and piping them through each other is a good principle and works in practice: some of these individual tools suck, but at least they may only suck at *one* specific thing rather than a multitude of them ;).

Lots of your tools on Linux are free. There are some very very good alternatives to the pay-for development tools on windows, but like all things of this vein, there aren't always good free analogues to the pay-for stuff and vice versa.

To mention some tools on both Windows AND Linux:

* I recommend at the very least having a good Text Editor that you know well, that's available on ANY platform. I naturally suggest you try either Vim or Emacs, because they are available on almost every operating system you could think of using, they're very powerful and free.
* Visual Studio 2008 (I use Team System) is awesome. I totally recommend it if you write code on Windows. I really like it as an IDE, it's very powerful, but it does come with a learning curve. The IntelliSense for C++ is apparently going to rock in VS2010 from what I hear, which will hopefully make working on our ~60kLOC C++ code base easier!
* For compilers: a recent GCC (like, 4.3 or 4.4 if possible,) or MSVC on Windows. MSVC sometimes lets non-standards compliant things slide in my experience and has some slightly annoying warnings, but they're both very good.
* I recommend you *do* learn a scripting language if you choose to use Linux. Even shell script would suffice (for an example, see: MapReduce as a Bash script!) It's useful so you can automate lots of shit you otherwise won't want to do again.
* IF you choose to write lots of C/C++ on Windows and/or Linux (but 'if' because you can probably spend your time better!,) despite what anybody tells you, the APIs have quirks and some apparent corner cases and are weird in some spots. There's also the problem sometimes of API fragmentation and you're not sure what's available where. Some of it just sucks.

For things about my linux system that I write code on at work:

I use Debian 5.0 (kernel 2.6.26 I think), with xmonad as my window manager. Soon I'll be getting a nice lenovo thinkpad x200 though, on which I'm planning on installing ArchLinux (so I have a recent enough kernel to have stable wireless drivers, Debian lags behind in that respect)

Use a profiler? Use a library that is designed to be highly performant for text processing?

Of course, it brings into question whether you really need to go from 9 seconds to 5 seconds - if you have to do a lot of contortions it might not be worth it, but that's a decision that's dependent on a lot of factors.

At 10/16/09 08:00 PM, smulse wrote: So in the next week or so, I'll own half of an Events Management Company. Along with carrying on freelancing as a Lighting Designer/Technician.

Hows everyone else doing for themselves?

Currently full-time employed at my company working as a software developer (and our company actually ships a software product, we programmers are not simply there because IT is a 'necessary evil' like so many other places) - 401k, stock options, full health coverage, a pretty good annual pay (well, it's only my first job as a developer so I think it's pretty good,) awesome & funny coworkers, etc.

For a sophomore in computer science @ uni I think this is pretty good. :)

The ease of development leads me to say python or ruby, because they are simple and abstract.

There are bindings to libraries like libSDL from these languages, which I would recommend using (SDL has been used commercially,) and it will severely cut down the amount of time you need to spend on unimportant details (like setting up a window etc.)

C++ is probably worth knowing if you're die-hard set on writing games I guess, but it is a very complicated language compared to almost any other and takes much time to master. It has a much better type system than C too, which, in comparison, is very very welcome.

I would recommend if you go this C++ route to use something like SDL and boost, because they will save you a lot of time doing stuff you really don't need to/shouldn't re-implement (we use boost at work - some parts of it are godawful horrific, others are incredibly useful.)

Hope it helps.

Haven't posted in a while. Been really busy lately. I do remember all you NGers!

Mainly busy because now I have a contract job writing software (with good pay + badass computer) that I started on Wednesday! There are boring and really cool parts to it, and there are only like 6 developers including me, so the work and hours are very flexible, all my coworkers are awesome, and yeah, the pay is pretty nice for a first real job!

I can't complain much other than that, only that school starts again in approximately 1 month, and my schedule this next semester will be busier than ever with work + this new job. But, you know,

TO INFINITY, AND BEYOND!

At 6/15/09 03:22 PM, kiwi-kiwi wrote: There is lots of BS on the internet, but you can easily filter it out if not by common sense at least by trial and error.

You can certainly spot erroneous information on the internet. 'easily' however, is not always the case. Also, you make the assumption that these erroneous pieces of information are things like API references or syntax examples or something I think, when that's not the point. An article can have perfectly correct, working code written in it, and still misrepresent ideas and information in a way that is stupid.

It's not simply that black and white, I'm afraid.

Also on the quality of information, you're making the assumption that if you pay for something you're going to get a better quality, which is not always correct.

When did I ever say you always get better information from books, because you pay for them? I don't believe I ever did.

Of course the guy that's making the book has to meet a high standard so he can sell the book and make money, but there are other reasons to make comprehensive guides ( for instance if you want people to use you software/programming language ) like adobe labs, msdn.

MSDN and a book on a specific topic or idea are not the same thing. There is a huge difference between API/library documentation on how to create a popup window, and things like actual written material that's about a particular subject of interest like such as artificial intelligence, or mechanized computer logic and theorem proving.

Let's say you're doing some project on .NET what would you rather do, search in a book you read or go to msdn?

This is a straw man argument. See above.

And as a conclusion. I am under the impression that programming isn't something you learn from books, you want to program, you have to stay in front of the compilers and try everything out, you just need a starting point and for that the internet is good enough

I never said programming is something you learn from books. I never said you should spend more time with books than in front of the compiler. I am merely pointing out that disregarding ALL books just because "the internet is there and free" or "some books are bad" is a pretty stupid conclusion to draw.

At 6/14/09 11:49 PM, kiwi-kiwi wrote: I'm just wondering why would someone pay for information freely available on the internet

Because maybe the same quality of information you would find in a book is not available on the internet whatsoever? Like, you know, just maybe the people who write stuff on the internet might sometimes be full of shit and know nothing, and just maybe the people who write books may write really good books because they are incredibly knowledgeable on the matter?

Yes, I know: it's insane to believe there could be bullshit on the internet, and maybe some books cover the literature better.

At 6/14/09 11:32 AM, kiwi-kiwi wrote: I never understood why people buy programming books when there are tons of free tutorials on the internet.
Personally I hate programming books because most of them are too basic and they assume a very low learning curve.

And a lot of stuff on the internet is either worthless tripe or just complete and utter bullshit. What's your point?

At 6/14/09 11:54 AM, Wonderful wrote: Structure and Interpretation of Computer Programs by Gerald Sussman

The first chapter compares programming to magic.

Be wary: SICP was and still is a mainstay in computer literature. Not many books - especially in the programming field - can literally stand on their legs for nearly two decades without any revision, when normally you get 3rd, 4th and 5th editions these days before you can say 'please.' It's a very good book and extremely deep with wide reaching implications, but it is also pretty fucking hard as well, so you may want to try something else like How To Design Programs first.

On the note of programming books, you may like other (seemingly unrelated) books on topics like Logic etc.. Logic I think is particularly easy to learn ('math without numbers,' those college students say) with a lot of real world applicability.

I also recommend you go pick up a book on Scheme/lisp or something while you can - most people tell me the book 'Lisp In Small Pieces' is excellent. It will probably be very 'different', but it will be very mind opening too hopefully, which is good in the end.

You know, I'd like to see a day where Linux suddenly got as much market share as windows has now. All the possible exploits would show as clearly as the sun, when your on the surface of it. I'd like to see the day when someone gives root access to a pirated application that has malicious code in it.

Well, by default linux DOES NOT give any regular user root permissions; you have to explicitly su or login via root to get that. So, by default a linux user is normally set up with little access outside their ~.

In the case of a regular user needing to do root-like shit, it's called 'sudo'. You can give fine-grained control to what applications users can and cannot run with root priviledges, so a user may be able to install packages with 'yaourt', but that's it - he can't "sudo rm -rf ~/", he can't "sudo cat /etc/passwd". He might not be able to uninstall packages either if you set up your sudo file properly.

On the other hand, with Windows, you pretty much were either Administrator or you weren't (not sure how much it has changed in vista/7, correct me if the following doesn't apply to them.) There is no point in denying the fact a large class of windows users simply have their account be the Administrator, BY DEFAULT, so they can do things like install software. There is no simple fine-grained access control mechanism to really give users certain permissions and deny them others in windows, and even then, the majority of people sure as hell probably don't use it, while in the case of e.g. ubuntu, sudo is by default the only way to get any sort of root priviledges. Yeah, okay, so even Administrator can't rm C:/WINDOWS/system32/ntoskrnl.exe, but by default Administrator can pretty much do everything else that could fuck up your system, so if someone has Administrator, there is still plenty they can do.

But other than that you seemed pretty right.

At 6/13/09 02:19 PM, iMate900 wrote: Read Wikipedia.

His point is that your trivialization isn't really all that important or worth mentioning.

No, essentially not a host application.

Have you read the Linux ptrace man page?

No.

Then ptrace helps the prog via kern.

Jesus christ, what the fuck does that even mean?

@CronoMan: Linux does have a viable way of looking into other processes and attaching to them - ptrace is the function which does exactly that (it's how you implement debuggers, etc..)

Read here if interested:
http://linux.die.net/man/2/ptrace

)

Stupid.

You clearly fail to see what Visual Basic did to how programmers relate to computer code

It is still too simple for newbies. I'd use FreeBASIC anyway.

Missed the point.

A simple way to hijack the host's request.

No its not, ActiveX is a module like anything else, if you say this, it goes for all "hookable" libraries, which includes.. well... everything.

The Story of Lib a, Lib b and Lib h
Lib a send request to Lib b
Lib h hooks Lib b
Lib h hijacks Lib a request and forges a fake return value

Just about every major operating system on earth has APIs, etc to do this programmatically in the general case, so you need to be more specific.

Visual Studio is too powerful for the newbie.

Agreed

Please explain, I'd love to hear this

VS.NET is now common, so .NET is in evrything.

What's your point?

Hahaha? Please provide me a link that says that.

SQL Server has tons of holes.

He said provide him a link, not 'your word' - either get results or shut up on the matter.

Also, go to #SQL on EFNet, start talking about MySQL and see how long it takes before they kick you out of the room. It's funny how the professional environment would never consider using MySQL for anything other than personal projects.

Go to #mysql, butt about SQL Server being powerful, and prepare to be kicked.

What? Of course you're going to get kicked out for shit talking MySQL in the #mysql IRC room. He's talking about going into a regular IRC channel and talking about getting kicked out for boasting mysql. Neither argument is really fair anyway because they don't say anything about the technical merits, but whatever.

phpMyAdmin is better.

Proof?

What fields would that be? Performance? No. Quantity of data? No. Transactional support? No. Full-text indexing? No. Reliability? No. Datatypes? No. Ease-of-use? No. Replication? No. Cross-domain support? No.

Performace: ubuntuforums.org's database has tons of writes/sec.
Reliability: ubuntuforums.org has to be 24/7. Mysql helps.
While not easy to be, still better than MSSQL.

What you just said means absolutely nothing.

Wow, yes, ubuntuforums is popular. Okay, how *many* database writes a second does it have? How do you know MySQL hasn't been their major bottleneck? How do you know MSSQL wouldn't give better throughput and connectivity/reliability?

You know none of these things, and furthermore you provide nothing in the case of MSSQL, making all your assertions all the more worthless (if that's possible.) On top of it all ubuntuforums is just one pathetic case of database usage in the real world (trivial to other places,) so you actually have no big-scale data to base your claims on either I imagine.

whoops!
Security is a problem in IIS. won't support mysql authection!

Then don't use MySQL?

XNA

Now destroying Macs, PSPs, PS3s, and Linux. From M$.

XNA is for Windows and XBox360, IT'S NOT FOR MAC, PS3 OR LINUX
Releasing it for the competition to use, is suicide. Do you like suicide?

a number of gamers are switching to mac and linux

Citation needed.

I write this from a security, user-friendly, anti-MS and Linux user's view, and some parts have one view placed in mind.

ARE YOU FUCKING SERIOUS?

You make me wait UNTIL THE END OF THE GODDAMN REPLY to tell me THAT NOTHING YOU SAID REALLY MEANT SHIT BECAUSE YOU'RE BIASED?

Fuck you kids these days.

Considering this isn't a remote exploit (or anything even close for that matter) that requires full access to a machine, I'm going to rate it as pretty much harmless, and say for the record that using it as the sole basis to determine the security of an entire system is pretty stupid (try more accurate titles next time.)

At 5/31/09 10:54 AM, kiwi-kiwi wrote: For instance if your average guy runs Linux at home just because he heard there are no viruses, you could just as easily bypass his security because he'll probably run everything as root.

Why in god's fucking name would you run everything as root on a Linux system? Are you insane? Pretty much every installer procedure (from ubuntu to gentoo) requires you to setup your own login credentials, and furthermore, there are few processes aside from system updaters/bootloader modifiers/daemons etc. that actually need root, and a lot of those processes may drop their credentials back down to being a regular user (lots of servers do this.)
I find this a wild hypothesis at best - finding people who run everything under administrator on Windows however isn't very difficult, considering people do it so they can install stuff; it's only that when it's like this, everything is run as administrator. Regardless, I still don't see how you can extrapolate that 'windows is insecure' from such a stupid trivial thing as this considering it isn't even a threat until after you already have full access.

I'm coming in late but whatever; fuck all who dare defy me.

a while ago, mongoid wrote

You can be the most creative person in the world, but you'll never be a doctor, engineer or a scientist without a college degree. And you sure as hell won't make it through college without a high school diploma.

Well, while in those very specific instances you are probably correct, it is worth keeping in mind that programming etc. is a field at which anyone can excel (google has some hired highschool dropouts, FYI) and do very well, even with a minimal amount of 'formal experience' in the craft (then again I feel kind of depressed about the CS programs at many US universities - this is why I'm thinking of changing my major to Pure Mathematics.)

I doubt you will be hired to work on the LHC or work with NASA's aeronautics team if you don't have a degree, but writing code - good code - and going about doing it is something that can be entirely self-taught. I've never taken a single computer science course in my life (well, okay, I did, but it didn't count because I didn't learn shit because it was in high school) and am completely self taught and I don't think I've come out that bad, to be honest (in fact I should hopefully be applying for my first actual code-writing-job-thing at the end of the summer - I WOULD GET TO WORK ON LINUX DRIVERS AND STUFF MAYBE LOLOLOL *nerdgasm*.)

I've also kind of come to the conclusion that most of the time when people put job postings up and they say stuff like 'Bachelors required,' they're many times very much willing to overlook that matter for people who will just take the fucking bait and apply with a good amount of confidence - I mean, in all honesty, some pretty shitty coders have come out of MIT's computer science department that not many people would hire. Nobody says it's all-or-nothing.

Also, many people are afraid of a question involving something like a degree (especially if they dropped out/didn't go at all,) because it seems very scary and critical, but it doesn't have to be - putting things in perspective (and on this note, showing confidence but not arrogance) can completely turn things around, e.g., if your interviewer asked you:

"Why did you drop out/not attend college?"

Which response would be better:

1. "I dropped out/didn't go because I hated school or it was dumb or something"
2. "I dropped out/didn't go because I didn't feel it comparatively had enough to offer me to working in the field/whatever, and I've compensated for the lack of my higher education in the field by doing... <blah blah blah>"

The first one shoots you to hell in a hand-basket pretty much right there. The second - if your case is stated correctly - could lead you to a job/second interview/whatever. Or maybe if you say the second one they'll tell you to just get the fuck out and never come back - but would you have had a BETTER chance if you went with response #1? Doubtful.

And with the advent of things like open-source software (and even cheap, open hardware like the Arduino or BeagleBoard,) it becomes much easier to publicly state your case, your skills and your contributions by listing off real-world projects. In fact, I think that's one of the best things you can possibly do to get better - find some open source project you like and work hard as hell at it. It really does make a good impression when you can gloat about your skills in a situation like that and you have conclusive proof.

I agree with your general point, but the field that people like us are involved in is a very strange one in comparison to most others, in a few ways. So, in a discussion like this, my point is that fact is worth keeping in mind.

Anyway,

mongoid also wrote:

On another note, college is incredibly useful in developing relationships with people who you may work with (or work for) in the future.

This I very much agree with - I think it's important people go to college because you meet a lot of really good people, which I think can weigh out to be worth *much* more in the long run (let's face it: much of it is just in who you know.) I've just finished my first full year, and I've made some very long lasting friendships and met many people who've helped me out, or want to work on projects with me. It's very much an amazing place to make friendships and meet people with similar interests, etc..

At 4/30/09 08:58 AM, Alphabit wrote: money.

It seems to come down to this most of the time, yes.

Anyway, I was working on a project which I plan to make some money on and I decided to incorporate that editor into my work - this will make the project much easier to complete. But the problem here isn't mine; I'm actually taking advantage of other people's work... I mean technically, I'm not allowed to explicitly charge a client for this section of open-source code; but that certainly doesn't mean that I can't leverage the value it brings to my system and thus make money from it indirectly.

Most open source licenses allow you to sell your code - there is absolutely 100% nothing illegal - or immoral - about this.

Certain licenses have different requirements. It is just like a drivers license: I give you this code, and it has this license attached. I hold the copyright, and it allows you to use the code, per some regulations. For example, the GPL requires that you must distribute changes to source code if you distribute the code in e.g. binary code. This means you must allow users to be able to see the changes. If you violate the terms of these licenses, the license can be revoked, which means you cannot use the code either.

Other licenses like BSD simply require that you keep the license and distribute it with the code (that you can't remove,) and give the author a mention. So the BSD is the more 'free' license. Neither prohibit you from selling the modified code.

It just seems to me that a lot of open source projects don't make sense and people are getting ripped off; maybe they think that they're being nice by letting other people use their code, but the truth is that all work has value and if the producers of the work aren't reaping the rewards; then you can be sure that someone else is.

Why do you say they are getting ripped off? Because someone else used their code and made money? They should be fully aware of the terms under which they release their own code - I don't think they feel 'ripped off'.

but I think a lot of them just don't make sense from a business perspective.

Then why are more and more governments talking about and in some cases adopting open source? Years ago, it was written off as a means of destroying all business credibility and was considered completely unrealistic, and even a threat; now, it's perfectly feasible to be paid for writing open source and for companies to adopt it and support development. It's happened many times. I'm not saying it's perfect, but it can get the job done.

It's been a very tiring semester. My mother died very early april, and now things are starting to grind as finals come up. :( I just want to get it over with.

How have other regs been doing?

on the other hand if the compiler can do constant propagation, it might only be a single instruction to get your results (and may be replaced with a bitshift or something,)

What I meant to say was:

"on the other hand if the compiler can do some optimization like constant propagation on the division instruction, it might only be a single instruction to get your results (and may be replaced with a bitshift or something,)"

At 4/29/09 02:42 PM, kiwi-kiwi wrote: I would go for the If on this one because when it comes down to binary level when you do an if, you computer just has to look at some bits.

Multiplication on the other hand...let's say you have two numbers on 8 bit precision: 10 - 00001010 and 11 - 00001011
Multiplication in binary goes mostly like multiplication in decimal:

00001010 *
00001011
------------
00000000
00001011
00000000
00001011
------------------
01101110

And you get 110 = 01101110 in 8 bit precision binary.
As for division, you don't want to go there, trust me.
So yeah I'm guessing if statements are faster than multiplication/division.

Right, but the speed of your programs isn't solely determined by the raw horsepower of your CPU: if statements can hurt branch prediction and cache locality if e.g. executed in the loop (ditto for a switch statement.) Cache eviction in particular can thrash your algorithms.

But normally, I'd say your comparison value will be in a register or something, so normally you're looking at about 3 instructions to do a comparison and jump - on the other hand if the compiler can do constant propagation, it might only be a single instruction to get your results (and may be replaced with a bitshift or something,) but there are are a huge amount of factors that can have an effect on the speed of any particular sequence of operations. This is partly why many people have studied things like 'cache-friendly algorithms,' and why some architectures - like the Cell Broadband Engine - do not have caches at all, only a local memory store with super fast access.

It will certainly work, and probably work pretty well. There are tons of libraries out there to speed up your development - you may want to look into the "Lightweight Java Game Library": http://lwjgl.org/documentation.php

At 4/24/09 09:03 AM, Alphabit wrote: Java is pretty much the godfather of modern OO and you'll find that most new OOP languages are based on it to some degree.

That's pretty debatable. Smalltalk is really the object-oriented language; as Alan Kay said, "when I invented object-oriented programming, C++ was not what I had in mind."

Either way, given what you define as OOP (that is, what Java has) I'm not sure you can really have a fully blow object-oriented language these days without taking some of those ideas, so, it also might not be fair to say that "new OOP languages are based on Java" as much as "new OOP languages have many features/similarities to java."

Why do people write C applications in VI / VIM on Unix? Are people purposely trying to keep progress back? What is the deal here?

As someone who uses vim to write C code on Linux, I find the assertion that I am 'trying to keep back progress' both ridiculous and inflammatory.

(I'm not saying C is the most uber-language there is - and don't even get me fucking started on editors - I just find this claim stupid.)

At 3/21/09 12:42 AM, ParadoxVoid wrote: I'm trying to load up a pre-written Haskell script written by my lecturer for an Assessment task - I've fixed up the parts that he wanted us to do, and when I load my edited script, it loads into GHCi fine.

However, when I try to run anything, it loads up all the required modules which are "Graphics.Plot" modules, but ultimately fails to load when it tries to load up GLUT-2.1.1.2. The error is:

Loading Package GLUT-2.1.1.2 ... can't load .so/.DLL for: glut32 (addDLL: unknown error)

Any ideas? I'm running on Windows Vista, if that's an issue.

It would probably be more appropriate to direct this question to the GHC developers themselves:

http://www.haskell.org/mailman/listinfo/
glasgow-haskell-users

Due to the... 'obtuse' error message, I think it could be a bug in the compiler. Ask the developers on the mailing list; it may need to be reported as a GHC bug (I actually fixed a bug in GHC today!) If you have further questions I might be able to help you, but you might want to post details about your GHC version. etc. as well as the program in question.

Personally, I just point people to python. It's intuitive, simple and you can get shit working quick. Beginners need motivation. Being able to build real apps fast gives you a sense of accomplishment, which is important. You can move when (if) you want to, but for a complete novice, first impressions are important. :)

At 3/9/09 09:59 AM, CronoMan wrote: Since java operates on bytecode, the code needs to be compiled in order to execute, there is no eval statement. Only scripting languages can do eval, because they are interpreted and not compiled.

A compiled language can very feasibly have an eval procedure - you just need compiler support at runtime (in the form of both a compiler and a runtime linker; these are normally both implemented as part of the langauge runtime itself.)

Many compiled languages just don't support this though, because that's a shit-ton of work and it's really hard to get right depending on the language (throw static typing in and it's a hell of a lot harder.) Scheme implementations in particular though are noticeable for both being able to compile code and have an eval procedure (since it's part of the spec, iirc.)

At 3/6/09 09:00 PM, iMate900 wrote: mov [es:edi], [eax-4]

What is wrong with this line?

This is Assembler using NASM as my assembler, not GAS.

There is no operand for x86 processors to mov from a memory location to a memory location; you can for example move from register to register and register to memory and vice/versa, but not memory to memory.

So the basic fix is to use two instructions - one to dereference eax-4, and one to move that value into edi.

At 3/7/09 01:57 AM, 066pop wrote: Hmm... I don't actually know anything about OSes

Then what the fuck are you in here for?! >:/

I don't know any but, I do know the experts will be avoiding NG...

Jesus christ, this is what you think?

People like you are the reason there's shit all over the wall here, because you're shitting on it, and not trying to help anything.

Arch Linux has been the distro I've used since I started using linux and I'm very much happy with it.

The package repositories may be smaller than e.g. apt-get (debian/ubuntu) or portage (gentoo,) but it's simple binary package installation so it's fast, and after the installer you're simply dropped into a shell where you can install a great deal of things and fit the system how you like it.

Like most distributions, there's great support for development tools and whatnot.

I recommend Arch Linux + xmonad as that's what I've been using for the past year or so. A very good minimalist setup that's simple and to the point. You can customize xmonad to whatever your liking (if you decide to go this route, you can steal my configuration file and I'll help you - the WM is written and extended in haskell.)

When you make a quick project, Java is great.

A blanketing statement that has no tangible support. Java isn't great for a lot of things; the language is verbose and rigid, and forced OOP isn't really great either. If I want access to all of Java's libraries and the ability to write code in a sound language, I would just choose Clojure.

Most people here on newgrounds can only do some really basic stuff in C++ involving variables, but they don't know how to make pointers and classes in it.

The amount of evidence you have for this claim is staggering, and your skill as a programmer should not be denied.

Shall I tremble with weak-knees at your ability to use things like, gasp, templates or multiple inheritance?

Yes. Yes I shall.

At 2/4/09 02:58 PM, DFox wrote: You're going to need to actually prove the use more Python than PHP because from the looks of things the site is driven by a LOT of PHP.

http://mail.python.org/pipermail/python-
dev/2006-December/070323.html

At 2/3/09 10:13 PM, littleMonsterGames wrote: Let me reiterate my post. All I was trying to say is that any language that is popular is probably worth learning for three reasons (maybe there's more, but w/e). One, it's probably popular because it's good, so you can expect a good language.

This one is wrong. No language or tool needs to actually be 'good' to take off - it just needs to scratch an itch; or, really, suck less in some other way. Java and especially PHP are not really 'good' languages by any means, but they're popular because they fill a need. Hopefully I never have to write PHP again for as long as I live, until maybe everybody running its show is killed and replaced by people who don't live to ruin things or something.

Being industrial/enterprise-y is not a good metric of whether something is 'good', in most cases you land up with something non-flexible, verbose and annoying (java is, again, an example.) Seriously, don't trust the status quo so much; it's more interesting and rewarding to challenge it anyway.

other stuff

I otherwise have no comments with this part of your post.

I don't particularly like Java as a language, but the JVM is IMO excellent; from a technical standpoint, it's easily the most advanced, optimizing, industrial-strength virtual machine there is. There are a lot of other languages out there for the JVM that are not java, and this gives you all the benefits of things like java's libraries, but without having to use it. :)

At 2/3/09 07:17 AM, littleMonsterGames wrote: Java isn't obsolete, but in my opinion, it's simply because it's so popular. I think c#/c++ are all anybody really needs. Of course, it could be argued that Java is popular because it isn't obsolete, but I honestly think it's the other way around.

I'm not really sure why you lump C# and C++ together, but the only reason people really use C++ today - outside of very niche domains - is probably the exact same reason people continue to use Java today: they have to support legacy code, because a rewrite of an old is not cost-effective economically. They don't use it because C++ is "all the world will ever need." All you ever really *need* to survive is really just a little water and bread, but that doesn't sound awesome either. :/

I can almost guarantee you for about 90% of the code you're going to want to write - unless it is a very niche domain like I said (think device drivers, real-time audio/graphics processing, number crunching, etc) - C++ realistically offers less benefits and will waste more of your time than a large swathe of other languages like Python (or C# or something.)

Want proof? About a year or so ago I was working on a trivial game server. This server could do a number of things:
* It could utilize multicore processors and the code was distributed amongst computers in a cluster, meaning it could scale up to hundreds of thousands/millions of concurrent connections
* There was an administration interface, so you could add nodes and remove nodes from the cluster while the system was running
* You could update all the code inside the server while it's running (hot-code loading.)
* The entry node had a load balencer and would balance between nodes that are going on and off
* Basic fault tolerance so in the case of failure, the other nodes take over and let it crash and recover gracefully.

I wrote this in not C++, and it only took 300 lines of code and 2 days to accomplish all of the above. What realistic benefit would C++ have offered me here? None.

Sorry for the rant, but I also haven't posted in a while it seems like. :)