Re: Programming?

[awilliam@xxxxxxxxxxxxx]

>>>>>BASIC is out of it;
>>>>I'm not convinced of this.  I've seen X-and some other VB like
>>>>>basics for Linux, and while I never really tried them, they might 
be
>>>>>perfect for low-end applications.  In this category I might also 
include
>>>>>Pascal (and certainly Delphi), as I think the demise of Pascal is 
greatly
>>>>>exaggerated.
>>>I seem to recall that this got started with a request for languages 
suitable
>>>for teaching.  IMO VB and cousins hide too much of the machinery to 
be really
>>>useful instructive tools, at least when it comes to a lot of CS 
topics, and
>>>disciplines like algorthms, data structures, etc. This is much less 
true of
>>>Pascal, but it is not "modern" (O-O) enough. For a first language, 
not a
>>>bad choice.  I'll pass on Delphi, since I simply don't know enough
>>about it.
>>Delphi Pascal is an update the Pascal by Borland to their Turbo 
Pascal
>>(which was OO, at least as much as C++), to be even more OO, include 
a
>>set of "toolkits" and I think some database (ODBC) libraries.  It is
>>closely affiliated with Borland's RAD package, but a free version of
>>Delphi Pascal by a GNU group is available.
>Doesn't sound bad, but it also doesn't sound like a "first language". 
It is
>**easy** to be more OO than C++

True, it's really intended for developing pieces of a business system, 
 and not really suited to "home" desktop apps.

>>>>>I suggest we NOT talk about real dinosaur languages such as 
ALGOL,
>>>>>COBOL,...
>>>>I wouldn't call COBOL a dinosaur,  it will probably still be 
around
>>>when I'm in about the same shape as the dinosaurs.
>>>There's still more COBOL around than everything else put together.
>>COBOL never got popular on UNIX for a number of reasons....
>>Eh?  I've got hundreds of thousands of lines of COBOL on my UNIX
>>server's at work.  And know lots of other unfortunate chaps you do 
as
>>well.
>Perhaps, but I wonder how many of those lines were originally written 
for
>UNIX? I'd put my money (a paltry sum) on all of this code (except a 
few lines
>modified for purposes of interfaces) having been written on some Big 
Iron.

No, these were written for multi-processor Turbo DOS Z-80 machines 
manufactured by DEC.  Each terminal hooked to a card with a Z80 chip 
and 64k of memory.  These card sat on a bus with a master controller 
(80186, 1Mb) that connected to the hard disks, and tape drive.  Also 
sported an Arcnet card.  It was then ported to Xenix, and then to AIX 
on the IBM RT series.  Then to the RS/AIX, Intel/SCO, and Alpha/DGUX.

>How many applications have been written on any UNIX in COBOL? I mean 
on purpose,
>where there was nothing like a network effect that madfe it a 
requirement??
>...
>>>>I think C is very often NOT the best approach to a problem,  but 
it
>>>>certainly is the safest.
>>>Eh? Unless you mean politically, ya lost me.
>>Yes I mean politically.  And it is doubtful their will ever be a
>>platform in the forseeable future that does not have a maintained C
>>compiler.
>Sure, but this is due to the rise of C instead of assemblers as the 
native
>language of the OS, not because it carries some special magic for 
applica-
>tions guys. A lot of the languages that are more abstract that C 
relative
>to the machine (e.g., Pascal, FORTRAN, COBOL, BASIC, PL/1, ADA, 
Eiffel, APL,
>PERL) represent attempts to represent the task of generating streams 
of mach-
>ine instructions in the domain of problem solvers in a reasonably 
compact and
>meaningful way. Each of the above is successful, in that they are 
created with
>different agendas, at different times, by people with differing world 
views.
>>Think if I had written the killer app in Logo, and now
>>wanted to run in on Windows CE,  I suspect I'd be outa luck.
>If any killer apps had been written in LOGO, there might well be 
demand for
>a LOGO run-time (or development system) for WINCE, or any other 
platform you
>can think of. This demand-supply effect weakens these kinds of 
statements.

True if were speaking of a mass-market app,  however "my" killer app 
isn't going to motivate anyone else to port a compiler.