Re: Your opinion on the Agner fog manuals.


o///annabee wrote:
På Sat, 09 Dec 2006 22:28:48 +0100, skrev <rhyde@xxxxxxxxxx>:

To date, you've offered no
evidence suggesting that Agner's material is incorrect other than state
that you have a difference of opinion as to whether large applications
should be written entirely in assembly.

No diffrence of opinion. Diffrence of facts.

Obviously, you confuse opinions with facts.


In the greater marketplace, I'm
afraid Agner's stance hold more credence than your's.

:)) Dont make me laugh.

Laugh all you want, that's your problem. If you don't recognize the
fact that most people feel that programming whole applications in
assembly language is foolish, then you are either being intellectually
dishonest or are too naive to even be having this discussion.


That doesn't suggest that he hasn't done so. Indeed, one of the reasons
he may be making this suggestion is exactly because he *has* done so
and understands the pain.

If its a pain to him writing assembly, why does he think he can go on
teaching it to others? Which is my point exactly.

You can be good at something, and be knowledgable about that topic, and
still prefer not to do it. What's your point?


I guess he takes his own advice for this one at least. But by all
means, I
could be wrong. He may have tried to make a program in assembly, but
unable to pull it off, he concluded that noone would be able to pull it
off. Maybe thats what he means?

I doubt it. He's probably making the common suggestion that assembly

He makes *Assumtions*.

How do you know? Have you asked him? How do you know he's not stating
this from experience? Again, you continue the personal attacks without
knowing the first thing about what you're talking about. You judge the
quality of an entire document that advanced programmers have respected
for years on the basis of a difference of opinion between you and the
author. You *assume* that the book is crap because the author states
that writing whole applications in assembly is foolish. It seems to me
that *you're* the one making assumptions here.


language should only be used for those components of an application
that require the power of assembly and other components of the app
should be written in a language suitable for those sections. No single
tool is appropriate for every job. That includes assembly as well as
any particular HLL.

Every program is build from building blocks which are assembly.

Not true.

Be
definition, your statement is logically absurd.

Whatever you say... You're just trolling again.



Perhaps the fact that those "advanced assembly language programmers"
desire the information he has to offer?

That one didnt even care to download it even. :)

Perhaps you should try citing some advanced programmers who've refused
to download the information? And while Rene may be more "advanced"
than you, he hardly qualifies in this department as he has steadfastly
refused to learn about optimization in assembly language.


answer: His level of intellectual honesty

Has nothing to do with it.

His (Agnars) lack of it then.

Agner's lack of intellectual dishonest has nothing to do with it.
You're just trolling again.

How did you come to that conclusion?

"Never make the whole program in assembly. " ?

How many times do you have to be told that you've come to the wrong
conclusion based on that statement? I realize you got your pedigree
from the "Rene Tournois school of debate", but making the same
incorrect statements over and over again won't win you any arguments.
You're just trolling again.



He is attacking the beginners.

???
By claiming that his book is for advanced programmers?
My, you're a bit paranoid, aren't you?

Paranoid? :)) How did you come to that conclusion?

Obviously, you feel that this paper is an attack on you.

When he offers a paper
no asm programmers would dirty their mind reading,

The facts prove otherwise. Agner's work is one of the most-cited tomes
on optimization. It didn't get that way by having no assembly language
programmers read it.

but claiming it is for
"advance programmers" then naturally beginners would tend to be
interessted in reading it,

It sucked you in, I suppose...

in the hope to learn something.

So now we get down to the real crux of the matter -- you're upset
because *you* haven't learned anything in two chapters of reading. Try
reading the whole thing first. If it still doesn't make sense to you,
you're probably one of those beginners that Agner's talking about who
needs a little more experience and reading to get to the point the
stuff makes sense. Hint: try reading some computer architecture
material too. That will really help you understand the nature of the
discussion in his documentation.

What they learn
is that "you should never".

Boy, you're really stuck on that *one sentence* in the documentation,
aren't you?

This is attacking their desire for knowledge
at its very root.

Not really. As the book is intended for advanced programmers, it's safe
to assume that the appropriate audience for the book has made up their
mind one way or the other long before reading Agner's documentation and
all he is doing is pandering to the vast majority of the people who
pick it up.


And you know what, that seem to me nothing but the
purest of evil.

So don't read it. And don't benefit from the "evil knowledge" contained
therein. You'll be the only one who suffers as a result.


How do you know? Do you know the man personally? Have you read his
assembly code? Have you even read beyond the introduction in his
documentation?

Yes I have, and may cover it later.

Yet you offer no evidence, other than he claims that whole apps
shouldn't be written in assembly, that his information is faulty?
Surely if the document is as bad as you claim it to be you can come up
with something more substantial than that.


Again, you're being a troll, nothing more.
You're attacking someone because you don't like the fact that they have
received a lot of attention for their hard work they've put into
documenting a subject like optimization; you wish you could get that
kind of attention for yourself. So, as a typical Rene underling, you
attack the person rather than doing something equally impressive
yourself. An old story with you and Rene.

Well, maybe thats what it looks like. I agree it could appear that way.
However, i think its my due to point out such things, when I see it. The
"you should never" part is particularly insane, considering the facts.

What "facts" are those? You do realize that about 99.9% of the
programming population out there fully agrees with Agner on his opinion
of language choice, don't you? Are you trying to claim that 99.9% of
the programming population is insane? That reflects poorly on your own
sanity.

If
he cared for those facts, he would since long (maybe 8 years at least)
have updated his documents. But I guess he just doesnt care for that....
and then he deserve a reminder.

If he wrote that statement 8 years ago, then it's probably even more
true today. Machines have gotten faster, storage has dramatically
increased, and HLLs have progressed in that time. All the more reason
that there is less need to write applications in assembly language
today.




Beginners would. And advanced asm programmers would not.

But advanced programmers do, and they approve of what Agner Fog has
written. So your premise is completely wrong.

Which one is that? So far only one advanced asm programmer posted in this
thread, and he refuses even to download it.

Well, if you are holding up Rene as an example of an "advanced assembly
language programmer" for whom Agner's documentation was written for, I
can understand why you think it's not that good. I've seen Rene's code.
He has a bit more to learn about assembly programming before he's ready
to tackle that kind of stuff.



As proven by the
very fact, that Betov didnt even want to look at it,

Earth to Wannabee -- Rene is not an "advanced" assembly language
programmer.

And who are you to decide?

Someone who has actually looked at Rene's code.

Did you ever write an assembler with an IDE

Yes, in the late 1970s and early 1980s as a matter of fact.

and
100s of tools, including a dissassmbler,

Yep.

all in an easy to use single
package. In full assembly language? No.

Wrong. Try searching for information about the LISA assembler sometime.
And, btw, that's when I learned that command-line tools are far more
acceptable to real programmers. That's when I learned that a strong
set of features is *far* more important to programmers than the speed
of the tool (I had the fastest tool around, by a couple of orders of
magnitude, and lost out to tools that were more feature-laden, that
supported things like static linking).



But Betov did. You wrote a text
converter in C, bison and god knows what.
Am I right?

No, you are not right. Indeed, the RosAsm system that Rene wrote is
amazingly similar to the LISA system I wrote over 25 years ago. All the
same problems -- no need for static linking because it's so fast, no
interoperability with other languages, big monolithic application
(well, big at the time), proprietary source format, etc., etc.. That's
why I laugh at RosAsm today; I made those same mistakes 25 years ago.


(its a rethorical question btw) :)

Only because of your ignorance you think so.


He has hacked out a lot of poorly designed code using the
"basic 25 instructions" out of the instruction set that he has bothered
to master. He is not the target audience for documentation like Agner
has written because he refuses to extend his x86 assembly language
programming knowledge. Indeed, he recommends that people *not* study
instruction-level optimization, a bizarre attitude for someone trying
to push assembly language programming.

You fail to understand what all masters seem to agree on, the importance
of simplifying.

Yeah, but note what Einstein said -- make things as simple as possible,
but no simpler. I'm afraid Rene has not followed that advice.


You also seem to be of the misinformed view that Betov would reject a
programming technich without first having examined it. Thats what
"advanced" tend to mean.

Rene rejects lots of things out of ignorance. Nothing new there.





and that the nuubs in
the asmcommunity, praising it a lot.

Why don't you post your complaints over in comp.arch and see what the
"nuubs" over there think of your comments?

I think its hard enough with you around. Two nuubs and I would have to
hire on a secretary.

Coward.


After all, the guy doesnt write his main applications in assembly.

So? What does that have to do with anything? As this is more of an
achitectural document rather than a programming tutorial, you could
conceivably write it without ever having written a single line of
assembly code, though that is clearly not the case for Agner Fog.
After all, this documentation is more about the hardware behavior than
anything else. But rest assured, Agner has written a *lot* of assembly
code to produce the timing information present in his documentation.

"Oh dear". :) So by timing his small snipplets, he taught us something?

Yes.

When you, in the actuall case of making some code run better, after been
thinking hard about the strategic optimizations, no longer are able to
think of a better strategy, sometimes you may try to turn to
cyclecounting. This is just the sign that you given up on the most
important part of optimization.

Now you're just parroting Rene's line about optimization. Don't you
have any original ideas of your own? Try reading Agner's book. You must
might learn something. But if you go into it with eyes blinded by
prejuidice you're just wasting your time.




(If he
even wrote anything in C) So how could he know enough about it to
dicourage others from trying?

How do you know he hasn't?

Hasnt what?

What you were claiming he hadn't done -- written assembly code. I
realize that all you're doing now is trolling to get attention (and
this will be my last follow-up, no need to further feed the troll), but
you should at least be aware of the argument you're trying to make.
Otherwise you wind up like Rene and start claiming exactly the opposite
of what you started out claiming when you started the thread.


Of course, you haven't even *read* it yet. So you're really not in a
position to judge the quality of the information therein, are you?

But I dont. I comment only on the parts I did read. I explisitly pointed
out that fact. Allthough it surprices me not that much that you couldnt
see it.

So you take things out of context and assume that you're stating
absolute truths?


The first two chapters stink a lot. The stink has the unmistakeable tint
of nuub all over it.

You keep saying all these things without providing a shred of evidence
to back up your statements. It's easy to say "the first two chapters
stink a lot." Now try telling us exactly what stinks, why it stinks,
and how it could be made to smell better. You're just using a lot of
negative words without backing up anything you're saying. IOW, you're
trolling to get attention.

He is saying "You should never", for all the wrong reasons. I have done
it, encouraged of the demonstration provided by RosAsm itself. And so I
have actually verified it. Therefore, what I say is verifiable correct,
and what Agnar says is pure bollocks.

Still stuck on that one sentence, eh?
So the first two chapters stink and the book contains no useful
information because Agner Fog repeats the well-established opinion that
whole applications must not be written in assembly language? My, how
your closed mind eliminates much of the world's knowledge of computer
science, then, because *most* books on programming subjects (that touch
on assembly language) make this claim.




And this also imply that the rest of it could be very bad as well. But I
havent read all of it yet, so it would be premature to make any claims
about the rest.

Amen.

So why didnt you go back and change your statement above, once you saw
this reservation? Maybe it allready slipped your mind?

I guess the sarcasm escaped you.



Looking forward to it :-)
The more information on optimization, the better.

The more live demonstrations that you _should_ take the path of writing
full PEs in assembly, the better.

Not my job. First of all, I don't equate an object code format
(Portable Executable, in this case) with assembly language programming.
Second of all, I don't particularly believe that everyone should learn
assembly so they can write all their apps in assembly -- I happen to
believe that most people should learn assembly so they can write better
programs in the high-level languages that they're actually going to
use.



When you actually get around to reading the optimization documentation,
you'll discover that much of Agner's information was determined through
experimentation, not by "quoting the Intel/AMD manuals." Agner's
documentation wouldn't be all *that* useful if all he did was quote the
existing manuals, now, would it?

Well. Again. I havent "got around to that" (sic).

So rather than shooting off your mouth so much, why don't you read the
rest of the document first and see what it's really all about? And do
yourself a favor, get past the "don't write apps in assembly"
statement.



Obviously, you haven't looked deeply enough. I'd advise you read
through the whole set before you stick your foot in your mouth again.

Its not needed to read the whole book, to comment on the first pages.

Then the title of your thread should have been:
"Re: Your opinion on the first few pages of the Agner fog manuals."

In
fact, if the first pages spell disaster, then, if not a 100% certainty,
the rest of it could very well also be a disaster.

Ever heard the line "you can't judge a book by its cover"?
Cheers,
Randy Hyde

.

Relevant Pages

  • Re: Teaching (and Learning) Assembly Language, Part 2
    ... individual's needs when learning assembly language. ... much in the way of software development tools, ... limitations of a "source editor" provided with an assembler that forces ... the "assembly programmers" could get up to speed faster ...
    (alt.lang.asm)
  • Re: the quality of assembly language code
    ... Intel's compiler is supposed to ... I have found that I prefer 'C' and assembler roughly in inverse ... >> code than human beings writing raw assembly language. ... > assembly programmers and compilers. ...
    (comp.lang.asm.x86)
  • Re: Faster HexToBuffer Routines
    ... assembly language and represents one of the main reasons for using ... programmers want doesn't mean everyone else is. ... they do use libraries when they are available to them. ... because your assembler has a broken design that doesn't allow the use ...
    (alt.lang.asm)
  • Re: Why Not Assembly?
    ... > A HLL shields you from the low level stuff of ... > to know to use assembler. ... HLLs don't shield you from this. ... Certainly one advantage of HLLs is that companies and programmers are ...
    (alt.lang.asm)
  • Library Design, the script kiddies nightmare.
    ... good across the board of assembler programming. ... With languages that have the capacity, libraries of reusable code make ... comes at a cost that some of the older assembler programmers will not ... The virtue of writing code in this modular format is that you can use ...
    (alt.lang.asm)
Loading