Re: Fast Bit Operations Contest (unmoderated)

På Mon, 08 Jan 2007 05:29:44 +0100, skrev rhyde@xxxxxxxxxx <rhyde@xxxxxxxxxx>:

//\\o//\\annabee <Free" wrote:

> The use of assembly code for optimizing software is relevant only for
> highly CPU-intensive programs such as sound and image processing,
> encryption, sorting, data compression and complicated mathematical
> calculations.

So many good points, but failing at the only one that matters. All of the
99% comes for free in asm. And by then you be able to deal with the rest
as well.

Actually, it does *not* come for free.
First of all, few people know assembly language. That means in order
for them to use assembly, they must first learn it.

same for any language, and easier to get started with asm. 3 weeks, would be enough time to learn the basics for which to have a small starting point.

And as you've
demonstrated, it takes a *long* time to get good at it (you've been at
it, what, two or three years now?

3 years. Yes. But I wrote applications during the first month.

And your code is still at the
beginning level according to your own admission).

Maybe, but my apps run fast even so. Which is one of my main arguments. Its not needed to be an expert in order to write apps with RosAsm. RosAsm takes care of the heavy stuff, for you.

Even if you're
already the expert assembly language programmer,

Seriosuly, I am not.

it's rather
ego-centric to think that all the time spent on an application is going
to be spent by *you*. Other people will have to deal with your source
code sooner or later. In a commercial environment, turnover guarantees
this; in an open-source environment, the openess of the source
generally causes this. Maybe you're the world's best assembly
programmer and you crank out assembly code at twice the speed a good
HLL programmer can do his job. But then the next person comes along and
get stuck by your horrible coding style and choice of a language that
they don't know.

Yes, true, (coding style). This is why I try to clear it up as I go. Did a total renovation this summer for instance. Uploading new stuff in may 2007.

You need to think outside your little box sometime.

Thats how I was able to dump 10 years of pascal code, in favor of RosAsm. Sometimes one must admit failure, and start over. If thats not "outside" the box, then what is? i know I repeat myself, but so are you. We been over this now a 100 times.

This has been the *real* failure of assembly language, not the
inability of any particular individual to learn or use it.

What i see is this : HLLs gives a powerful demonstration, which is very hard to ignore. Very hard to argue against! But you of all people, who are the author of "Art of assembly" should see clearly that asm gives a programmer glasses that are clearer then the one HLLs use. If you cannot admit that asm is the better choise, then why the hell are you haunting in theese parts of the net at all. Thats the question we should be asking. And when answering, Betovs way of treating you no longer seems so strange, huh ???

Randy Hyde


Relevant Pages

  • Re: Unions in Assembly Language
    ... This mean less experience cranking out asm code. ... beginner asm programmer can most times do better than an experienced ... but a beginner asm programmer. ... that experience is the only facts, arguments can go back and forth. ...
  • Re: Disobeying jet engines - why?
    ... if the programmer is not up to the task. ... I know this, some programs I wrote for several processors, in asm. ... In such a case libc, and all the other libs, are a must. ... protocols have different libraries. ...
  • Re: Ideal computer language from scratch?
    ... The thing is totally diffrent with asm. ... In asm, even a beginner will ... And certainly in terms of the usage of machine instructions, ... The average RotAsm programmer, using Rene's suggestions, ...
  • Re: Rene cant handle AoAs Success
    ... An fully educated ASM-coder must be able to program any HW/CPU without ... the ballast of whatsover existiong Libraries around. ... A "fully educated" ASM programmer is dead. ... Pick up a copy of a book like "The Pragmattic Programmer". ...
  • Re: purely relative addressing
    ... > jmp 0xAddressOfCall ... Buffer overflow vulnerability is based on pure programmer stupidity, i.e., ...