Oberon-Growth-Limit was Re: Oberon: beginners questions 
Author Message
 Oberon-Growth-Limit was Re: Oberon: beginners questions


Quote:
>It would be nice if Oberon had a postscript viewer, especially
>since ETH distributes a lot of documents in PS format, but
>there is none yet that I'm aware of.

No PostScript viewer?

Yet another symptom of the Oberon-Growth-Limit. Interesting how that
thread turned into a question about whether or not to conditionally
compile.

I notice a few beginners asking questions on this group, so the language/OS
seems to be attracting some notice. Too bad it also seems that there is
little for Oberon to offer other than its own elegance.

I need to keep this short because I have to get back to my thesis
work where I use:

        (1) Maple for handling symbolic algebra,
        (2) C++/C because it is a joy compared to fortran 77,
        (3) FORTRAN 77, because rewriting that code is a waste of time
        (4) IDL to graph, and plot the results acquired with the above
        (5) AMSTeX/AASTeX/LaTeX/TeX for writing & formatting
        (6) all, of course, in a distributed, multi-platform UNIX/X11
            environment.

Sound familiar? The question to be answered is: How will oberon help me
do what I do better, and at what cost?

Time, money, and risk are involved in writing code, and at this time,
I do not see that Oberon is competitive with the computing environment
I have outlined above. Not if I have to recreate the above utility myself.
Furthermore, the Oberon community does not seem sufficiently large to
produce a rich, attractive variety of useful applications.

If I'm wrong, please let me know. The ideas and philosophy of Oberon are
quite attractive. It is a pity to see them languishing like Lisp, another
failing, yet good idea.

Oh well, perhaps some of those ideas will filter into future versions of
the more popular programming languages.

Regards,

Mike Rilee

Cornell Univ., Lab. of Plasma Studies, 369 Upson Hall, Ithaca, NY 14853-7501
Ph: 607 255 4723 Fx: 607 255 3004
<A HREF=" http://www.*-*-*.com/ ;>http</A>



Fri, 21 Nov 1997 03:00:00 GMT  
 Oberon-Growth-Limit was Re: Oberon: beginners questions

As Popeye would say:
 "I've had alls I's can take, and I can't takes no more."

This thread has run the both ends of the threshold of validity, but
very few postings have been down the middle of the road.

Quote:

>> does not seem to be extending its user community as one might hope.
>> If and when it fails, it will be important to understand why the
>> theory that we agree upon in principle has had difficulty in practice.
>The points you and others have already mentioned.
>>           I recently resampled doing the same things
>> in V4 (SPARC), V4 for MS-Windows, System3 for SPARC, and System3 for
>MS-Windows, and Oberon/F.   Why ? This shouldn't be necessary.
>> too many "improving upon" Oberon-2.  As discussed in the previous paragraph,
>> part of the problem could be too much grouth, that is, growth that is more
>> cancerous than healthy.

Yes, this is indeed a problem but it is only a minor problems and
belies the fact that most people don't want to do programming.  In
fact, this argument, while valid and important, has little impact on
the reason Oberon is not mainstream.  According to many of the people
posting to this thread, and herein lies the big inconsistency in the
arguments, people want to _use_ their computer, not program it.

This is a problem of 'perceived audience' and the only thing that a
person intending on purveying an Oberon system can do is define
'target audience' for themselves and work with that.  Completely
ignore the naysayers who want all the applications in the world (one
argument) and ignore all those people who want interoperability
between all the different versions (another argument).

The world is not going to become Oberon-aware if either of these two
pieces of advice are taken, and here's why:

1) I want all the tools in the world.

[Apologies to all who have read this before] As a short discourse,
there was a guy at the Oakwood meeting in '93 who wanted to add all
sorts of C++-isms to Oberon.  After a heated argumented ensued two
things were discovered about this person: first, he never had really
used Oberon and secondly that, even if all his proposals were
adopted, he probably still wouldn't use Oberon.

The m{*filter*}of the story is what I would like to put forth as Hutt's
Law of Computer Users:

   A user of a particular software package will not switch to a new
   operating system to use a different version of the same software.

   Adenda:

   A user of a piece of Microsoft software will always switch to a
   new OS to get the latest version of anything (in hopes that it
   fixes at least 15% of the problems with the current version).

That being said, it's plain to see that users (and I use the word
strongly: user! Not hacker, not hobbiest, not techno-nerd, not
solder-gun-weilding HAM radio operator)

2) I want to have >85% compatbility between different versions of the
   Oberon systems -- even though they are by different vendors and
   are internally different and have different target audiences.

Gee, I'd like Unix to do the same thing.  I'd like MS-DOS to do the
same thing.

Oberon is not going to make great inroads into commercial development
any time soon -- period.  Oberon is going to be driven by a few
dedicate people developing (and distributing) worthwhile software at
a reasonable cost -- and this core group is going to have their
favorite flavor of Oberon.  This group is not going to run around
with different configurations to make sure that their software works
on all different platforms and all different compilers -- that's the
result of working in a small group like Oberon.

While I'm on the topic of compatibility, I'll toss out my pet peeve
about V4: it's Text model is too high level.  Texts, IMHO, should be
derived from a lower-level text abstraction.  The reason for this
complaint is that some of the Oberon.Mod messages have a text field
in them - and if you are writing a different text model (not derived
from Texts), it is more difficult (and less efficient) to implement
some of the messages (like CopyOverMsg).

Quote:
>Concentrate an creating more common code, running under all Oberon Systems.
>> This is where Wojtek's analogy to Linux is not at all foolish.  
>But you can't compare Oberon to Linux, Linux is from the application viewpoint
>just a new clone of the dino UNIX. Therefore a lot of code is already available
>for free. gcc, bison, emacs, X11, shell commands, ...
>Oberon must still get this repository of generic code.
>> we are probably too fragmented and individualistic to generate and build
>> what is necessary.  In the commercial, "wanna-be" commercial, or "other" world
>we have OM2, POW, Oberon/F, Logic Magicians V4, ...  Which, if any, should
>> be backed?

Choose the one which best fits your goals and desires.
Do you want to make stand-alone executables on a variety of hardware? (OM2)
... end-user, prettified apps? (POW, System3/Gadgets)
... with the widest array of hardware and software specs? (Oberon/F)
... to have lots of source code and an open environment? (LM V4)

Quote:
>> Oberon must still get this repository of generic code.

>That is a major part of the point.  The availablity of repositories of code
>inside and outside ETHZ is not growing, as one might hope, for a system that is
>technically one of the best and fastest software development
>languages/environments.
>Why this seems to be true is part of what we are debating on this thread.

This is an interesting observation; something which I must think
about.  Should I release low-level datastructures code for the good
of the community or be more elitest (err, capitalistic) and only sell
it.  I can see good and bad for both arguments.  Any input?

Quote:
>> > Above all, I believe we need the leadership (or management, which
>> > I have less faith in) that Wojtek was basically asking for.  The Internet
>> > and c.l.o. is probably very poor at providing that, but as a forum for
>> > expression of real leadership it can speak to a lot of us.

Very poor, indeed.  You should have seen how fast the NetM2 project
fizzled -- it was as if a group of people has lots of input on how to
use a piece of flint and steel -- but few had any real hands-on
experience building a fire that way.  Unfortunately, as soon as it
was apparant that no one really knew if they wanted a big fire or a
little fire, everyone turned away.

One drawback to using usenet as a means of following a leader is the
(unspoken) ban on advertising.  It would be somewhat unethical for a
vendor (like ModulaWare or OUS or me) to come along and tell everyone
to follow their flag.

However, the converse is not true: the public can adopt a vendor and
guide the vendor (provided its ears are tuned correctly) in the
direction the public wants to go.  So, rather than being reactive,
the public ought to be outlining what they want from the vendors
(within reason -- in a market as small as Oberon, don't expect one
vendor to adopt the conventions of another; don't expect emulation of
OLE2, etc., etc.)  Any input?

Quote:

>>It would be nice if Oberon had a PostScript viewer, especially
>>since ETH distributes a lot of documents in PS format, but
>>there is none yet that I'm aware of.

>No PostScript viewer?

>Yet another symptom of the Oberon-Growth-Limit. Interesting how that
>thread turned into a question about whether or not to conditionally
>compile.

Hogwash and a red herring.  There is little correlation between
Oberon's growth limit and a postscript viewer.

Quote:

>I notice a few beginners asking questions on this group, so the language/OS
>seems to be attracting some notice. Too bad it also seems that there is
>little for Oberon to offer other than its own elegance.

>I need to keep this short because I have to get back to my thesis
>work where I use:

> (1) Maple for handling symbolic algebra,
> (2) C++/C because it is a joy compared to FORTRAN 77,
> (3) FORTRAN 77, because rewriting that code is a waste of time
> (4) IDL to graph, and plot the results acquired with the above
> (5) AMSTeX/AASTeX/LaTeX/TeX for writing & formatting
> (6) all, of course, in a distributed, multi-platform UNIX/X11
>     environment.

>Sound familiar? The question to be answered is: How will oberon help me
>do what I do better, and at what cost?

Pure and simple: tripe, garbage, baloney.  This argument about Oberon
is as old as the newsgroup.  If you want these things, take some time
to write them.  If they existed, would you drop all that you have now
and only use Oberon?  Have you done anything to get _those_ vendors
to support the Oberon system? (No? then why are you {*filter*}ing about
it?)  Have you personally done anything to help the Oberon community
move to having these types of applications?  If you want to be at the
leading edge of a new system, with no real commercial support, then
you must be willing to sacrifice everything which most people call
'personal life'.

Quote:
>Time, money, and risk are involved in writing code, and at this time,
>I do not see that Oberon is competitive with the computing environment
>I have outlined above. Not if I have to recreate the above utility myself.
>Furthermore, the Oberon community does not seem sufficiently large to
>produce a rich, attractive variety of useful applications.

I'm sick and tired of this type of crap.  Why does everyone have the
agenda that Oberon must somehow compete on the level of large,
multimillion dollar corporations?  Oberon is NOT anywhere near the
market saturation nor does it have anywhere near the money to do
these things.  Unless people like you are willing to shut up and
start doing something about it, we are always going to have exactly
what _YOU_ put into it: NOTHING.

- Show quoted text -

Quote:

>If I'm wrong, please let me know. The ideas and philosophy of Oberon are
>quite attractive. It is a pity to see them languishing like Lisp, another
>failing, yet good idea.

>Oh well, perhaps some of those ideas will

...

read more »



Fri, 21 Nov 1997 03:00:00 GMT  
 Oberon-Growth-Limit was Re: Oberon: beginners questions
  [About Oberon growth limitations]

Quote:
> Yes, this is indeed a problem but it is only a minor problems and
> belies the fact that most people don't want to do programming.  In
> fact, this argument, while valid and important, has little impact on
> the reason Oberon is not mainstream.  According to many of the people
> posting to this thread, and herein lies the big inconsistency in the
> arguments, people want to _use_ their computer, not program it.

I would say: a kind of user we are talking about (like myself) wants
to use a computer through programming, but wants to program as little
as s/he possibly can. (In my case it means "up to a few klines of code
per month".) By "as little as I can" I mean "no whole new applications",
though avoiding this is not always possible. My desire is to inherit
as much code as I can, in order to limit painful programming as much
as I can. My own pain with Oberon is too little code "out there" for
my own consumption. I am not saying there is no code. I am saying there
is no right code for my purpose.

Having confessed this much, I want to ask: where is the inconsistency?

Quote:

> The m{*filter*}of the story is what I would like to put forth as Hutt's
> Law of Computer Users:

>    A user of a particular software package will not switch to a new
>    operating system to use a different version of the same software.

 You are right on target, as usual !

 However, the most recent System-3 delivers "different version of the
same software", like Web browser and the like. Any inconsistency here?

Quote:
> That being said, it's plain to see that users (and I use the word
> strongly: user! Not hacker, not hobbiest, not techno-nerd, not
> solder-gun-weilding HAM radio operator)

I am not sure I understand?

 [...]

Quote:
>>> Oberon must still get this repository of generic code.

>>That is a major part of the point.  The availablity of repositories of code
>>inside and outside ETHZ is not growing, as one might hope, for a system
>> that is technically one of the best and fastest software development
>>languages/environments.
>>Why this seems to be true is part of what we are debating on this thread.

> This is an interesting observation; something which I must think
> about.  

I think this is the most basic and most troublesome question:
if it is so good, then why the hell it fares so badly?

I am glad the Oberon Paladin will think of it. If only ETH could also
devote some thought to the issue, we might be all set. They took water
in their mouth, though.

Quote:
> Should I release low-level datastructures code for the good
> of the community or be more elitest (err, capitalistic) and only sell
> it.  I can see good and bad for both arguments.  Any input?

Releasing sources is what drove Linux where it is now. Not releasing
sources is what (probably) drove Oberon System where it is now.
On the other hand, why should you be more Papal than the Pope?

Quote:
>>> > Above all, I believe we need the leadership (or management, which
>>> > I have less faith in) that Wojtek was basically asking for.  The Internet
>>> > and c.l.o. is probably very poor at providing that, but as a forum for
>>> > expression of real leadership it can speak to a lot of us.

> Very poor, indeed.  You should have seen how fast the NetM2 project
> fizzled -- it was as if a group of people has lots of input on how to
> use a piece of flint and steel -- but few had any real hands-on
> experience building a fire that way.  Unfortunately, as soon as it
> was apparant that no one really knew if they wanted a big fire or a
> little fire, everyone turned away.

What about Linux? It is also net-coordinated effort, is it not?

Quote:
> One drawback to using usenet as a means of following a leader is the
> (unspoken) ban on advertising.  It would be somewhat unethical for a
> vendor (like ModulaWare or OUS or me) to come along and tell everyone
> to follow their flag.

Did you ever look at comp.os.linux.announce? There are strict rules
about advertising there, but people do advertise.

Quote:
> However, the converse is not true: the public can adopt a vendor and
> guide the vendor (provided its ears are tuned correctly) in the
> direction the public wants to go.  So, rather than being reactive,
> the public ought to be outlining what they want from the vendors
> (within reason -- in a market as small as Oberon, don't expect one
> vendor to adopt the conventions of another; don't expect emulation of
> OLE2, etc., etc.)  Any input?

Not from me for you, unless you are willing to work on Alpha port :-)
The point is, that you are probably talking to wrong audience in this
group. We are all Internet-based, workstation-based, or VMS-based (myself).
You are targeting the hobbyist PC market. Not many hobbyists here.

Quote:
>>Sound familiar? The question to be answered is: How will oberon help me
>>do what I do better, and at what cost?

> Pure and simple: tripe, garbage, baloney.  This argument about Oberon
> is as old as the newsgroup.  If you want these things, take some time
> to write them.  

Taylor, please! Be reasonable. You are a professional programmer,
who wants to make your living by programming computers. You are talking
to a Ph.D student of completely different specialty. Please, do not
blame him for not writing his own compiler. I think his point is a very
good one. You are the provider, he is the user. Nothing wrong about
the user who wants to be a user. (Maybe you should discuss a purchase
order instead?)

I had a similar discussion with one of ETH developers, who told me
"why do not you write your own VMS port". Why, indeed? Because I am
a nuclear physicist. I can see, why he was able to write his port.
He is a CS scientist, working in a CS institute. But proposing that
everybody else writes compilers? It is a bit too much. (You would not
expect my compiler to compile, would you?)

Quote:
> If they [the things] existed, would you drop all that you have now
> and only use Oberon?  Have you done anything to get _those_ vendors
> to support the Oberon system? (No? then why are you {*filter*}ing about
> it?)  Have you personally done anything to help the Oberon community
> move to having these types of applications?  If you want to be at the
> leading edge of a new system, with no real commercial support, then
> you must be willing to sacrifice everything which most people call
> 'personal life'.

We in science have sacrificed our personal lives _once_, to do our
particular flavour of science. This is true of scientists and Ph.D
students worldwide, because such is the reality of doing science.
Having done it once, we have not much more to sacrifice. This is why
we say, what he is is saying.

The only difference between us and (say) ETH scientists, is that
they sacrificed their personal lives to do Oberon, and they can say
to others "why do not _you_ do the same" (see above about VMS port).
It is as unreasonable, as for example myself saying "Taylor, why do not
you take a few owl shifts at the accelerator?". If I take owl shifts,
it does not mean everybody else has to.

Quote:
>>Time, money, and risk are involved in writing code, and at this time,
>>I do not see that Oberon is competitive with the computing environment
>>I have outlined above. Not if I have to recreate the above utility myself.
>>Furthermore, the Oberon community does not seem sufficiently large to
>>produce a rich, attractive variety of useful applications.

> I'm sick and tired of this type of crap.  Why does everyone have the
> agenda that Oberon must somehow compete on the level of large,
> multimillion dollar corporations?  

The paragrapf marked ">>" above is from Bob's posting. He never talked
about multimillion corporations. He is an astronomer sitting at the top
of his mountain. He wants a few useful graphical modules to built upon.
Oberon cannot offer him even this.

Taylor, please check to what exactly you are responding.

Quote:
> If you are not willing to contribute anything constructive, why
> bother contributing?

I would propose: let us stop throwing accusations at each other, let us
try to find an answer to the basic question: if it is so good, then
how to make it more useful than it is now?

Quote:

> Taylor Hutt, a grumpy kind of guy

Wojtek


Fri, 21 Nov 1997 03:00:00 GMT  
 Oberon-Growth-Limit was Re: Oberon: beginners questions

Quote:


>> While I'm on the topic of compatibility, I'll toss out my pet peeve
>> about V4: it's Text model is too high level.  Texts, IMHO, should be
>> derived from a lower-level text abstraction.  The reason for this
>> complaint is that some of the Oberon.Mod messages have a text field
>> in them - and if you are writing a different text model (not derived
>> from Texts), it is more difficult (and less efficient) to implement
>> some of the messages (like CopyOverMsg).

>Do you think that System-3 has the answer? Or do you want to suggest
>some entirely different solution of your own ?

In the sense that most (if not all) things are derived from
Objects.Object in System3, it presents a better solution.

The solution for V4 could be visualized like so:

DEFINITION BaseTexts;

  TYPE
    Text = POINTER TO TextDesc;
    TextDesc = RECORD END;

(* more stuff possibly *)
END BaseTexts;

DEFINITION Texts;
IMPORT BaseTexts;

  TYPE
    Text = POINTER TO TextDesc;
    TextDesc = RECORD (BaseTexts.TextDesc) (* more *) END;
END Texts.

Then, all current references to Texts.Texts in the current implementation
of Oberon (generally speaking, the messages) would be changed to
BaseTexts.Text.

All references to variables of type BaseTexts.Text should be guarded with
a type test to see if they are of the proper Texts.Text type -- to
preserve the current behavior.

But, I do not advocate this change at all (contemplated, but not advocated).

Taylor Hutt, contemplative
Contemplative is not the same as constipated.



Fri, 21 Nov 1997 03:00:00 GMT  
 Oberon-Growth-Limit was Re: Oberon: beginners questions

Quote:

> While I'm on the topic of compatibility, I'll toss out my pet peeve
> about V4: it's Text model is too high level.  Texts, IMHO, should be
> derived from a lower-level text abstraction.  The reason for this
> complaint is that some of the Oberon.Mod messages have a text field
> in them - and if you are writing a different text model (not derived
> from Texts), it is more difficult (and less efficient) to implement
> some of the messages (like CopyOverMsg).

Do you think that System-3 has the answer? Or do you want to suggest
some entirely different solution of your own ?

Wojtek



Fri, 21 Nov 1997 03:00:00 GMT  
 Oberon-Growth-Limit was Re: Oberon: beginners questions
Sorry to jump in so much today...


Quote:

>If only ETH could also
>devote some thought to the issue, we might be all set. They took water
>in their mouth, though.

Huh?

Quote:
>>>Sound familiar? The question to be answered is: How will oberon help me
>>>do what I do better, and at what cost?

Why does Oberon have an _obligation_ to do this?  It is a workbench,
and if someone gives you a nice workbench for free, you don't walk over to
it and say: well, the thing I need made isn't sitting on this workbench
already waiting for me...

---

This is a good point:

Quote:
>We in science have sacrificed our personal lives _once_, to do our
>particular flavour of science. This is true of scientists and Ph.D
>students worldwide, because such is the reality of doing science.
>Having done it once, we have not much more to sacrifice.

I understand this, and understand that you can't be expected to
write compilers and equation typesetters and so forth. However,
_this_ is nonsense:

Quote:
>It is as unreasonable, as for example myself saying "Taylor, why do not
>you take a few owl shifts at the accelerator?". If I take owl shifts,
>it does not mean everybody else has to.

Taylor (at least inasmuch as I have read here) is not publically
complaining about the way you run your accelerator, or demanding that
you run experiments not on your schedule that he might be curious about.
What happens when (and if) Taylor says:

Taylor: "Hey, I want to know what happens if you fling highly charged
         subatomic particles through Swiss-coffee-marinated beets? Why
         aren't you running this experiment?  It's not reasonable to
         ask _me_ to do it, is it? _I_'m not a nuclear physicist!"

And IMHO, the scientific payoff for the above experiment roughly matches
that which ETH would get from a VMS port... :-)

Quote:
>[Bob] is an astronomer sitting at the top
>of his mountain. He wants a few useful graphical modules to built upon.
>Oberon cannot offer him even this.

Sheesh. I've built I-don't-know-how-many graphical gizmos on top
of the Oberon primitives.  And they work on a half-dozen platforms,
which is a pretty good payback for the time invested.

But I think a request for the graphical modules is a reasonable one.
Extrapolating from the lack of Bob's modules to Oberon-is-a-failure
is kind of risky, IMHO.

Quote:
>If it is so good, then how to make it more useful than it is now?

Good question. I'd like to see more math and IP tools, and I'm
writing some now that won't be encumbered by corporate policy so
I can release them...

Quote:
>Wojtek

-greg
---
Greg DeLozier/Senior Computer Scientist, Aristar SDI


Sat, 22 Nov 1997 03:00:00 GMT  
 
 [ 9 post ] 

 Relevant Pages 

1. Frage Oberon 3 / Question Oberon 3

2. Oberon growth limitations (was: tracks please)

3. Oberon growth limitations

4. Oberon growth limitations (was tracks

5. New Books on Oberon (Short summary: Three Questions on Oberon)

6. Beginner's question on Oberon

7. Beginner's questions concerning Oberon V4 from J.Templ-CD

8. Oberon: beginners questions

9. I am excited about Oberon again!!!!!

10. CD-Oberon - Oberon/F

11. CD-Oberon Oberon 3 Printer Problem

12. Oberon System 3 / Native Oberon projects

 

 
Powered by phpBB® Forum Software