1. Challenge for speed freaks
On Thursday 14 February 2002 10:34 pm, C. K. Lester wrote:
> That's what I said, so I emphasized the contrast: EUPHORIA is faster
> than Python, so, all things being equal (each using a C graphics
> library), EUPHORIA would be the faster one.
I'll agree, Euphoria is generally (but not always) faster than Python, or
Perl, or Java or......[feel free to add almost any other language here]
I would like to see comments, especially from Rob, on these questions:
Why, if Euphoria is so much faster, are there hundreds or thousands of people
who use perl, python, or java, etc. for every person who uses Euphoria.
Why do the bookshelves contain dozens of books on perl, python, and java,
etc. but none for Euphoria?
Why do computer magazines run articles on these other languages, but
make no mention of Euphoria?
Could it be that *fast* just isn't the most important criteria for most
programmers?
I don't buy the argument that marketing makes the difference, either -
a lot of money has been spent to promote Java, but the rest? Not that
I know of.
Regards
Irv
2. Re: Challenge for speed freaks
Ray Smith writes:
> With the processing power of pc's these days is speed a real issue?
> Except for specialized applications like action games, 3d modelling,
> simulations etc speed is almost a non issue.
I just finished crunching 17 Mb of log data from RapidEuphoria.com
The log records each page hit, each .zip downloaded, etc, along with
the referring URL, IP address etc. etc.
My Euphoria program took 1 minute to give me lots of interesting
highly customized information. Would I want to wait half an hour
for Python or Perl? (My ISP has a free log analyzer. It provides lots of
data, but little useful information that I need to evaluate sources of
advertising.)
Speed will always be valuable.
Irv Mullins writes:
> I'll agree, Euphoria is generally (but not always) faster than Python, or
> Perl, or Java or......[feel free to add almost any other language here]
> I would like to see comments, especially from Rob, on these questions:
> Why, if Euphoria is so much faster, are there hundreds or
> thousands of people who use perl, python, or java, etc. for every
> person who uses Euphoria.
Java:
Sorry, I don't have one billion dollars to spend on marketing.
Perl:
I used to work on UNIX systems back in the 80's when
Perl came out. It was something of a novelty, being open source
and free. It filled the gap between slow shell scripts and compiled C.
There weren't many alternatives. Since there were so many
different flavors of UNIX and so many different CPU's, it was
a practical necessity for any widespread language to be open source.
Nobody had the time to personally support binary executables for dozens
of UNIX platforms. In short, by giving away Perl for free in source form,
Larry Wall filled a gap that was there. It didn't really matter if Perl was
particularly good, or what programming philosophy it had.
Python:
Someone said Python was started 3 years before Euphoria.
So I guess it has been around 12 years.
Python is another a UNIX-based language. It seems to be
gaining market share from Perl. Perl is a grab bag of
ad hoc features with no concept behind it. Python tries
to be more sophisticated.
Euphoria was released for Linux (a subset of the UNIX world)
only 2 years and 2 months ago. It will take time to win over
the open source, everything must be totally free, crowd.
> Why do the bookshelves contain dozens of books
> on perl, python, and java, etc. but none for Euphoria?
Of course they have more users, and
maybe things have changed, but the free, on-line documentation
for Perl and Python, as of a couple of years ago, was terrible
(probably to promote books sales) Also, these languages
are much "bigger" and much more complicated,
so people feel a strong need to run out and buy a book.
Paper is a rather archaic medium, and as soon as a new release
comes out, a paper book is obsolete.
> Why do computer magazines run articles on these
> other languages, but make no mention of Euphoria?
Computer magazines have mentioned Euphoria,
but again you are just milking the number of users issue.
> Could it be that *fast* just isn't the most important
> criteria for most programmers?
That's what I say to all the C/C++ programmers 
> I don't buy the argument that marketing makes the difference,
> either - a lot of money has been spent to promote Java,
> but the rest? Not that I know of.
OK Irv, you win.
Since Euphoria isn't the most "popular" language in the world,
there must be something tragically wrong with it,
and I might as well quit.
You should go to Salt Lake City and tell all the athletes
who didn't get a medal that they've wasted their time
and should quit too.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
3. Re: Challenge for speed freaks
On Friday 15 February 2002 09:58 am, jaspers_post at hotmail.com wrote:
> First of all I think it's not really fair towards Rob to ask him why
> Euphoria isn't used more than other languages. It's like asking the
> author of a good book why his book isn't sold better.
And yet, that is exactly the question the publisher and the author of
a book which doesn't sell ask themselves. An honest answer can
lead to the next book being (and selling) much better.
They do this by reading reviews, and by talking with people who have
read the book. Talk is not enough, however. if they want the next book
to sell better, they act on what they've learned.
Regards,
Irv
4. Re: Challenge for speed freaks
On Friday 15 February 2002 01:29 pm, C. K. Lester wrote:
>
> He indicates at times that he has secret and proprietary code...
> Probably valuable to other C coders.
>
The speed of Euphoria would indicate that he does.
Common sense would say that with the millions of C coders out there,
someone else would have discovered the secret as well.
Perhaps they aren't trying very hard.
Regards,
Irv
5. Re: Challenge for speed freaks
On Friday 15 February 2002 01:29 pm, C. K. Lester wrote:
>
> Andy Serpa wrote:
> > There is *no way* a "one-man" language
> > existing as a commercial entity will ever be huge.
>
> Not without given a lot of time? Even slow growth is growth, and as long
> as that doesn't stop, the language is very popular in, say, 10 years? 15
> years? :)
Not unless it keeps up with current needs.
10 years ago, who knew we'd want to download mpeg's of Britney Spears?
Nobody - it wasn't even possible.
Yet, somehow, software has been created to let us do just that.
If slow means falling farther and farther behind, then more and more people
will have move on to software which does keep up. That's ok too, if your
definition of "one-man language" means "one user".
Regards,
Irv
6. Re: Challenge for speed freaks
On Friday 15 February 2002 09:24 am, Ray Smith wrote:
> You mention 3 popular languages. If you do a web search there are
> probably a couple of hundred other languages around all struggling to
> find users and developers. You'll also find alot of languages with
> "this page last updated" sometime well into the last century!
Last time I looked, I found a list of over 1,700 languages.
A large number of these are probably useless: either designed
to solve a very specific problem, and unsuited for anything
else, or the result of someone's master's thesis in CS, theoretically
sound, but unbearably slow.
There's always a chance that a truly outstanding language has been
developed, and no one but the author has tried it. It would be a shame
to overlook something like that.
Regards,
Irv
7. Re: Challenge for speed freaks
On 15 Feb 2002, at 13:28, Robert Craig wrote:
<snip>
> OK Irv, you win.
> Since Euphoria isn't the most "popular" language in the world,
> there must be something tragically wrong with it,
> and I might as well quit.
>
> You should go to Salt Lake City and tell all the athletes
> who didn't get a medal that they've wasted their time
> and should quit too.
Or do like i do, and ignore nearly everyone. Rob, the language could use a
few more commands. The gui came along to enable easier use. Basic was
not meant as a mainstream language, but even microsoft has a Basic
language, two: vbs and vb. People want to get a programming task done,
they don't wanna learn a new complex puter language philosophy.
Karl, can goto use the contents of a variable as a target, please? sorta
pseudo-case-like?
s = {"money","monkey","trap","tree"}
if match(input,s)
then goto eval(input)
else -- process error
end if
-- targets follow --
:money:
:monkey:
etc
Like i said 2yrs(?!) ago here, it should be easy to make interpreters smart,
and make radical code for the programmer easy. I mean really, i have code
out there that lets channel ops spec what word they wish to use to control
script behaviour in a bot, in whatever language they wish. But it's written in
mirc.
Kat
8. Re: Challenge for speed freaks
On Friday 15 February 2002 11:10 am, SR.Williamson wrote:
> Just curious, but what can Eu NOT do that it needs to do to attract more
> people? Irv has talked before about the limits of Eu, but I don't recall
> anything specific, probably because I'm not a programmer and it just
> goes right over my head. What keeps people from building stuff other
> than toys with it?
It's not so much what Euphoria can't do - you can do almost anything with
enough work. The same statement applies to assembler, or C, or whatever.
It does lack a straightforward way to do many of the things that programmers
take for granted. Those things have been discussed ad infinitum on this list,
(look thru the archives) and Rob is adamant that he isn't ever going to add
them.
I've supported and even promoted Euphoria for a long time, but I can't do so
any longer. Any programmer with 6 months' experience can find something to
laugh about in Euphoria. I prefer a slower language which is taken seriously.
Regards,
Irv
9. Re: Challenge for speed freaks
Dear EU users,
I am not very new list's correspondent and it is strange enough
for me to read the things I can read here periodically.
Say, there were no translators Euphoria to C, but there was cry
of some people about, so to say, "low" speed Euphoria.
Now we can just tranclate/compile and speed up our old programs
to compiled C from the best in the World interpreted speed.
But now we have just the same thing -- cry about too high
speed of Euphoria and Rob must prove to some people,
to programmers, why the speed is just a very good thing
in the computer world.
If you do not want high speed programs, just do not program
in Euphoria.
You want Python ?
What a problem ? Delete EUPHORIA directory from HD,
install Python, write programs in Python.
You want Java ?
What a problem ? Delete EUPHORIA directory from HD,
install Java, write programs in Java.
You want Perl ?
What a problem ? Delete EUPHORIA directory from HD,
install Perl, write programs in Perl.
You want Lua ? ... Hmmm ...
What a problem ? Delete EUPHORIA directory from HD,
install Lua, write programs in Lua.
You want APL ? ... Hmmm ...
What a problem ? Delete EUPHORIA directory from HD,
install APL, write programs in APL.
You want Ruby ? ... Hmmm ...
What a problem ? Delete EUPHORIA directory from HD,
install Ruby, write programs in Ruby.
You want ... What do you want ? I can not understand ...
Hey, it seems to be, you want famous and rich Rob, and famous
and rich youself for all your work about Euphoria ...
But with PD full featured Euphoria Rob may not become
famous and rich as Bill of MS.
So my recipe is :
1. No any Euphoria PD packages.
2. Price UP to $1000 for *1 floppy disk*.
3. No any downloads of free software from RDS.
BUT there is one great BUT ...
Speed of Euphoria -- gives us the cycles of our CPUs,
just our time, just our life, just our main property.
You want give away your life to Python, Perl, Java ... Yes ?
Are you crazy ? No ?
Why, No ?
Well.
BUT there is one additional BUT ...
If someone think Euphoria is not popular, this problem is not
an Euphoria's problem or an Euphoria programmer's problem.
This problem is just that someone's problem.
If your train is ran away from you then this is your problem,
not train's problem.
Is it clear for Euphoria community ?
Just some thoughts.
Regards,
Igor Kachan
kinz at peterlink.ru
10. Re: Challenge for speed freaks
On Friday 15 February 2002 01:28 pm, Robert Craig wrote:
>
> OK Irv, you win.
> Since Euphoria isn't the most "popular" language in the world,
> there must be something tragically wrong with it,
> and I might as well quit.
Perhaps you already have. You hinted as much in your recent post:
"People should not focus so much on what I'm going to do.
My job is to stimulate, and harness the creativity of the
Euphoria community. I can't predict what's going to happen."
Isn't that another way of saying "if you want it done, do it yourself"?
Over the years, I've watched many dozens of well-qualified
people make well-meaning and polite suggestions about how to improve
Euphoria, how it might be made more useful, and how it might better be
promoted, (so you would make more money from it). Usually your
response was some version of "I don't see the need for that".
Most of them eventually gave up trying to be helpful and moved on.
I thought that, out of a sense of pride in your creation, you would clear up
the few rough spots in Euphoria, instead, you refuse to admit that there are
any rough spots, no matter how many people point them out.
I thought that, out of a sense of competitiveness, you would realize the
importance of Windows support, and be quick to add it. But Dave, Derek,
Judith and others have had to do most of that.
I thought perhaps you were tired of Euphoria, and just didn't want to bother
adding anything new - a charge you hotly denied.
I thought perhaps you weren't making any real money from Euphoria, maybe it
was just an annoying sideline to your real occupation - you claimed this was
not true, yet you apparently see no need and make little effort to make it
more popular. (A move which would translate directly into more money for you.)
I even thought you might have done what I often do - programmed yourself into
a corner, and don't dare make any major changes for fear of ruining
the whole thing. I have been assured this is not the case.
I have thought all these things, only to be told I was wrong.
I'm glad to know I was wrong.
Yet your actions, along with your words quoted above, give the
impression that you'd rather be doing something else.
Let me suggest a career in undertaking. Your customers never
talk back, and you get to bury your mistakes.
> You should go to Salt Lake City and tell all the athletes
> who didn't get a medal that they've wasted their time
> and should quit too.
Other than the 2 or 3 who will get TV contracts from this, yes,
I'd say the majority have wasted years of their time.
But that's another subject.
Irv
11. Re: Challenge for speed freaks
- Posted by rforno at tutopia.com
Feb 16, 2002
Just my two cents...
I studied Java and two years I taught a Java course. But I cant call me a
proficient Java programmer. Moreover, I hate Java.
A year ago I started using Euphoria. I like using it. Two months from now I
expect to be teaching an Euphoria course at a local University.
Problem is, in how much time one can proficiently program in a language?
----- Original Message -----
From: "Ray Smith" <smithr at ix.net.au>
To: "EUforum" <EUforum at topica.com>
Subject: RE: Challenge for speed freaks
>
>
> Irv Mullins wrote:
>
> > Why, if Euphoria is so much faster, are there hundreds or thousands of
> > people
> > who use perl, python, or java, etc. for every person who uses Euphoria.
>
> Perl, Python and Java are free.
> This is a huge factor!!!!
> Everyone wants a free lunch. Even though the price of Euphoria
> is inexpensive it's the concept behind it.
>
> Perl I have heard is very fast at text processing.
> I'm sure there would be a large number of programs (text manipulation)
> that would run faster in Perl. I haven't seen any action games in
> Perl ... but that isn't the aim of the language.
>
> Python and Java have been built from the ground up as OOP languages.
> Perl has been adapted to support OOP.
> This is another huge factor.
> As much as everyone on this list keeps saying they can live without
> classes everyone else in the world says they can't.
> Classes are a very powerful tool and I'd be very surprised if you
> could name one "newly" developed NON OOP language that has been
> successful in the last 10 years.
> Do you think a couple of dozen people on a mailing list are going to
> change the rest of the worlds mind re OOP?
>
> The 3 languages you mention are highly portable.
> I know that Java and Python have threads support, exception handling,
> many many more libraries available ... most freely downloadable.
>
> > Why do the bookshelves contain dozens of books on perl, python, and
> > java,
> > etc. but none for Euphoria?
>
> A simple question of user base.
> It wasn't that long ago (2 years - probably 3!) that the first
> Python book was released.
>
>
> > Why do computer magazines run articles on these other languages, but
> > make no mention of Euphoria?
>
> user base again.
>
>
> > Could it be that *fast* just isn't the most important criteria for most
> > programmers?
>
> Three of the really big selling points for Euphoria are:
> "fast",
> "simple", and
> "small".
>
> With the processing power of pc's these days is speed a real issue?
> Except for specialized applications like action games, 3d modelling,
> simulations etc speed is almost a non issue.
> For these specialised cases you can still write sub routines in C.
> Just like we do for Euphoria.
>
> Simple is a great thing to have. Simple doesn't have to mean you
> don't have threads, or classes, or exception handling. It just means
> when you do have these things they are implemented in a consistant
> well structured way.
> If they "are" implemented and people don't want to use them ...
> then they don't!
>
> Small! - who copies files by floppy disk anymore?
> All these other languages have 10mb or more downloads,
> no one cares. It takes maybe an hour on a standard modem to download
> them.
> How much effort do people put into developing applications?
> Is an extra hour to setup, and an extra 50MB of disk space going to
> stop anyone from using Euphoria? I'd say no.
>
> > I don't buy the argument that marketing makes the difference, either -
> > a lot of money has been spent to promote Java, but the rest? Not that
> > I know of.
>
> How many of these popular languages are developed by one person?
> It's impossible, can't be done.
>
> I don't mean this is an offense way Rob (if your reading!),
> but what major software development project has become successful
> by the authors just "seeing" what happens and not having any long
> term plans or goals? That's Robs way, that's his choice and no one
> can judge him and say he's wrong. What I can say is success won't
> come knocking on your door!
>
>
> For the PRO side:
>
> It's a bit harsh to look at some popular languages and say
> "why isn't Euphoria like language X!"
>
> Java is a product developed by hundreds of people.
> Python has commercial backing and has a number of full time staff
> developing as well as a large band of eager users.
>
<snip>
>
>
>
12. Re: Challenge for speed freaks
- Posted by rforno at tutopia.com
Feb 16, 2002
Oopps... I meant "two years ago I taught a Java course"
----- Original Message -----
From: <rforno at tutopia.com>
To: "EUforum" <EUforum at topica.com>
Subject: Re: Challenge for speed freaks
>
> Just my two cents...
> I studied Java and two years I taught a Java course. But I cant call me a
> proficient Java programmer. Moreover, I hate Java.
> A year ago I started using Euphoria. I like using it. Two months from now
I
> expect to be teaching an Euphoria course at a local University.
> Problem is, in how much time one can proficiently program in a language?
> ----- Original Message -----
> From: "Ray Smith" <smithr at ix.net.au>
> To: "EUforum" <EUforum at topica.com>
> Sent: Friday, February 15, 2002 11:24 AM
> Subject: RE: Challenge for speed freaks
>
>
> > Irv Mullins wrote:
> >
> > > Why, if Euphoria is so much faster, are there hundreds or thousands of
> > > people
> > > who use perl, python, or java, etc. for every person who uses
Euphoria.
> >
> > Perl, Python and Java are free.
> > This is a huge factor!!!!
> > Everyone wants a free lunch. Even though the price of Euphoria
> > is inexpensive it's the concept behind it.
> >
> > Perl I have heard is very fast at text processing.
> > I'm sure there would be a large number of programs (text manipulation)
> > that would run faster in Perl. I haven't seen any action games in
> > Perl ... but that isn't the aim of the language.
> >
> > Python and Java have been built from the ground up as OOP languages.
> > Perl has been adapted to support OOP.
> > This is another huge factor.
> > As much as everyone on this list keeps saying they can live without
> > classes everyone else in the world says they can't.
> > Classes are a very powerful tool and I'd be very surprised if you
> > could name one "newly" developed NON OOP language that has been
> > successful in the last 10 years.
> > Do you think a couple of dozen people on a mailing list are going to
> > change the rest of the worlds mind re OOP?
> >
> > The 3 languages you mention are highly portable.
> > I know that Java and Python have threads support, exception handling,
> > many many more libraries available ... most freely downloadable.
> >
> > > Why do the bookshelves contain dozens of books on perl, python, and
> > > java,
> > > etc. but none for Euphoria?
> >
> > A simple question of user base.
> > It wasn't that long ago (2 years - probably 3!) that the first
> > Python book was released.
> >
> >
> > > Why do computer magazines run articles on these other languages, but
> > > make no mention of Euphoria?
> >
> > user base again.
> >
> >
> > > Could it be that *fast* just isn't the most important criteria for
most
> > > programmers?
> >
> > Three of the really big selling points for Euphoria are:
> > "fast",
> > "simple", and
> > "small".
> >
> > With the processing power of pc's these days is speed a real issue?
> > Except for specialized applications like action games, 3d modelling,
> > simulations etc speed is almost a non issue.
> > For these specialised cases you can still write sub routines in C.
> > Just like we do for Euphoria.
> >
> > Simple is a great thing to have. Simple doesn't have to mean you
> > don't have threads, or classes, or exception handling. It just means
> > when you do have these things they are implemented in a consistant
> > well structured way.
> > If they "are" implemented and people don't want to use them ...
> > then they don't!
> >
> > Small! - who copies files by floppy disk anymore?
> > All these other languages have 10mb or more downloads,
> > no one cares. It takes maybe an hour on a standard modem to download
> > them.
> > How much effort do people put into developing applications?
> > Is an extra hour to setup, and an extra 50MB of disk space going to
> > stop anyone from using Euphoria? I'd say no.
> >
> > > I don't buy the argument that marketing makes the difference, either -
> > > a lot of money has been spent to promote Java, but the rest? Not that
> > > I know of.
> >
> > How many of these popular languages are developed by one person?
> > It's impossible, can't be done.
> >
> > I don't mean this is an offense way Rob (if your reading!),
> > but what major software development project has become successful
> > by the authors just "seeing" what happens and not having any long
> > term plans or goals? That's Robs way, that's his choice and no one
> > can judge him and say he's wrong. What I can say is success won't
> > come knocking on your door!
> >
> >
> > For the PRO side:
> >
> > It's a bit harsh to look at some popular languages and say
> > "why isn't Euphoria like language X!"
> >
> > Java is a product developed by hundreds of people.
> > Python has commercial backing and has a number of full time staff
> > developing as well as a large band of eager users.
> >
> <snip>
>
> >
>
>
>
13. Re: Challenge for speed freaks
Ray Smith writes:
> * has (or will) the "Translater" stifle the developemnt
> of Euphoria as any "offical" feature will need twice as much work?
The interpreter and translator share the same source code
for scanning, parsing, symbol table lookups, and run-time routines.
Where they differ is that the interpreter uses immediate
execution of intermediate-language op codes, whereas the
translator converts those op codes into C source code, using
a multi-pass process with various optimizations (eliminating
unused routines, eliminating unnecessary Ref/DeRef's,
deducing information about the possible types of data stored
in a variable to reduce if-statements and unnecesary code).
There are even some cool global optimizations, such as
checking all calls to a routine and figuring out the possible
parameter types, e.g. you might declare a parameter as "object",
but if the translator finds that you are only passing integers
it will use that info inside the routine to streamline the code.
If you are only passing a specific hard-coded value, say 99,
it will take advantage of that too.
I would guess that most future changes are likely to be in the
shared code, and not require any duplication of design or coding effort,
just a bit of extra testing.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
