NEXTSTEP : Don't let your company miss it. Don't let your career miss it 
Author Message
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Hi Folks !

Are you ready to see the most beautifull peace of advertis{*filter*}t you've ever  
seen in any newspaper for any product ?

Check InfoWorld November 8, 1993, Volume 15, issue 45 !

NeXT publish a 8 full glass-paper pages of messages for American's CEO/CIO, and  
Information System managers that could change their life for the better.

I've never seen myself a so perfect advertisment. From the quality of paper to  
the clarity of the message. Everything is there ! The respect of all the  
respected standards (from Dos/Windows to MacOS, and NFS, Nowel, C, C++ and  
Objective-C, to Oracle and Sybase, from HP to Dell, from IBM 3270 to AS/400),  
the competitive advantage, the large corporate references, the shrinkapps  
coherence solution. Nothing is missed to illustrate, explain and demonstrate  
their coherente strategy. This advertis{*filter*}t is the message from a winner who  
understand the needs of business corporations, and give them the most  
attractive solution for their present and their futur.

This advertis{*filter*}t shows that NeXT Computer, Inc is the best winner StartUp in  
the computer/software industry on the market.

Buy InfoWorld to discover the most pleasant news at the COMDEX time and show it  
to your manager. Like Mike Adelson (Project Manager, MIS, Chrysler Financial)  
who has decided to implement NEXTSTEP in 100 financial centers across North  
America, like Ingvar Peturson (Chief Information Officer, MacCaw Cellular) who  
has decided to deploy NEXTSTEP on 4.000 users, like Dwight Koop (Executif  
Director Information Technology, Swiss Bank Corportation), who has allready  
installed 1.500 NEXTSTEP users in one of the world's leading optins trading  
companies, take the opportunity to become the adviser of your CEO/CIO or direct  
manager to let them take the competitive advantage of the NeXT generation, and  
give youself a chance to get, the quicker as possible, the same promotion that  
those engineers who at Swiss Bank, MacCaw Cellular, Chrysler Financial and lots  
others, got by choosing NEXTSTEP.

Do like PanCanadian Petroleum Limited, which was at 90% down the road to  
standardizing on Windows with PowerBuilder in creating their client/server  
development environment, discover NEXTSTEP and make the switch.

This news at the COMDEX is one of the most important turn in the computer  
science history. Don't let your company miss it. Don't let your career miss it.

Best regards,

        - Eric

---
_____________________________________________________________________
                                |
Eric de la Tribouille           | Advanced Collaborative Systems Lab.

                                | University of Illinois at
                                |               Urbana-Champaign
                                | 1304W. Springfield Avenue
NeXT-Mail very welcome !        | Urbana, ILL 61801, USA
?;?)/
                        TOMORROW's TOOLS TODAY !
_____________________________________________________________________
Opinions expressed are my own.

                "The trouble with our times is
                that the future is not what it  
                used to be."  - Paul Valery



Mon, 29 Apr 1996 02:43:49 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:

>Are you ready to see the most beautifull peace of advertis{*filter*}t you've ever  

                                                   ^^^^^^^^^^^^^^

Quote:
>seen in any newspaper for any product ?

I guess their spelling checker still needs a little work, though. :-)
--



Mon, 29 Apr 1996 05:42:03 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:



> >Are you ready to see the most beautifull peace of advertis{*filter*}t you've ever  
>                                                    ^^^^^^^^^^^^^^
> >seen in any newspaper for any product ?

> I guess their spelling checker still needs a little work, though. :-)
> --


and Heaven protect me from the NextSteppers who can't say "Hello World!"
with less than 1 Gbyte of RAM and a terabyte HD.  :-)

--jvn
--
Julian V. Noble



Tue, 30 Apr 1996 05:47:02 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:



>> >Are you ready to see the most beautifull peace of advertis{*filter*}t you've ever  
>>                                                    ^^^^^^^^^^^^^^
>> >seen in any newspaper for any product ?

>> I guess their spelling checker still needs a little work, though. :-)

  I think the rest of their system needs a little work.

Quote:

>Hi Folks !

>Are you ready to see the most beautifull peace of advertis{*filter*}t you've ever  
>seen in any newspaper for any product ?

  He then tells us how great NeXTSTEP is...

  Well, let me tell you how NeXTSTEP for Intel 3.1 really is.

  Well, it installs easily, and uses a lot of disk space for things
like an on-line Webster dictionary, quotations and Shakespeare plays.
Seems to me it would make more sense to leave the dictionary etc.
on the CD. (BTW, It's easily done with a symbolic link but you
have to delete the dictionary etc. after loading...)
You need at least 300mb. A very good chunk of that is taken up
with the above mentioned dictionary, quotations and plays. I'd rather use
a real dictionary on my desk. Well, to each their own.

  The 3.1 release I was evaluating would not hold a serial connection
long enough to do any uucp transfers. It would hard lockup the
PC forcing a hard reboot which would {*filter*}disk files left right and centre.
3.1 for Intel is useless without the serial port driver update. It should
never have left the factory with a bug that serious. (3.2 for Intel
is available now, and is delivered with the serial driver update)
The uucp system is worse than that available on linux or NetBSD/FreeBSD.
They would be better off putting Taylor uucp on their system.
They are using a lot of the GNU stuff already...

  It would mount a ISO9660 CD, but NFS clients could not use it when
it was exported via NFS. Makes it kind of useless for a small office for
resource sharing doesn't it? We queried NeXTSTEP about it, and was
told that although other unix systems did it (After all there _is_
the VFS layer in most modern unix, except for NeXTSTEP) that the
other systems were using a "work around." pfftthh
Yeah right. A "work around."
BTW, They are working on adding it.... They have no VFS layer. wow.

  It is display postscript. No X11 support at all. X11 support
costs you extra. So you can't use code right off the net. You
also can't use any of the X11 servers such as an NCD X term
to add another low cost seat... I know, X11 *sucks* eggs
and we all want dedicated workstations, but it is a consideration.

  To be fair, everything above can be worked around or is
being addressed in a later release. Though paying extra for
X11 support which comes bundled in with other versions of
unix for the same price seems a bit stupid.

  The kicker was when NeXTSTEP wrote back to our complaints and
told us not to compare it to a server. That they envisioned
NeXTSTEP as a development machine for client-server (as the client)
applications perhaps with a NT machine as the server. So, you are
telling me you are using a BSD system (on top of the mach kernel)
that pioneered networking and you can't do networking?
right. Whats the point then of buying a NeXTSTEP so you can
have an easily maintained homogeneous network when you can't
use it as server as well? You might as well go all NT (yech)
or SCO or something else that can do it right. The attitude
from the company *sucked*.

  The desktop is wonderful. The development system is wonderful.
Objective c is wonderful. It _is_ a great system for producing quality
production code.  No argument. But the rest of the system *sucks*
Perhaps for certain niche applications it is the way to go.
But I don't think it will be for very long. I am certain that there
must be other companies doing it _right_, but on X11 or Windoze.
If NeXTSTEP ported their environment to X11 (Probably heresy to them,
and possibly it is too late) and sold the development environment
for use on other platforms that can do unix right....
(I read something about an HP deal. perhaps this is what
they are doing...) But with COSE staring us in the face,
where does that leave NeXTSTEP?

...

Quote:

>This news at the COMDEX is one of the most important turn in the computer  
>science history. Don't let your company miss it. Don't let your career miss it.

  I'd say give it a pass.... Don't let it ruin your career.... That's
_my_ personal opinion.

--

29 Vanson Ave. Nepean On, K2E 6A9       Am I the only one who has trouble
613-225-9920                            thinking up witty sayings for here?



Tue, 30 Apr 1996 22:28:52 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:

>>Are you ready to see the most beautifull peace of advertis{*filter*}t you've ever
>                                                ^^^^^^^^^^^^^^
>>seen in any newspaper for any product ?

>   I guess their spelling checker still needs a little work, though. :-)

        Not if the language is free-form franglais, in which the
original article was written.

--



Wed, 01 May 1996 09:45:52 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:
>    Well, let me tell you how NeXTSTEP for Intel 3.1 really
>    is.

>    Well, it installs easily, and uses a lot of disk space
>  for things like an on-line Webster dictionary,
>  quotations and Shakespeare plays.  
>  ... You need
>  at least 300mb. A very good chunk of that is taken up
>  with the above mentioned dictionary, quotations and plays.

To run NEXTSTEP you need less than 100Mb...not unreasonable with
respect to Unix, NT, or the going price of about $1/Mb. The install
process will allow you to reject literature if you desire. (I did.)
Documentation, too. Another large (and optional) part of the
installation is support for German, French, and Swedish in the OS and
standard apps...a commendable effort.

Quote:
>    The 3.1 release I was evaluating would not hold a serial
>  connection long enough to do any uucp transfers. It
>  would hard lockup the PC forcing a hard reboot which
>  would {*filter*}disk files left right and centre.  3.1 for
>  Intel is useless without the serial port driver update.
>  It should never have left the factory with a bug that
>  serious.

This was indeed a black spot on the initial release. However,
"useless" is out of line. I use a standalone Intel at home and
networked Intels elsewhere. UUCP is a non-issue on the network and an
annoyance at home. Though I crashed dozens of times at home using UUCP
1) I never corrupted a disk and 2) I was able to maintain normal UUCP
activity with the net community.

Quote:
>    It is display postscript. No X11 support at all. X11
>  support costs you extra. So you can't use code right
>  off the net. You also can't use any of the X11 servers
>  such as an NCD X term to add another low cost seat... I
>  know, X11 *sucks* eggs and we all want dedicated
>  workstations, but it is a consideration.

>    To be fair, everything above can be worked around or
>  is being addressed in a later release. Though paying
>  extra for X11 support which comes bundled in with other
>  versions of unix for the same price seems a bit stupid.

I disagree. PostScript is the best and most broadly supported imaging
model out there. Allocating scarce resources to X11 is certainly not
what I want. The X11 approach has always been controversial, and I
believe it owes what success it has to the fact that the UNIX
alternatives are just as bad.

NEXTSTEP, instead, has been working on Distributed Objects as the core
for client/server apps, rather than Distribute Bitmaps. This support
*is* bundled, and is lot more valuable.

Quote:
>    The desktop is wonderful. The development system is
>  wonderful.  Objective c is wonderful. It _is_ a great
>  system for producing quality production code.  No
>  argument. But the rest of the system *sucks* Perhaps for
>  certain niche applications it is the way to go.  But I
>  don't think it will be for very long. I am certain that
>  there must be other companies doing it _right_, but on
>  X11 or Windoze.

Windows, perhaps, but if so, NEXTSTEP will have lots of company as an
extinct OS.

By the way, the HP effort is two-fold...first, to provide distributed
object support between NEXTSTEP and HP-UX for client/server
applications and second, to port NEXTSTEP to the PA-RISC family.
---

Corona Design, Inc.
P.O. Box 51022
Seattle, WA 98115-1022



Wed, 01 May 1996 19:43:50 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

:   The desktop is wonderful. The development system is wonderful.
: Objective c is wonderful. It _is_ a great system for producing quality
: production code.  No argument. But the rest of the system *sucks*

Although "flaming" is not considered appreciable in these newsgroups,
at least your addition "No argument" provokes my remark:
There can never be any quality software produced in C or any of its
derivates. C *sucks*.




Sat, 11 May 1996 18:54:28 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:

> There can never be any quality software produced in C or any of its
> derivates.

        This is silly.

Quote:
> C *sucks*.

        This is true.

Whitney



Sun, 12 May 1996 00:34:24 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:

>:   The desktop is wonderful. The development system is wonderful.
>: Objective c is wonderful. It _is_ a great system for producing quality
>: production code.  No argument. But the rest of the system *sucks*

Your prayers have been answered. NeXT and Sun are putting NeXT Step on Solaris.
NeXT is also making NeXT Step available for other OSs (nothing more specific).
See todays Wall Street Journal.

This is the best thing for NS and NeXT should have been working toward this
a long time ago. (Hindsight and all that...)

--



Sun, 12 May 1996 00:59:47 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it
In this morning's (Nov 23, 93) San Francisco Chronicle business section is
an announcement that Sun is buying 1.5 percent of Next for $10 million.
The suggestion is that Sun will be getting the interface builder technology.
The article also reviewed Next and Hewlett Packard have an agreement; and
that Intel has invested in Next so that the Next software would be ported
to the Intel architecture.

IMHO, Next is dying and these are desperation moves.

Randall Neff



Sun, 12 May 1996 01:34:08 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:


>:   The desktop is wonderful. The development system is wonderful.
>: Objective c is wonderful. It _is_ a great system for producing quality
>: production code.  No argument. But the rest of the system *sucks*
>Although "flaming" is not considered appreciable in these newsgroups,
>at least your addition "No argument" provokes my remark:
>There can never be any quality software produced in C or any of its
>derivates.

This is ridiculous.  Quality software is a function of quality
engineering, not of the computer language used in the implementation.
Reliance on the language to magically "do the right thing" typically
leads to shoddy engineering.

Quote:
>C *sucks*.

Yeah, probably, it's just the alternatives that get me _really_ worried.

FLAMMENWERFER FREI!
Na, wenn das der Stand der Dinge bei den FU-Informatikern ist, dann
gute Nacht!
UND SICHERN...

Marcel



Sun, 12 May 1996 06:19:21 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:


>:   The desktop is wonderful. The development system is wonderful.
>: Objective c is wonderful. It _is_ a great system for producing quality
>: production code.  No argument. But the rest of the system *sucks*
>Although "flaming" is not considered appreciable in these newsgroups,
>at least your addition "No argument" provokes my remark:
>There can never be any quality software produced in C or any of its
>derivates. C *sucks*.

Flaming is indeed not considered appreciable, especially this kind of
flaming. If you think that C is not fit to be used to produce quality
software then you should explain:
1) Why do you think C is not usable
2) What other language you think IS suited to produce quality
software.
3) Why do you think that the reasons under 1) also apply to Objective-C,
which IMHO has enough extra qualities above those of C that a blind
reduction to C, as you make, is not justifiable.

I am interested in your answers, especially those to 2)

Rene'

--

------------------------------------------------------------------------
If life seems jolly rotten, there's something you've forgotten!
And that's to laugh and smile and dance and sing...



Sun, 12 May 1996 17:01:36 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it


:

: >
: > :   The desktop is wonderful. The development system is wonderful.
: > : Objective c is wonderful. It _is_ a great system for producing quality
: > : production code.  No argument. But the rest of the system *sucks*
: >
: > Although "flaming" is not considered appreciable in these newsgroups,
: > at least your addition "No argument" provokes my remark:
: > There can never be any quality software produced in C or any of its
: > derivates. C *sucks*.
: >

:
: FLAME ON!!!!!
:
: Mr. Wolff seems to forget that the quality of one's software depends more
: on one's algorithms, care and intelligence than on one's language.
More, yes, but not only. A good craftsman could probably make a good
piece of work with an old rusty hammer but he/she would also legitimately
consider this beneath his/her professional honour. Moreover, an apprentice
could never learn how to make good work with old rusty tools. The same
holds with programming: as soon as C becomes a generally accepted language
more and more, even fewer people will get the education that enables them
to abstract from the obstacles of C and produce good things in spite of them.

: Mr. Seymour Cray wrote the software for the PP's attached to {*filter*}
: supercomputers during the '70s in assembler. Not one bug was *EVER* found
: and the software did what it was supposed to do.
I would not deny that better software can be produced in assembler than in C.

: I don't personally like C and its derivates, but I have done
: development using them
So have I (actually not from scratch but I worked on software in C).
: and I don't think my software sucks.
Nor does mine but it was an inappropriately big effort to achieve that.

: It probably took me longer to get done, but then, you can't have everything.
No but still much more than usual.
: I suggest that you to go alt.religion with your comments on languages.
I hoped to have made clear by my introduction why I felt compelled to make
my remark. I thought the poster I responded to was pretty religious.

I think the omnipresence of the machine-independent assembly language
we are bothered with and the nuisance I feel whenever it is taken as a
matter of course that people ought to accept C entitle me do be a
little bit polemic about it at times.

: >There can never be any quality software produced in C or any of its
: >derivates.
:
: This is ridiculous.  Quality software is a function of quality
: engineering, not of the computer language used in the implementation.
See rusty hammer comment above.
: Reliance on the language to magically "do the right thing" typically
: leads to shoddy engineering.
I did not say a good language is sufficient for good software. I said
a bad language is (almost) sufficient for bad software.
Bad programmers (e.g. those who take NOT A -> NOT B for A -> B) are
also sufficient for bad software.

: Flaming is indeed not considered appreciable, especially this kind of
: flaming.
See comments about the cause of my remark and about being polemic above.

: If you think that C is not fit to be used to produce quality
: software then you should explain:
: 1) Why do you think C is not usable
There has been much discussion about this in the literature.
I refrain from quoting specific details as that might suggest I
consider these worse than others.
Personal experience as mentioned above (inadequate effort to get
software right around the problems of C) adds to my opinion.

: 2) What other language you think IS suited to produce quality
: software.
: 3) Why do you think that the reasons under 1) also apply to Objective-C,
: which IMHO has enough extra qualities above those of C that a blind
: reduction to C, as you make, is not justifiable.
As with C--, extra qualities do not remove previous deficiencies.

: I am interested in your answers, especially those to 2)
Alright, this is a very good question indeed. I can enumerate a couple
of languages that incorporate good structure or good concepts (Algol68
to begin with, Pascal of course, Ada for "programming in the large",
ML perhaps among the functional languages, Eiffel if object structure
is important). Unfortunately, all of them have their own problems and
deficiencies. Algol68 is virtually unavailable nowadays and also a
little bit complex, Pascal as an educational language may not have
enough "features" for professional use, Turbo-Pascal which adds
important features is proprietary, Ada is the modern PL/I in terms
of complexity, ML is too unfamiliar for most contemporary programmers,
Eiffel is a dogmatic object-oriented language which I did not at all
wish to promote by mentioning it.
Similar criticism can probably be found (and be formulated more
substantially) for other current or recent developments. Does this
mean a language which can do everything just because it is so low-level
should be agreed on and its own really gross problems ignored?
I'd rather think the situation means that a big effort of research still
has to be put into the development of a modern programming language for
the future. Apart from valuable developments in special areas which are
being done this includes design of just a language which meets the
needs of programming on the level of the state of the art (just as the
effort for Ada was meant for but with a more agreeable result). I think
it is a shame that this task is being neglected by the scientific
community probably because it is considered boring or not bearing enough
merits.




Wed, 15 May 1996 00:51:55 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:
(Marcel Weiher) writes:

> This is ridiculous.  Quality software is a function of quality
> engineering, not of the computer language used in the implementation.
> Reliance on the language to magically "do the right thing" typically
> leads to shoddy engineering.

Quality software is a function of quality engineering and quality  
implementation by quality engineers using quality tools. C is not a  
quality tool, it is even degrading the work of a quality engineer, thus  
the overall quality of a product is usually severely damaged by the use  
of C especially under the general time pressure under which today's  
products have to be developed.

Nobody would argue that a bad design but a "good" language would  
automagically lead to a quality product. But the quality of the  
programming language is an important prerequisite to a good software  
product.

--
------------------------------------------------------------------

Brooks's Law:
        Adding manpower to a late software project makes it later
------------------------------------------------------------------



Mon, 13 May 1996 18:30:56 GMT  
 NEXTSTEP : Don't let your company miss it. Don't let your career miss it

Quote:


>: Mr. Wolff seems to forget that the quality of one's software depends more
>: on one's algorithms, care and intelligence than on one's language.
>More, yes, but not only. A good craftsman could probably make a good
>piece of work with an old rusty hammer but he/she would also legitimately

[...rest of rusty hammer analogy deleted...]

I think the rusty hammer is not a good analogy.  I prefer the fast
sports car:  A good driver will be able to go very fast with it,
a beginner will probably get himself killed.  C is certainly not
suitable for beginners, unless they are unusally talented and
disciplined.

[...]

Quote:
>: I don't personally like C and its derivates, but I have done
>: development using them
>So have I (actually not from scratch but I worked on software in C).

So you worked on other people's code in C?  Maybe some PD UNIX
utilities.  Most of those make me puke as well (pardon the
language).  Don't hold the language responsible for people
who cannot use it.  (Gee, I whenever I go faster than 20 km/h
in my Porsche Turbo, it skids.  I guess it must be a really
bad car ).

[...]

Quote:
>I think the omnipresence of the machine-independent assembly language
>we are bothered with and the nuisance I feel whenever it is taken as a
>matter of course that people ought to accept C entitle me do be a
>little bit polemic about it at times.

Actually, around here (comp.lang.objective-c) we do not put up with
C at all, we do most of our work in a language that was derived
largely from Smalltalk and implemented on top of C, so to speak.

To compare Objective-C with C or C++, you must be unaware of
the language, IMHO.

[Your comment, (Thomas Wolff)]

Quote:
>: >There can never be any quality software produced in C or any of its
              ^^^^^
>: >derivates.

[My comment (Marcel) to your comment]

Quote:
>: This is ridiculous.  Quality software is a function of quality
>: engineering, not of the computer language used in the implementation.
>See rusty hammer comment above.

Well, I disagree with the rusty hammer (see fast sports car), but this
is not what you said.  You stated, CATEGORICALLY, that it is impossible
to produce quality software in C and any of its derivative.  I stand by
my opinion that this statement is utterly ludicrous and exposes its
author's ignorance.

Quote:
>: Reliance on the language to magically "do the right thing" typically
>: leads to shoddy engineering.
>I did not say a good language is sufficient for good software. I said
>a bad language is (almost) sufficient for bad software.
>Bad programmers (e.g. those who take NOT A -> NOT B for A -> B) are
>also sufficient for bad software.

Thank you for the little logic lesson. :-)  My comment was not about
logic, but programmer psychology, something that you seem to be
unaware of.
Now how about a vocabulary session in turn.  Please, in 2 sentences
or less, explain the difference betwen "never" and "sometimes".

[Question: what IS a quality tool?]

Quote:
>Alright, this is a very good question indeed. I can enumerate a couple
>of languages that incorporate good structure or good concepts (Algol68
>to begin with, Pascal of course, Ada for "programming in the large",
>ML perhaps among the functional languages, Eiffel if object structure
>is important). Unfortunately, all of them have their own problems and
>deficiencies. Algol68 is virtually unavailable nowadays and also a
>little bit complex, Pascal as an educational language may not have
>enough "features" for professional use, Turbo-Pascal which adds
>important features is proprietary, Ada is the modern PL/I in terms
>of complexity, ML is too unfamiliar for most contemporary programmers,
>Eiffel is a dogmatic object-oriented language which I did not at all
>wish to promote by mentioning it .

Please explain what makes Pascal "better" than ANSI C.  They are
both block structured, procedural languages, except that Pascal
tends to get in your way.  Ada adds (again, for me) packages and
generics, but I much prefer Obj-C classes to achieve encapsulation,
data hiding and modularity.  Ada also has exceptions, but those
can be, and are, implemented in C/Objective-C.  Never did any
work in Algol or PL/I, but people who have tell me I haven't
missed much.

ML:  trying to keep the type-checker/inference engine happy instead
of programming.  Lack of true polymorphism.  Funktors are nice,
but I really prefer FP's funktors.

Eiffel:  well, one thing we agree on.

You forgot:

Prolog (nice declarative semantics, but "sucks" once you
have to start using it operationally, which you almost always do.)
I just had a look at Prolog III (and other CLP languages) and
it looks quite nice.

Smalltalk:  really nice, can do functors just (or almost?) as well
as Functional Languages without being so constraining.  The
environment is kinda big, though.

and of course, Lots of Insanely neSted Parentheses, fortran and COBOL.

Quote:
>Similar criticism can probably be found (and be formulated more
>substantially) for other current or recent developments. Does this
>mean a language which can do everything just because it is so low-level
>should be agreed on and its own really gross problems ignored?

NO, not "agreed on", but I'll use it until something better comes along,
because if there is a defect (and there always is) I CAN FIX IT.
With most of the languages above, if the designer didn't think of it,
you're stuck.  
For example, Eiffel has assertions as a built-in feature (though
I'm not sure wether they are actually compiled), in C someone
wrote an assert Macro, and anyone could do it themselves if they
didn't like it.
Ada has exceptions, again you can implement them in C.
Full runtime dynamic method dispatching can be implemented in
ANSI-C and hidden with Macros.  It's just more efficently done
in the compiler, and you get a nicer syntax.
People have added garbage collectors, though I don't use one
myself.

With Objective-C, you get all that low-level flexibility with
a fairly good OO-Subsytem thown in on top to address modularity
(I like to call it composeability).

[...Big research effort to find/create "perfect" language...]

Well, take some OO-base system with garbage collection, dynamic
and static type checking/inference and delegation as well as
actors (possibly make it a prototype based language).  Make
expressing higher order concepts natural.  It should be a
'literate' system, allowing me to produce documentation and
code at the same time, in a WYSIWYG environment, possibly
even with graphical notation.

Add the ability to state almost arbitrary constraints, either
numeric or symbolic.

In short, let me express what I want to get done, and minimize
what I have to tell the machine in order to get it done.  On the
other hand, if I'm not happy with how the machine gets it done,
let me tell it.  Easily.

Make it a dynamically compiled environment that is, nonetheles,
capable of producing stand-alone programs.

Make it available on all computing platforms (or just on mine...)

In the meantime, I'll stick with Objective-C and play with other
things.




Wed, 15 May 1996 20:28:58 GMT  
 
 [ 42 post ]  Go to page: [1] [2] [3]

 Relevant Pages 

1. NEXTSTEP : Don't let your company miss it. Don't let your career miss it

2. what windows don't let to do

3. On Windows, don't let the mouse timer interfere with a size/move

4. FS/A: Cool rare SMALLTALK stuff, don't miss out, NeXT and BeOS stuff too

5. Don't miss it!

6. ESUG98: Don't miss this unique event !

7. Developers Graphics Special Ending Soon...don't miss out

8. Last day...don't miss it!

 

 
Powered by phpBB® Forum Software