[OT] Gibson's vision of computer languajes

[OT] Gibson's vision of computer languajes

[Jano]

I'm pretty sure many of us know Steven Gibson. Today I've tested a 
freeware from him. In the instructions page he made these statements:

"All of my programs are so small and so fast because I write only in the 
one, true, computer language: Assembly Language. All other languages 
reduce to assembly language, but they lose a lot in the translation. 
(Judge for yourself.)"

"People who can't program in assembly language (just about everyone) get 
really annoyed with me when I talk about how much better it is than their 
pet language. But it's not my fault if they just say they care about 
quality."

My blood is boiling. I want only to share to cool me a little down.

-- 
----------------------
Jano
402450@cepsz.unizar.es
----------------------

Re: [OT] Gibson's vision of computer languajes

[Dave Poirier]

Jano wrote:
> I'm pretty sure many of us know Steven Gibson. Today I've tested a 
> freeware from him. In the instructions page he made these statements:
> 
> "All of my programs are so small and so fast because I write only in the 
> one, true, computer language: Assembly Language. All other languages 
> reduce to assembly language, but they lose a lot in the translation. 
> (Judge for yourself.)"
> 
> "People who can't program in assembly language (just about everyone) get 
> really annoyed with me when I talk about how much better it is than their 
> pet language. But it's not my fault if they just say they care about 
> quality."
> 
> My blood is boiling. I want only to share to cool me a little down.

Well, I'm an assembly freak, and I must say that I pretty much agree 
with him <g>.  While assembly does create software that are small and 
fast, and can be made as reliable as any other programming language 
(using proper development techniques), portability can be applied at 
best by rewriting the entire thing using a generic algorithm.

Assembly is my language of choice when portability is not involved, but 
when you need to make it run on more than once Cpu, it's a no brainer.

EKS - Dave Poirier

Re: [OT] Gibson's vision of computer languages

[Larry Kilgallen]

In article <MPG.16ee10f0509986e9896f2@news.cis.dfn.de>, Jano <402450@cepsz.unizar.es> writes:

> "All of my programs are so small and so fast because I write only in the 
> one, true, computer language: Assembly Language. All other languages 
> reduce to assembly language, but they lose a lot in the translation. 
> (Judge for yourself.)"

Obviously he is a programming wimp, depending on a crutch like
assembly language.  If he were a _real_ programmer he would be
using machine language - raw 1's and 0's like nature intended.

And of course those programs should be entered in the only real
authentic method, toggling them in with switches on the front
panel.

Re: [OT] Gibson's vision of computer languajes

[Darren New]

Jano wrote:
> 
> I'm pretty sure many of us know Steven Gibson. Today I've tested a
> freeware from him. In the instructions page he made these statements:
> 
> "All of my programs are so small and so fast because I write only in the
> one, true, computer language: Assembly Language. All other languages
> reduce to assembly language, but they lose a lot in the translation.
> (Judge for yourself.)"
> 
> "People who can't program in assembly language (just about everyone) get
> really annoyed with me when I talk about how much better it is than their
> pet language. But it's not my fault if they just say they care about
> quality."

Well, I think knowing how to program in at least one assembly language
is important, since you know how the machine works. However, there's a
difference between knowing how to program in assembler and programming
in assembler. Indeed, some of the most recent advances in computers
(RISC, etc) are predicated on the idea that you won't write in
assembler, and indeed it is less efficient to do so.

And no, not all other languages reduce to assembler. All other languages
reduce to machine code. Whassamatta? Can't toggle in the binary like a
*real* programmer?

-- 
Darren New 
San Diego, CA, USA (PST). Cryptokeys on demand.
  To the user, everything works just as expected,
    assuming the user's expectations are correct.

Re: [OT] Gibson's vision of computer languajes

[Al Mole]

"Jano" <402450@cepsz.unizar.es> wrote in message
news:MPG.16ee10f0509986e9896f2@news.cis.dfn.de...
> I'm pretty sure many of us know Steven Gibson. Today I've tested a
> freeware from him. In the instructions page he made these statements:
>
> "All of my programs are so small and so fast because I write only in the
> one, true, computer language: Assembly Language. All other languages
> reduce to assembly language, but they lose a lot in the translation.
> (Judge for yourself.)"
>
> "People who can't program in assembly language (just about everyone) get
> really annoyed with me when I talk about how much better it is than their
> pet language. But it's not my fault if they just say they care about
> quality."
>
> My blood is boiling. I want only to share to cool me a little down.

It's a long time since I did any but yes, I 'll go along with Steve Gibson.
It's not practical for most situations but a programmer that can't do it is
a programmer not worth having on your team.

A friend had a job recoding Ada modules into assembler for EFA quite a few
years ago. Shows the value of Ada as an HDL as well as the shortcomings of
any high level language when it comes to performance.

Re: [OT] Gibson's vision of computer languajes

[tmoran]

> "All of my programs are so small and so fast because I write only in the
> one, true, computer language: Assembly Language.
  In this day of massive APIs and no access to the hardware, I wonder
how much of his programs' time is spent in his careful asm code, vs
in the OS somewhere.  OTOH, maybe he's onto something, marketing-wise.
---------------
*real* programmers know the cycle time of each instruction.

Re: [OT] Gibson's vision of computer languages

[Richard Riehle]

Larry Kilgallen wrote:

> In article <MPG.16ee10f0509986e9896f2@news.cis.dfn.de>, Jano <402450@cepsz.unizar.es> writes:
>
> > "All of my programs are so small and so fast because I write only in the
> > one, true, computer language: Assembly Language. All other languages
> > reduce to assembly language, but they lose a lot in the translation.
> > (Judge for yourself.)"
>
> Obviously he is a programming wimp, depending on a crutch like
> assembly language.  If he were a _real_ programmer he would be
> using machine language - raw 1's and 0's like nature intended.
>
> And of course those programs should be entered in the only real
> authentic method, toggling them in with switches on the front
> panel.

Brings back fond memories of the good old days when we could
enter entire programs through the front panel switches.

Richard Riehle

Re: [OT] Gibson's vision of computer languajes

[Adrian Hoe]

Jano wrote:
> 
> I'm pretty sure many of us know Steven Gibson. Today I've tested a
> freeware from him. In the instructions page he made these statements:
> 
> "All of my programs are so small and so fast because I write only in the
> one, true, computer language: Assembly Language. All other languages
> reduce to assembly language, but they lose a lot in the translation.
> (Judge for yourself.)"


Well, he says his programs are small, doesn't he? I agree that assembly
language becomes handy for certain programming purpose that requires
speed and no other programming languages on earth can beat assembly.
Assembly becomes a huge monster if the program is huge. Ask Steven to
write an ERP (Enterprise Resource Planning) software, which contains
more than 250K function points. The task will be beyond Steven's
ability, but doesn't mean that is impossible.

The speed of execution may become irrelevent, if not today, at least 15
years later when 30GHz microprocessors (as predicted by Intel's Craig
Barrett) will be available.

An ERP software written in Ada can parse 50 complex business rules and
extract information (based on 50 business rules) from 6TB database in
just 3 minutes. Running on a 1.5GHz P4 workstation and a Sun Blade 1000
(dual CPUs) DB server.

At least 3 minutes are satisfactory for our clients. :)
-- 
                                       -- Adrian Hoe
                                       -- http://adrianhoe.com

Re: [OT] Gibson's vision of computer languajes

[Chad R. Meiners]

No reason to get your blood boiling.  Remember high level languages have can
have some speed trade-offs; this is risk you take when you allow the machine
to do the compiling.  Of course with modern pipelined processors I would
rather the compiler do all the fancy tricks to tweak out all the speed
instead of me manually scheduling instructions to avoid stalls.

-CRM

"Jano" <402450@cepsz.unizar.es> wrote in message
news:MPG.16ee10f0509986e9896f2@news.cis.dfn.de...
> I'm pretty sure many of us know Steven Gibson. Today I've tested a
> freeware from him. In the instructions page he made these statements:
>
> "All of my programs are so small and so fast because I write only in the
> one, true, computer language: Assembly Language. All other languages
> reduce to assembly language, but they lose a lot in the translation.
> (Judge for yourself.)"
>
> "People who can't program in assembly language (just about everyone) get
> really annoyed with me when I talk about how much better it is than their
> pet language. But it's not my fault if they just say they care about
> quality."
>
> My blood is boiling. I want only to share to cool me a little down.
>
> --
> ----------------------
> Jano
> 402450@cepsz.unizar.es
> ----------------------

Re: [OT] Gibson's vision of computer languajes

[Thomas Koenig]

 <tmoran@acm.org> wrote:

>---------------
>*real* programmers know the cycle time of each instruction.

If that helps you any, in the days of RISC, pipeling, caches,
etc... and even then, that may only work for a given combination
of processor, intermediate cache, main and virtual memory.

Very often, it's cache misses where cycles are burned these days,
not individual machine instructions.

Re: [OT] Gibson's vision of computer languajes

[Preben Randhol]

On Mon, 4 Mar 2002 23:17:42 +0100, Jano wrote:
> "All of my programs are so small and so fast because I write only in the 

More correctly he should simply have stated:

"All of my programs are so small so I can write them in assembly"

-- 
Preben Randhol         �For me, Ada95 puts back the joy in programming.�

Re: [OT] Gibson's vision of computer languajes

[Pascal Obry]

Dave Poirier <instinc@users.sf.net> writes:

> Well, I'm an assembly freak, and I must say that I pretty much agree with him
> <g>.  While assembly does create software that are small and fast, and can be
> made as reliable as any other programming language (using proper development
> techniques), 

Of course "can be", I just can't resist... But you are certainly a far
superior guy to achieve that, from which planet ?. BTW, "as reliable" for which
size of projects (here on earth we have many programs with more that 100k lines
of high level language), and for which cost ?

Just a side note, I've done also lots of assembly, just that in my field
(Information Systems) I do not see assembly as a possible choice :)

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|         http://perso.wanadoo.fr/pascal.obry
--|
--| "The best way to travel is by means of imagination"

Re: [OT] Gibson's vision of computer languajes

[Warren W. Gay VE3WWG]

Dave Poirier wrote:

> Jano wrote:
>> I'm pretty sure many of us know Steven Gibson. Today I've tested a 
>> freeware from him. In the instructions page he made these statements:
>>
>> "All of my programs are so small and so fast because I write only in 
>> the one, true, computer language: Assembly Language. All other 
>> languages reduce to assembly language, but they lose a lot in the 
>> translation. (Judge for yourself.)"
>>
>> "People who can't program in assembly language (just about everyone) 
>> get really annoyed with me when I talk about how much better it is 
>> than their pet language. But it's not my fault if they just say they 
>> care about quality."
>>
>> My blood is boiling. I want only to share to cool me a little down.
> 
> Well, I'm an assembly freak, and I must say that I pretty much agree 
> with him <g>.  While assembly does create software that are small and 
> fast, and can be made as reliable as any other programming language 
> (using proper development techniques), portability can be applied at 
> best by rewriting the entire thing using a generic algorithm.
...
> EKS - Dave Poirier

There is nothing wrong with "liking assembler", but you're foolish to
believe that assembler programs "can be made as reliable as any other
programming language".  Having used operating systems written in
assembly language (anyone remember Honeywell's GCOS8/DPS8?), you
would not want to go back there!  Using those systems I learned very
quickly to save my edit session every few minutes (if not seconds),
because it was not unusual for the system to crash between 1-5 times
a day. There were always new oodles of assembler patches being issued
to correct prior patches and on and on it went. MULTICS and UNIX were
a big advance in reliability because they went AWAY from assembly
language. Now it is a good time to move away from C to Ada for
operating systems (note that BSD and Linux are still written in
C, not C++).

The argument you make is the exact same argument that C/C++ programmers
make WRT Ada. They always state "using proper development techniques",
but the underlying problem is that this is subject to human error and
is not reliably done. It kinda reminds one of the
saying "the pathway to hell is paved with good intentions".

-- 
Warren W. Gay VE3WWG
http://home.cogeco.ca/~ve3wwg

Re: [OT] Gibson's vision of computer languages

[Warren W. Gay VE3WWG]

Larry Kilgallen wrote:

> In article <MPG.16ee10f0509986e9896f2@news.cis.dfn.de>, Jano <402450@cepsz.unizar.es> writes:
> 
>>"All of my programs are so small and so fast because I write only in the 
>>one, true, computer language: Assembly Language. All other languages 
>>reduce to assembly language, but they lose a lot in the translation. 
>>(Judge for yourself.)"
> 
> Obviously he is a programming wimp, depending on a crutch like
> assembly language.  If he were a _real_ programmer he would be
> using machine language - raw 1's and 0's like nature intended.
> 
> And of course those programs should be entered in the only real
> authentic method, toggling them in with switches on the front
> panel.


That's right, and using octal or hexadecimal is cheating!


-- 
Warren W. Gay VE3WWG
http://home.cogeco.ca/~ve3wwg

Re: [OT] Gibson's vision of computer languajes

[Dave Poirier]

Pascal Obry wrote:
> Dave Poirier <instinc@users.sf.net> writes:
> 
> 
>>Well, I'm an assembly freak, and I must say that I pretty much agree with him
>><g>.  While assembly does create software that are small and fast, and can be
>>made as reliable as any other programming language (using proper development
>>techniques), 
>>
> 
> Of course "can be", I just can't resist... But you are certainly a far
> superior guy to achieve that, from which planet ?. BTW, "as reliable" for which
> size of projects (here on earth we have many programs with more that 100k lines
> of high level language), and for which cost ?
> 
> Just a side note, I've done also lots of assembly, just that in my field
> (Information Systems) I do not see assembly as a possible choice :)
> 
> Pascal.

Oh, don't get me wrong, for any commerically viable product of such
major size, assembly should be avoided except where you require really
tight loop.  But if you are simply a computer passionate who create on
his free time, then assembly can certainly be used for any sized
project.

Software reliability in any programming language doesn't come by itself,
but by an entire development process which include this never ending
serie of tests on each individual components then on those components
once put together.

I have programmed assembly in projects well over 25,000 lines of
without getting lost or more confused than when the project was
only 100 lines.  The key element is proper documentation and commenting,
which unfortunately most hobbyist asm developers nowadays seems to so
easily forget. 100k lines is certainly possible, even 200k lines if one
have sufficient time.

Again, all this is not necessarily, read most unlikely, the best choice.
HLL does provides many benefits when it comes to commercially viable
products, it lower the costs of production and maintenance, and open the
doors to larger pool of programmers.

In my opinion, Assembly is the best suited language for any project, but
our society is unfortunately based on money, and to be commercially
viable assembly must of then be traded for less human-time consuming
alternatives.

EKS - Dave Poirier

Re: [OT] Gibson's vision of computer languajes

[Dave Poirier]

Warren W. Gay VE3WWG wrote:
> Dave Poirier wrote:
> 
>> Jano wrote:
>>
>>> I'm pretty sure many of us know Steven Gibson. Today I've tested a 
>>> freeware from him. In the instructions page he made these statements:
>>>
>>> "All of my programs are so small and so fast because I write only in 
>>> the one, true, computer language: Assembly Language. All other 
>>> languages reduce to assembly language, but they lose a lot in the 
>>> translation. (Judge for yourself.)"
>>>
>>> "People who can't program in assembly language (just about everyone) 
>>> get really annoyed with me when I talk about how much better it is 
>>> than their pet language. But it's not my fault if they just say they 
>>> care about quality."
>>>
>>> My blood is boiling. I want only to share to cool me a little down.
>>
>>
>> Well, I'm an assembly freak, and I must say that I pretty much agree 
>> with him <g>.  While assembly does create software that are small and 
>> fast, and can be made as reliable as any other programming language 
>> (using proper development techniques), portability can be applied at 
>> best by rewriting the entire thing using a generic algorithm.
> 
> ...
> 
>> EKS - Dave Poirier
> 
> 
> There is nothing wrong with "liking assembler", but you're foolish to
> believe that assembler programs "can be made as reliable as any other
> programming language".  Having used operating systems written in
> assembly language (anyone remember Honeywell's GCOS8/DPS8?), you
> would not want to go back there!  Using those systems I learned very
> quickly to save my edit session every few minutes (if not seconds),
> because it was not unusual for the system to crash between 1-5 times
> a day. There were always new oodles of assembler patches being issued
> to correct prior patches and on and on it went. MULTICS and UNIX were
> a big advance in reliability because they went AWAY from assembly
> language. Now it is a good time to move away from C to Ada for
> operating systems (note that BSD and Linux are still written in
> C, not C++).
> 
> The argument you make is the exact same argument that C/C++ programmers
> make WRT Ada. They always state "using proper development techniques",
> but the underlying problem is that this is subject to human error and
> is not reliably done. It kinda reminds one of the
> saying "the pathway to hell is paved with good intentions".

True, the reliability of an assembly built project is largely based on 
the skills of the human writing the code and testing it.  Since humans 
are due to fail, so is the software.  Why would I otherwise learn Ada95 
myself if it wasn't for this exact reason?

I do like to believe that programmers should be skilled professionals 
rather than script kiddies.  Note I'm not calling you or anybody else in 
this channel as such, but we do notice that most of the asm programmers 
still programming in asm are young uneducated people, at least those 
visible in the Open Source community.

Software development techniques have evolved a great deal since those 
times that you refer, some of those techniques and tools can be used not 
only in higher level language like Ada, but also by lower level 
languages like Assembly.

By combining skilled assembly programmers and those new development 
techniques, as well as an all time careful approach, I do believe it is 
possible to create tighter/smaller programs that are as reliable (read: 
not developed in the same amount of time nor with the same resources) as 
their higher level counterparts.

EKS - Dave Poirier

Re: [OT] Gibson's vision of computer languajes

[chris.danx]

> "People who can't program in assembly language (just about everyone) get
> really annoyed with me when I talk about how much better it is than their
> pet language. But it's not my fault if they just say they care about
> quality."
>
> My blood is boiling. I want only to share to cool me a little down.

In the past few weeks there have been a few threads where the OP never
contributed to the thread past the original post (roughly once every two or
three weeks).  Huge threads resulted from these, mainly because the post was
designed to provoke such an argument, with little accomplished because most
ppls minds are made up on the issue in question.  Seems it's all too easy to
do.


Appologies if this is a "genuine" poster, rather than a "troll" (is that the
correct term?)
Chris

Re: [OT] Gibson's vision of computer languajes

[Marin David Condic]

"Dave Poirier" <instinc@users.sf.net> wrote in message
news:3C8503D9.30209@users.sf.net...
>
> In my opinion, Assembly is the best suited language for any project, but
> our society is unfortunately based on money, and to be commercially
> viable assembly must of then be traded for less human-time consuming
> alternatives.
>

Why do you think assembly language is the best suited language for any
project? Most people would contend that assembly is occasionally necessary
to accomplish things that can't be done (easily) in a high level language,
but that it is to be avoided because of its developmental inefficiency, high
probability of injecting errors and difficulty of maintenance. I'm curious
as to what reasons you'd give to go against the conventional wisdom.

High level languages of almost any stripe generally mean you get the job
done faster/better/cheaper & that's a good thing. What advantage do you get
from assembler other than the occasional need to dip into it for some
specific task not easily done otherwise - or the pure asthetic pleasure of
doing something one might find interesting? (Liking a language is one thing.
I like Ada for various asthetic reasons. Giving sound engineering reasons
for using a particular language is another. Most of us here could give such
reasons for using Ada in a given domain. What's the argument for favoring
assembler?)

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/

Re: [OT] Gibson's vision of computer languajes

[Darren New]

"Warren W. Gay VE3WWG" wrote:
> programming language".  Having used operating systems written in
> assembly language (anyone remember Honeywell's GCOS8/DPS8?), you
> would not want to go back there!  Using those systems I learned very
> quickly to save my edit session every few minutes (if not seconds),
> because it was not unusual for the system to crash between 1-5 times
> a day.

On the other hand, CP-V, running on the same honeywell hardware, also in
assembler, ran years without crashing. Indeed, I *never* saw it crashing
except when privledged users were mapping kernel memory structures into
their own address space to do things they weren't supposed to and hit a
race condition with other privledged users doing something similar
(which happened maybe once a year). I even had long compiles survive
power failures without a hiccup, thanks to the wonders of core memory.
:-)  Oh, well, there was the one time the system programmer
single-stepped an interrupt-off instruction in the kernel. That was
pretty amusing.

But this is getting really, really OT. 

  Id? Logon please!
  A603: Load module does not exist.

-- 
Darren New 
San Diego, CA, USA (PST). Cryptokeys on demand.
  To the user, everything works just as expected,
    assuming the user's expectations are correct.

Re: [OT] Gibson's vision of computer languages

[Larry Kilgallen]

In article <3C84FF8A.3060607@home.com>, "Warren W. Gay VE3WWG" <ve3wwg@home.com> writes:
> Larry Kilgallen wrote:
> 
>> In article <MPG.16ee10f0509986e9896f2@news.cis.dfn.de>, Jano <402450@cepsz.unizar.es> writes:
>> 
>>>"All of my programs are so small and so fast because I write only in the 
>>>one, true, computer language: Assembly Language. All other languages 
>>>reduce to assembly language, but they lose a lot in the translation. 
>>>(Judge for yourself.)"
>> 
>> Obviously he is a programming wimp, depending on a crutch like
>> assembly language.  If he were a _real_ programmer he would be
>> using machine language - raw 1's and 0's like nature intended.
>> 
>> And of course those programs should be entered in the only real
>> authentic method, toggling them in with switches on the front
>> panel.
> 
> 
> That's right, and using octal or hexadecimal is cheating!

Well let's hold on there.  I would be willing to give full marks
to someone who can accurately manipulate a toggle switch that
has 8 different meaningful positions along the same length path
as a normal toggle switch.  And extra credit for 16 !

Re: [OT] Gibson's vision of computer languages

[Wes Groleau]

> And of course those programs should be entered in the only real
> authentic method, toggling them in with switches on the front
> panel.

Naah -- I punch them in on paper tape with an
ice pick and feed it through the reader.

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: [OT] Gibson's vision of computer languajes

[sk]

Hi,

>In the past few weeks there have been a few threads where the 
>OP never contributed to the thread past the original post 
>(roughly once every two or three weeks).  Huge threads resulted 
>from these, mainly because the post was designed to provoke such 
>an argument, with little accomplished because most ppls minds 
>are made up on the issue in question.  Seems it's all too 
>easy to do.

At least it makes no attempt to claim a relationship to Ada 
and is clearly marked OT for a change :-)

I think cla has done remarkably well so far this year. It 
is already march and the perenial

* Ada vs C ...
* Ada should do it my way ...
* ACT/Aonix etc should do it my way ...

wars havn't broken out yet (holding breath and crosssing 
fingers :-)

Luckily, the "Future with Ada" thread has not escalated 
and actually moved into identifying various "Rosen"
posters.

 o o
  |
 \_/

-- 
-------------------------------------
-- Merge vertically for real address
-------------------------------------
s n p @ t . o
 k i e k c c m
-------------------------------------

Re: [OT] Gibson's vision of computer languages

[Wes Groleau]

> That's right, and using octal or hexadecimal is cheating!

Unless your front panel has octal rotary switches.

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: [OT] Gibson's vision of computer languajes

[David Starner]

On Tue, 05 Mar 2002 12:43:53 -0500, Dave Poirier <instinc@users.sf.net> wrote:
> But if you are simply a computer passionate who create on
> his free time, then assembly can certainly be used for any sized
> project.

But that's true for any programming language and project. I'm sure most
of us have heard about the Fortran compiler written in Teco? 
 
> In my opinion, Assembly is the best suited language for any project, but
> our society is unfortunately based on money, and to be commercially
> viable assembly must of then be traded for less human-time consuming
> alternatives.

I don't think we could have our networked society if everyone used
assembly. There's just not enough programmers who are good enough to
write the programs people need in the two decades since the PC
revolution.

"Best suited language" implies that running speed is the goal. It's
frequently not, and sometimes where it is, assembly just isn't going to
make enough difference. I have a couple programs that need to be sped
up. 3 orders of magnitude should get them fast enough that I can
actually get results sometime. I may be able to make one faster - but
only if I can reason about it easily, which assembly doesn't help. The
primary goal for a lot of scientists is getting a correct answer in a
reasonable amount of time, with minimal programming knowledge.
Assembly's not the tool for that job, either.

-- 
David Starner - starner@okstate.edu
"It's not a habit; it's cool; I feel alive. 
If you don't have it you're on the other side." 
- K's Choice (probably refering to the Internet)

Re: [OT] Gibson's vision of computer languajes

[Dave Poirier]

Marin David Condic wrote:
> "Dave Poirier" <instinc@users.sf.net> wrote in message
> news:3C8503D9.30209@users.sf.net...
> 
>>In my opinion, Assembly is the best suited language for any project, but
>>our society is unfortunately based on money, and to be commercially
>>viable assembly must of then be traded for less human-time consuming
>>alternatives.
>>
>>
> 
> Why do you think assembly language is the best suited language for any
> project? Most people would contend that assembly is occasionally necessary
> to accomplish things that can't be done (easily) in a high level language,
> but that it is to be avoided because of its developmental inefficiency, high
> probability of injecting errors and difficulty of maintenance. I'm curious
> as to what reasons you'd give to go against the conventional wisdom.
> 
> High level languages of almost any stripe generally mean you get the job
> done faster/better/cheaper & that's a good thing. What advantage do you get
> from assembler other than the occasional need to dip into it for some
> specific task not easily done otherwise - or the pure asthetic pleasure of
> doing something one might find interesting? (Liking a language is one thing.
> I like Ada for various asthetic reasons. Giving sound engineering reasons
> for using a particular language is another. Most of us here could give such
> reasons for using Ada in a given domain. What's the argument for favoring
> assembler?)

I think there is really only one thing that could explain why I value 
more assembler than other languages, it's that I probably value more the 
  time the machine spend executing the code than the time the human 
spend writing the code.

Lately I started considering myself more of an artist than a programmer, 
which is probably why I see the beauty in pure binary and take pleasure 
in counting cpu cycles and bytes of memory used.  For me, a program is 
"good" if it does the job without failing, a program is "nice" if it 
does the job fast and still does it "good", and a program is "awesome" 
if it's using one of the best sequence of instructions possible and fits 
in tight places while achieving the set goal.

I'm happy when I get a program done in HLL, but I'm not satisfied.  I'm 
happy and satisfied if I get a program done in assembly and I know it's 
the smallest/fastest thing I could create.

I just keep seeing all we could do with those computers we had in the 
'80s and read about what was done in the '60s and '70s (I wasn't there 
back then), and I find it amazing at the amount of work that could be 
done on those slow beasts.  Now we seem to have each a supercomputer 
sitting on our desk and we seem to be barely able to edit a text document.

Sure, the document is now fitting in a page, that we actually "see" as 
it will be printed, and some other stuff, but when I come to think about 
what the cpu actually execute, I just see billion of wasted cycles. 
Instead of improving programs, ppl buy bigger computers.

sorry for rambling on, hope I didn't confuse anybody (sorry for my 
english also).

EKS - Dave Poirier

Re: [OT] Gibson's vision of computer languajes

[John English]

Dave Poirier wrote:
> 
> I think there is really only one thing that could explain why I value
> more assembler than other languages, it's that I probably value more the
> time the machine spend executing the code than the time the human
> spend writing the code.

Oh, I value *my* time more highly than any mere processor -- an hour
of my time is worth much more than a few nanoseconds of my processor's.
However, I'm happy to value the time "the human" spends writing the
code more cheaply, just as long as I'm not the human in question...

-----------------------------------------------------------------
 John English              | mailto:je@brighton.ac.uk
 Senior Lecturer           | http://www.it.bton.ac.uk/staff/je
 Dept. of Computing        | ** NON-PROFIT CD FOR CS STUDENTS **
 University of Brighton    |    -- see http://burks.bton.ac.uk
-----------------------------------------------------------------

Re: [OT] Gibson's vision of computer languajes

[Marin David Condic]

"David Starner" <dvdeug@x8b4e53cd.dhcp.okstate.edu> wrote in message
news:a63js4$9qe1@news.cis.okstate.edu...
>
> "Best suited language" implies that running speed is the goal. It's
> frequently not, and sometimes where it is, assembly just isn't going to
> make enough difference. I have a couple programs that need to be sped
> up. 3 orders of magnitude should get them fast enough that I can
> actually get results sometime. I may be able to make one faster - but
> only if I can reason about it easily, which assembly doesn't help. The
> primary goal for a lot of scientists is getting a correct answer in a
> reasonable amount of time, with minimal programming knowledge.
> Assembly's not the tool for that job, either.
>

Many compilers are capable of producing code that is as tight - or nearly as
tight - as a skilled assembly programmer could do. (Given the same semantic
requirements, of course. No fair requiring runtime checks of the compiler
and not from the assembly expert...) And while with one given small segment
of code, the assembly programmer might just be able to beat the compiler, I
have faith that the compiler will do a consistently better job across the
whole body of code more often than the programmer.

So if speed is the only advantage, I'd contend that for reasonably large
programs, you won't find any significant improvements to doing the whole
thing in assembler versus doing it in a high level language. Clearly, one
might hand-optimize the small percentage of the code doing the really hard
work & buy an improvement, but it isn't going to change things much to do
the whole job in assembler. Machines work consistently better at producing
high quality stuff than do humans. A skilled human with a file and a raw
chunk of steel might produce a really fine automobile engine one day. But
stamping them out with CNC machinery is likely to produce a lot more of them
at a consistently higher quality.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/

Re: [OT] Gibson's vision of computer languajes

[Marin David Condic]

Well, I can't hardly argue with that. :-) Seeing the challenge as creating
an artistic thing of beauty negates any of the engineering arguments against
that technique. I used to know an artist back in college who did these very
large paintings that usually consisted of a monochrome background with some
squiggly lines on them of a different color. He'd labor for days/weeks with
oil paint, carefully getting the monochrome background absolutely smooth and
consistent in color & texture. It was really impressive that he could get it
to look just like the shiny paint job on a new car - all by hand & with
relatively primitive equipment.

But Detroit was getting the same appearance (or better) in a fraction of the
time and cost. Would you buy the car that was painted by the artist rather
than the computer-controlled spraying equipment? You might if money was no
object. It might amuse you that you got the same paint job as a Ford
Escort - but all laboriusly done by hand. But that would mean that only a
very small number of people with lots of money could have cars like that.

I won't say it has no value to do software entirely in assembler -
especially if viewed as an artistic endeavor. Its just that people don't buy
software to marvel at how perfectly it squeezes out every last nanosecond of
execution time or every last byte of memory. They buy it to get a job done.
Hence they want a reliable, affordable solution to their problems - not a
work of art.

Code bloat, OTOH, is a different issue. I'll agree that lots of software is
designed inefficiently and includes thousands of features that are
interesting from a marketing perspective, but don't really move the mission
forward. Sometimes I just want to type up some text and a command line
interface & glass-teletype would do the job just as well as a really spiffy
GUI based program that consumed megabytes of my disk drive. But I don't see
that as an Assembler versus HLL issue.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/


"Dave Poirier" <instinc@users.sf.net> wrote in message
news:3C85AA9A.7050201@users.sf.net...
>
> I think there is really only one thing that could explain why I value
> more assembler than other languages, it's that I probably value more the
>   time the machine spend executing the code than the time the human
> spend writing the code.
>
> Lately I started considering myself more of an artist than a programmer,
> which is probably why I see the beauty in pure binary and take pleasure
> in counting cpu cycles and bytes of memory used.  For me, a program is
> "good" if it does the job without failing, a program is "nice" if it
> does the job fast and still does it "good", and a program is "awesome"
> if it's using one of the best sequence of instructions possible and fits
> in tight places while achieving the set goal.
>
> I'm happy when I get a program done in HLL, but I'm not satisfied.  I'm
> happy and satisfied if I get a program done in assembly and I know it's
> the smallest/fastest thing I could create.
>
> I just keep seeing all we could do with those computers we had in the
> '80s and read about what was done in the '60s and '70s (I wasn't there
> back then), and I find it amazing at the amount of work that could be
> done on those slow beasts.  Now we seem to have each a supercomputer
> sitting on our desk and we seem to be barely able to edit a text document.
>
> Sure, the document is now fitting in a page, that we actually "see" as
> it will be printed, and some other stuff, but when I come to think about
> what the cpu actually execute, I just see billion of wasted cycles.
> Instead of improving programs, ppl buy bigger computers.
>
> sorry for rambling on, hope I didn't confuse anybody (sorry for my
> english also).
>
> EKS - Dave Poirier
>

Re: [OT] Gibson's vision of computer languajes

[Marin David Condic]

"John English" <je@brighton.ac.uk> wrote in message
news:3C85EE99.CDBCA9D6@brighton.ac.uk...
>
> Oh, I value *my* time more highly than any mere processor -- an hour
> of my time is worth much more than a few nanoseconds of my processor's.
> However, I'm happy to value the time "the human" spends writing the
> code more cheaply, just as long as I'm not the human in question...
>
You might not if you were cutting the paycheck for "the human" in exchange
for the programmed result. :-)

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/
> -----------------------------------------------------------------

Re: [OT] Gibson's vision of computer languajes

[Wes Groleau]

> I'm happy when I get a program done in HLL, but I'm not satisfied.  I'm
> happy and satisfied if I get a program done in assembly and I know it's
> the smallest/fastest thing I could create.

I'm happy when I meet all the requirements and move on
to something else.  If the requirements are that it
fit in 3K and execute in 10 milliseconds, I am not happy
to be spending time making it fit in 2K or execute
in 5 milliseconds.  That's time that could be spent
meeting additional requirements.

To put it another way, I don't get a lot of pleasure
fixing something that's not broken.

(On the other hand, I have a fairly broad definition
of 'broken')

-- 
Wes Groleau
http://freepages.rootsweb.com/~wgroleau

Re: [OT] Gibson's vision of computer languajes

[Warren W. Gay VE3WWG]

Dave Poirier wrote:

> Marin David Condic wrote:
>> "Dave Poirier" <instinc@users.sf.net> wrote in message
>> news:3C8503D9.30209@users.sf.net...
>>
>>> In my opinion, Assembly is the best suited language for any project, but
>>> our society is unfortunately based on money, and to be commercially
>>> viable assembly must of then be traded for less human-time consuming
>>> alternatives.
...
>> Why do you think assembly language is the best suited language for any
>> project? Most people would contend that assembly is occasionally 
>> necessary
>> to accomplish things that can't be done (easily) in a high level 
>> language,
>> but that it is to be avoided because of its developmental 
>> inefficiency, high
>> probability of injecting errors and difficulty of maintenance. I'm 
>> curious
>> as to what reasons you'd give to go against the conventional wisdom.
...
> 
> I think there is really only one thing that could explain why I value 
> more assembler than other languages, it's that I probably value more the 
>  time the machine spend executing the code than the time the human spend 
> writing the code.

...
 > EKS - Dave Poirier

If you're young, you can afford to spend time being "nostalgic". But that's
a different angle than saying "assembler is the best tool for the job". As
you get older, you will get more concerned with correct results. You will
not care that CPU cycles are wasted. The important thing is are you wasting
your OWN cycles?

By way of illustration, I would much rather have Windows-2010 on my desktop

running sluggishly (if necessary) as long as it was reliable and secure.
Being fast, doesn't buy me anything if I keep losing time and work because
of crashes and inept programs/security. Sluggish I'll accept, because I
know that by this time next year, new CPUs will be available at low cost,
for me to upgrade to. It is now becoming easier to throw money at hardware,
than it is to get good quality software.  It is easier to buy faster equipment,
than it is to hire good software engineers.

That is why I for one, have embraced the "Ada way". My time spent writing
new Open Sourced software is much better spent getting it right, and getting
the quality. Call me old fashioned, or inept if you like, but I do like the
having the compiler straightening me out, when I need it. It saves me a
great deal of my OWN limited time.

Who really cares about idle/overworked CPUs? I am the master of my CPU, and
not slave to it's needs. Crack the whip on the CPU and make it earn its keep!

Caveat: I am not saying that you never have to be concerned about CPU budgets
in certain applications. But in the run-of-the-mill PC desktop case, it often
is the least of your concern outside of semi-realtime media based programming.

-- 
Warren W. Gay VE3WWG
http://home.cogeco.ca/~ve3wwg

Re: [OT] Gibson's vision of computer languajes

[Marin David Condic]

"Warren W. Gay VE3WWG" <ve3wwg@home.com> wrote in message
news:3C8651EF.40104@home.com...
>
> Caveat: I am not saying that you never have to be concerned about CPU
budgets
> in certain applications. But in the run-of-the-mill PC desktop case, it
often
> is the least of your concern outside of semi-realtime media based
programming.
>
All engineering is about tradeoffs. Agonizing over CPU cycles is
occasionally important, but usually not. Spending time optimizing CPU cycles
at the expense of optimizing other things when you've got more than enough
CPU cycles to do the job is not focusing on the mission. It can be
entertaining, but it doesn't move the mission forward. It can even be moving
the mission backward. Its important to remember that when selecting the
tools to do a job - hence my preference for Ada in most cases. It makes good
tradeoffs between speed, safety, reliability, ease of implementation, etc.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/