Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #32971 > unrolled thread

Java primitive variable objects

Back to article view | Back to comp.lang.java.programmer

Contents

  Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-02 10:11 -0800
    Re: Java primitive variable objects Eric Sosman <esosman@comcast-dot-net.invalid> - 2015-11-02 14:15 -0500
      Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-02 13:03 -0800
        Re: Java primitive variable objects Eric Sosman <esosman@comcast-dot-net.invalid> - 2015-11-02 16:29 -0500
    Re: Java primitive variable objects Jan Burse <janburse@fastmail.fm> - 2015-11-02 22:46 +0100
      Re: Java primitive variable objects Jan Burse <janburse@fastmail.fm> - 2015-11-02 22:52 +0100
      Re: Java primitive variable objects Eric Sosman <esosman@comcast-dot-net.invalid> - 2015-11-02 17:47 -0500
    Re: Java primitive variable objects Joerg Meier <joergmmeier@arcor.de> - 2015-11-05 22:19 +0100
      Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-05 13:31 -0800
        Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-07 22:21 -0500
          Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-09 04:52 -0800
            Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 08:54 -0500
              Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-09 07:27 -0800
                Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 10:51 -0500
                  Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-09 08:28 -0800
                    Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 11:57 -0500
                      Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-09 10:13 -0800
                      Re: Java primitive variable objects Eric Sosman <esosman@comcast-dot-net.invalid> - 2015-11-09 13:28 -0500
                        Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 14:53 -0500
                        Re: Java primitive variable objects Nigel Wade <nmw@ion.le.ac.uk> - 2015-11-10 09:27 +0000
                      Re: Java primitive variable objects "Jeffrey H. Coffield" <jeffrey@digitalsynergyinc.com> - 2015-11-09 10:44 -0800
                      Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-09 18:59 +0000
                        Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 15:02 -0500
                          Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-09 22:57 +0000
                            Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 19:16 -0500
                      Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-09 20:55 -0500
                        Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 22:13 -0500
                  Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-09 20:40 -0500
                    Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 22:23 -0500
            Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-09 14:00 +0000
              Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-09 07:18 -0800
                Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-09 20:43 -0500
            Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-09 19:38 -0500
              Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 20:26 -0500
                Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-10 02:10 +0000
                  Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-09 22:12 -0500
                    Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-10 13:01 +0000
                      Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-10 12:23 -0500
                        Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-10 10:10 -0800
                          Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-10 22:08 -0500
                        Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-10 20:05 +0000
                          Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-10 13:21 -0800
                            Re: Java primitive variable objects Martin Gregorie <martin@address-in-sig.invalid> - 2015-11-10 23:30 +0000
                            Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-10 20:58 -0500
                            Re: Java primitive variable objects Gene Wirchenko <genew@telus.net> - 2015-11-10 22:30 -0800
                    Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-10 21:58 -0500
                Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-10 21:56 -0500
                  Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-10 22:19 -0500
                    Re: Java primitive variable objects Arne Vajhøj <arne@vajhoej.dk> - 2015-11-10 22:42 -0500
                      Re: Java primitive variable objects Jerry Stuckle <jstucklex@attglobal.net> - 2015-11-10 23:07 -0500
              Re: Java primitive variable objects Eric Douglas <e.d.programmer@gmail.com> - 2015-11-10 06:06 -0800
    Re: Java primitive variable objects Lew <lewbloch@gmail.com> - 2015-11-16 14:47 -0800

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

#33054

On 11/09/2015 08:57 AM, Jerry Stuckle wrote:

> You never used FORTRAN on punched cards then :)
>
> Columns 1-6  were numeric for statement number; 7 a continuation column
> (non-blank means the statement continued from the previous card).  8-72
> for the code itself, and 73-80 "free form" - generally used to
> sequentially number the cards for when you dropped the deck :)
>

I wrote a Fortran program in the late 60's that evaluated different
strategies on playing solitaire and the data deck was 52 cards (ace, 
two, etc.). I will never forget the look on the operators face when I 
shuffled the deck and handed it to him to run. He was even more 
surprised when the program ran and produced output.

[toc] | [prev] | [next] | [standalone]

#33055

On Mon, 09 Nov 2015 11:57:35 -0500, Jerry Stuckle wrote:

> But you couldn't beat RPG for rapid report generation, even by a
> non-technical user.  When I learned it, there was no SQL or relational
> databases, no MS Access or similar.  Without RPG the manager would have
> to find a programmer (probably COBOL at the time) who would need several
> days to create the same report the manager could in a couple of hours.
> But all it was good for was generating reports (hence the name "Report
> Generating Program").  You could fetch some data, filter it and spit it
> out in a pretty column format, but not much else.
>
Unfortunately, by the time I met it on an AS/400, it had become RPG3 and 
spatchcocked/extended/twisted (choose one) so it could be a universal 
language. We used it to drive fixed format 80x24 green screens (actually 
IBM 5250 emulators running on PCs - doing the sort of things you do with 
a 4CL, e.g Scuptor,dBase,Clipper, or PowerBuilder. 

I much prefer Sculptor, etc because they're designed for screen handling 
and usually have a separate report writer. RPG3 wasn't nice or fast to 
use when used for screen handling because you had to disable the file 
reading loop and cobble up your own screen handling logic - and even then 
the screen layout was defined in a separate file. Fortunately, we were 
able to use PL/1 for the main logic in this application.


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#33058

On 11/9/2015 1:59 PM, Martin Gregorie wrote:
> On Mon, 09 Nov 2015 11:57:35 -0500, Jerry Stuckle wrote:
> 
>> But you couldn't beat RPG for rapid report generation, even by a
>> non-technical user.  When I learned it, there was no SQL or relational
>> databases, no MS Access or similar.  Without RPG the manager would have
>> to find a programmer (probably COBOL at the time) who would need several
>> days to create the same report the manager could in a couple of hours.
>> But all it was good for was generating reports (hence the name "Report
>> Generating Program").  You could fetch some data, filter it and spit it
>> out in a pretty column format, but not much else.
>>
> Unfortunately, by the time I met it on an AS/400, it had become RPG3 and 
> spatchcocked/extended/twisted (choose one) so it could be a universal 
> language. We used it to drive fixed format 80x24 green screens (actually 
> IBM 5250 emulators running on PCs - doing the sort of things you do with 
> a 4CL, e.g Scuptor,dBase,Clipper, or PowerBuilder. 
> 
> I much prefer Sculptor, etc because they're designed for screen handling 
> and usually have a separate report writer. RPG3 wasn't nice or fast to 
> use when used for screen handling because you had to disable the file 
> reading loop and cobble up your own screen handling logic - and even then 
> the screen layout was defined in a separate file. Fortunately, we were 
> able to use PL/1 for the main logic in this application.
> 
> 

Sounds like I was fortunate in that I got away from it before it became
RPG3 - this was before video displays.  The operator interface was
pretty much punch cards and printouts; you could get remote access via
teletype machines (some even had card readers!) but they were slow - 300
baud or less.

And PL/1 - I remember it fondly.  But that was in the early 80's, long
after I was dealing with RPG.

-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33059

On Mon, 09 Nov 2015 15:02:09 -0500, Jerry Stuckle wrote:

> And PL/1 - I remember it fondly.  But that was in the early 80's, long
> after I was dealing with RPG.
>
I've used it on two projects  - initially on an IBM S/88. This was re-
badged Stratus fault tolerant box using their own near-UNIX OS and 
(probably) PL/1 subset G. The second time was regular IBM PL/1 on an 
AS/400.

I thought PL/1 was OK, but didn't much like the way it used exception 
handlers. They were fine for trapping errors, but I thought that also 
using them for normal events like end-of-file just disrupted the program 
flow by separating the error trap from the loop where the end-of-file 
occurred. I thought then and still think that COBOL did that better with 
its 'PERFORM ... UNTIL ..' construct, particularly now that it allows the 
controlled statements to be embedded in the PERFORM sentence.
 

-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#33060

On 11/9/2015 5:57 PM, Martin Gregorie wrote:
> On Mon, 09 Nov 2015 15:02:09 -0500, Jerry Stuckle wrote:
> 
>> And PL/1 - I remember it fondly.  But that was in the early 80's, long
>> after I was dealing with RPG.
>>
> I've used it on two projects  - initially on an IBM S/88. This was re-
> badged Stratus fault tolerant box using their own near-UNIX OS and 
> (probably) PL/1 subset G. The second time was regular IBM PL/1 on an 
> AS/400.
> 
> I thought PL/1 was OK, but didn't much like the way it used exception 
> handlers. They were fine for trapping errors, but I thought that also 
> using them for normal events like end-of-file just disrupted the program 
> flow by separating the error trap from the loop where the end-of-file 
> occurred. I thought then and still think that COBOL did that better with 
> its 'PERFORM ... UNTIL ..' construct, particularly now that it allows the 
> controlled statements to be embedded in the PERFORM sentence.
>  
> 

Probably because exceptions weren't made for handling normal events,
just like in C++ and Java.  They were designed to handle abnormal
events.  However, they were a rather new concept, and misused more than
they were correctly used.

I agree, the COBOL PERFORM ... UNTIL .. construct was better.  But then
COBOL was better than most languages at the time for handling file input.

-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33065

On 11/9/2015 11:57 AM, Jerry Stuckle wrote:
> On 11/9/2015 11:28 AM, Eric Douglas wrote:
>> I learned about RPG in school, never had to use it for work.  The
>> language was so ugly, and the teacher for it told war stories
>> instead of teaching, that no one bothered to learn it.  They just
>> copied the labs from someone who had the class before.  I can't
>> imagine a worse 'language' than one that tells you what characters
>> you can type and hard codes columns to type them in.
>>
>
> You never used FORTRAN on punched cards then :)
>
> Columns 1-6  were numeric for statement number; 7 a continuation
> column (non-blank means the statement continued from the previous
> card).  8-72 for the code itself,

As Eric already pointed out then it is 1-5, 6 and 7-72.

>                                 and 73-80 "free form" - generally
> used to sequentially number the cards for when you dropped the deck
> :)

73-80 was just ignored.

But please don't call it free form, which has a completely different
meaning in Fortran (the new source format introduced with Fortran 90).

Arne

[toc] | [prev] | [next] | [standalone]

#33068

On 11/9/2015 8:55 PM, Arne Vajhøj wrote:
> On 11/9/2015 11:57 AM, Jerry Stuckle wrote:
>> On 11/9/2015 11:28 AM, Eric Douglas wrote:
>>> I learned about RPG in school, never had to use it for work.  The
>>> language was so ugly, and the teacher for it told war stories
>>> instead of teaching, that no one bothered to learn it.  They just
>>> copied the labs from someone who had the class before.  I can't
>>> imagine a worse 'language' than one that tells you what characters
>>> you can type and hard codes columns to type them in.
>>>
>>
>> You never used FORTRAN on punched cards then :)
>>
>> Columns 1-6  were numeric for statement number; 7 a continuation
>> column (non-blank means the statement continued from the previous
>> card).  8-72 for the code itself,
> 
> As Eric already pointed out then it is 1-5, 6 and 7-72.
> 
>>                                 and 73-80 "free form" - generally
>> used to sequentially number the cards for when you dropped the deck
>> :)
> 
> 73-80 was just ignored.
> 
> But please don't call it free form, which has a completely different
> meaning in Fortran (the new source format introduced with Fortran 90).
> 
> Arne
> 
> 
> 

In FORTRAN II, which is what we used, it was known as free form.
FORTRAN 90 wouldn't exist for another 20 years.

-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33063

On 11/9/2015 10:51 AM, Jerry Stuckle wrote:
> On 11/9/2015 10:27 AM, Eric Douglas wrote:
>> On Monday, November 9, 2015 at 8:54:58 AM UTC-5, Jerry Stuckle wrote:
>>> PASCAL isn't used much in business for the same reason it makes a great
>>> learning tool.  It places you in a straightjacket, and results in a
>>> slower development cycle.  OTOH, that same straightjacket teaches good
>>> programming techniques that can be applied to virtually any language.
>>
>> I've never written PASCAL but I've never seen a programming straightjacket language worse than RPG.
>>
>
> Oh, RPG was fun. A simple language that allows simple managers to
> quickly create reports from a database.  Not much for computational
> power, but it didn't need it for what it did.
>
> PASCAL is much worse.  For instance, two variables, each defined as
> char[80] are not considered the same type.

Sure they are.

>                                            You need to define a new
> type such as "char80" then have the two variables of that type.  Now
> they can be compared, assigned, etc.

That works too. And are in general a good idea. But it is not required.

Arne

[toc] | [prev] | [next] | [standalone]

#33069

On 11/9/2015 8:40 PM, Arne Vajhøj wrote:
> On 11/9/2015 10:51 AM, Jerry Stuckle wrote:
>> On 11/9/2015 10:27 AM, Eric Douglas wrote:
>>> On Monday, November 9, 2015 at 8:54:58 AM UTC-5, Jerry Stuckle wrote:
>>>> PASCAL isn't used much in business for the same reason it makes a great
>>>> learning tool.  It places you in a straightjacket, and results in a
>>>> slower development cycle.  OTOH, that same straightjacket teaches good
>>>> programming techniques that can be applied to virtually any language.
>>>
>>> I've never written PASCAL but I've never seen a programming
>>> straightjacket language worse than RPG.
>>>
>>
>> Oh, RPG was fun. A simple language that allows simple managers to
>> quickly create reports from a database.  Not much for computational
>> power, but it didn't need it for what it did.
>>
>> PASCAL is much worse.  For instance, two variables, each defined as
>> char[80] are not considered the same type.
> 
> Sure they are.
> 

Not in a standards compliant compiler they are not.

Example (might have a typo or two - it's been a while).

var
  array1 : array[1..80] of char;
  array2 : array[1..80] of char;

array1 and array2 are considered different types.

type
  line = array[1..80] of char;
var
  array1: line;
  array2: line;

array1 and array2 are now of the same type.

>>                                            You need to define a new
>> type such as "char80" then have the two variables of that type.  Now
>> they can be compared, assigned, etc.
> 
> That works too. And are in general a good idea. But it is not required.
> 

It is in a standards compliant compiler.

> Arne
> 
> 



-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33042

On Mon, 09 Nov 2015 04:52:57 -0800, Eric Douglas wrote:

> On Saturday, November 7, 2015 at 10:21:54 PM UTC-5, Arne Vajhøj wrote:
>> In the 90's you should have been taught Pascal and/or C++ (which both
>> has enums) if the school had been following trends.
>> 
>> Arne
> 
> I've never heard of either of those being used for business
> applications.
> I went to college from 1993 to 1996.  Most companies were using COBOL.
> I did have one intro to C++ course.
> Let me guess, you're one of those people, who gets a BS in CS for a
> career in CIS.

Where was your college?

I'm asking because, back in the mid '70s I was working with a very good, 
though fairly inexperienced, programmer/designer in NYC. He'd picked up a 
BSc in programming at one of the NY colleges and had been taught to 
program in COBOL: that was pretty much the entire course. This surprised 
me because I knew that in the UK a computing BSc covered a lot of stuff 
besides programming, such as writing operating systems and compilers as 
well as quite a lot of application systems design. If you were at a US 
college, all I can say is that not much seems to have moved on there 
between the mid 70s and the mid 90s.

Me? I'd learnt to program in Algol 60 at University, as part of a 
chemistry degree. After graduation I ended up at ICL where I was 
initially taught PLAN assembler and application design followed a year or 
so later by COBOL. Meanwhile I was teaching myself to operate a 1900 and 
to be sysadmin on the then-current OSen (George 1,2, and 3). 

But the main skill I learnt at ISC was how to handle the "We need 
somebody who knows XXXX. You once saw the manual? The very man! You're on 
the project next week." situation. That's a skill that's served me well 
ever since.


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#33046

On Monday, November 9, 2015 at 9:03:02 AM UTC-5, Martin Gregorie wrote:
> Where was your college?

Ohio
They change their curriculum every year to match what area businesses say they're looking for.

[toc] | [prev] | [next] | [standalone]

#33064

On 11/9/2015 10:18 AM, Eric Douglas wrote:
> On Monday, November 9, 2015 at 9:03:02 AM UTC-5, Martin Gregorie wrote:
>> Where was your college?
>
> Ohio
> They change their curriculum every year to match what area businesses say they're looking for.

Not good. That is chasing yesterdays demand.

Arne

[toc] | [prev] | [next] | [standalone]

#33061

On 11/9/2015 7:52 AM, Eric Douglas wrote:
> On Saturday, November 7, 2015 at 10:21:54 PM UTC-5, Arne Vajhøj wrote:
>> In the 90's you should have been taught Pascal and/or C++ (which
>> both has enums) if the school had been following trends.
>
> I've never heard of either of those being used for business applications.
> I went to college from 1993 to 1996.  Most companies were using COBOL.
> I did have one intro to C++ course.

Universities are supposed to teach students something that will help
them through a career of 40-45 years.

At that time I think it both made sense and was common to use
Pascal or C++.

Even though COBOL was still very popular at that time, then it
was obvious that it would not stay that way.

C and C++ was already making serious inroads into business. And
everybody knew that OO would be the big thing - it was just
not clear what languages it would be besides C++.

> Let me guess, you're one of those people, who gets a BS in CS for a career in CIS.

Not quite.

But I probably should have taken a MSc in CS instead of what I did.

Arne

[toc] | [prev] | [next] | [standalone]

#33062

On 11/9/2015 7:38 PM, Arne Vajhøj wrote:
> On 11/9/2015 7:52 AM, Eric Douglas wrote:
>> On Saturday, November 7, 2015 at 10:21:54 PM UTC-5, Arne Vajhøj wrote:
>>> In the 90's you should have been taught Pascal and/or C++ (which
>>> both has enums) if the school had been following trends.
>>
>> I've never heard of either of those being used for business applications.
>> I went to college from 1993 to 1996.  Most companies were using COBOL.
>> I did have one intro to C++ course.
> 
> Universities are supposed to teach students something that will help
> them through a career of 40-45 years.
> 

That's hard to do in any of the sciences today.  The world is changing
too quickly.

> At that time I think it both made sense and was common to use
> Pascal or C++.
> 

PASCAL has never been big in business.  C++ is to some extent, but still
not as popular as C.

> Even though COBOL was still very popular at that time, then it
> was obvious that it would not stay that way.
>

It is still the most commonly used language in the world.

> C and C++ was already making serious inroads into business. And
> everybody knew that OO would be the big thing - it was just
> not clear what languages it would be besides C++.
> 

Not in the financial sector, it isn't - which includes banks, insurance
companies, governments, and many, many more.  It still can't beat
COBOL's packed decimal datatype for numeric operations.  And OO is good
for a lot of things, but it's definitely not the answer to everything
programming.

>> Let me guess, you're one of those people, who gets a BS in CS for a
>> career in CIS.
> 
> Not quite.
> 
> But I probably should have taken a MSc in CS instead of what I did.
> 
> Arne
> 
> 



-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33066

On Mon, 09 Nov 2015 20:26:53 -0500, Jerry Stuckle wrote:

> On 11/9/2015 7:38 PM, Arne Vajhøj wrote:
>> Universities are supposed to teach students something that will help
>> them through a career of 40-45 years.
>> 
>> 
> That's hard to do in any of the sciences today.  The world is changing
> too quickly.
>
Yes, but its what a *University* should be doing - and apparently a 
number of US State universities weren't doing it in the 70s and haven't 
changed their ways 20 years later, as confirmed by ED. The question is: 
have they changed since then.

In NZ, where I got my degree, and in the UK computer science degrees have 
certainly taught programming languages, but as tools that let students 
get hands-on experience of the non language-specific skills that the 
degree is meant to teach them. Things such as the elements of hardware 
and software architecture, systems analysis principles (not how to use 
SSADM, Bootch or Rational Rose), modular program decomposition (not how 
to look up design patterns). IOW, they're teaching general principles and 
problem solving - skills which *can* serve you well throughout a career.

I'd call any school that simply taught coding in a programming language a 
polytech: its certainly not a University, at least not the sort of 
University I attended.
 
>> At that time I think it both made sense and was common to use Pascal or
>> C++.
>>
Yes, Pascal has always been a popular teaching language. In addition to 
the reasons already given, algorithms expressed in it are very easy to 
read by anybody who knows any block-structured language. Take a look at 
Sedgewick: "Algorithms" - a book that belongs in any serious programmer's 
library. It uses Pascal and I know, from personal experience that I can 
transcribe its algorithms directly into C. It should be equally easy to 
transcribe its Pascal code into Java, C++ or Python.

> PASCAL has never been big in business.  C++ is to some extent, but still
> not as popular as C.
>
I think it depends what you're doing, though I agree about Pascal. If 
you're writing, say, business domain-specific message switches, ATM or 
EFTPOS network controllers or almost any very high volume application 
you'll probably choose C with Java and C++ in 2nd rank due to their 
generally larger support libraries and related overheads.
   
>> Even though COBOLwas still very popular at that time, then it was
>> obvious that it would not stay that way.
>>
>>
> It is still the most commonly used language in the world.
>
And, within the application domain it was designed for, it may still be 
the best. As I've said before, how many lines would it take in any other 
language to format a financial in the way described by 

 77 MY-VALUE      COMP SYNC RIGHT PIC $$,$$$,$$0.99DB BLANK WHEN ZERO.


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#33067

On 11/9/2015 9:10 PM, Martin Gregorie wrote:
> On Mon, 09 Nov 2015 20:26:53 -0500, Jerry Stuckle wrote:
> 
>> On 11/9/2015 7:38 PM, Arne Vajhøj wrote:
>>> Universities are supposed to teach students something that will help
>>> them through a career of 40-45 years.
>>>
>>>
>> That's hard to do in any of the sciences today.  The world is changing
>> too quickly.
>>
> Yes, but its what a *University* should be doing - and apparently a 
> number of US State universities weren't doing it in the 70s and haven't 
> changed their ways 20 years later, as confirmed by ED. The question is: 
> have they changed since then.
>

It's hard for a university to teach students something that won't even
exist for 20 years.  For instance, when I was in college, there was no
C, much less C++.  So how was my university supposed to teach it?

> In NZ, where I got my degree, and in the UK computer science degrees have 
> certainly taught programming languages, but as tools that let students 
> get hands-on experience of the non language-specific skills that the 
> degree is meant to teach them. Things such as the elements of hardware 
> and software architecture, systems analysis principles (not how to use 
> SSADM, Bootch or Rational Rose), modular program decomposition (not how 
> to look up design patterns). IOW, they're teaching general principles and 
> problem solving - skills which *can* serve you well throughout a career.
> 

We learned that, also.  But we didn't learn object oriented programming,
for instance - another concept which didn't exist at the time.  The same
with relational databases, SQL and any number of concepts we use
regularly today.

> I'd call any school that simply taught coding in a programming language a 
> polytech: its certainly not a University, at least not the sort of 
> University I attended.
>  

It wasn't the type I attended, either.

>>> At that time I think it both made sense and was common to use Pascal or
>>> C++.
>>>
> Yes, Pascal has always been a popular teaching language. In addition to 
> the reasons already given, algorithms expressed in it are very easy to 
> read by anybody who knows any block-structured language. Take a look at 
> Sedgewick: "Algorithms" - a book that belongs in any serious programmer's 
> library. It uses Pascal and I know, from personal experience that I can 
> transcribe its algorithms directly into C. It should be equally easy to 
> transcribe its Pascal code into Java, C++ or Python.
> 
>> PASCAL has never been big in business.  C++ is to some extent, but still
>> not as popular as C.
>>
> I think it depends what you're doing, though I agree about Pascal. If 
> you're writing, say, business domain-specific message switches, ATM or 
> EFTPOS network controllers or almost any very high volume application 
> you'll probably choose C with Java and C++ in 2nd rank due to their 
> generally larger support libraries and related overheads.
>    
>>> Even though COBOLwas still very popular at that time, then it was
>>> obvious that it would not stay that way.
>>>
>>>
>> It is still the most commonly used language in the world.
>>
> And, within the application domain it was designed for, it may still be 
> the best. As I've said before, how many lines would it take in any other 
> language to format a financial in the way described by 
> 
>  77 MY-VALUE      COMP SYNC RIGHT PIC $$,$$$,$$0.99DB BLANK WHEN ZERO.
> 
> 

Or, take it 4 more places to each the left and right, yet  maintain
accuracy.


-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33073

On Mon, 09 Nov 2015 22:12:40 -0500, Jerry Stuckle wrote:

> On 11/9/2015 9:10 PM, Martin Gregorie wrote:
>> Yes, but its what a *University* should be doing - and apparently a
>> number of US State universities weren't doing it in the 70s and haven't
>> changed their ways 20 years later, as confirmed by ED. The question is:
>> have they changed since then.
>>
>>
> It's hard for a university to teach students something that won't even
> exist for 20 years.  
Yes, of course, but what I was trying to say is that a University should 
be teaching their students to learn, not just pounding facts into their 
heads with rote learning.

> For instance, when I was in college, there was no
> C, much less C++.  So how was my university supposed to teach it?
>
They weren't, but if they did turned out properly educated graduates, 
they'd be equipped to learn new languages by themselves.

For instance, nobody ever formally taught me to write TAL, the language 
Tandem wrote the Guardian OS and its utilities in, but I quickly spotted 
its similarity to PL/9 (a derivative of PL/M) which I'd taught myself and 
was fluent in, so I quickly picked up enough TAL to write a module that 
transferred messages between SNA LU.2 sessions and TCP/IP sessions. We 
needed to make an AS/400 financial application talk to banking package on 
an HP UNIX box via a Tandem NonStop box. 
 
>> In NZ, where I got my degree, and in the UK computer science degrees
>> have certainly taught programming languages, but as tools that let
>> students get hands-on experience of the non language-specific skills
>> that the degree is meant to teach them. Things such as the elements of
>> hardware and software architecture, systems analysis principles (not
>> how to use SSADM, Bootch or Rational Rose), modular program
>> decomposition (not how to look up design patterns). IOW, they're
>> teaching general principles and problem solving - skills which *can*
>> serve you well throughout a career.
>> 
>> 
> We learned that, also.  But we didn't learn object oriented programming,
> for instance - another concept which didn't exist at the time.  The same
> with relational databases, SQL and any number of concepts we use
> regularly today.
> 
Same thing - you evidently went to a pukka university which equipped you 
to learn. 

>> I'd call any school that simply taught coding in a programming language
>> a polytech: its certainly not a University, at least not the sort of
>> University I attended.
>>  
>>  
> It wasn't the type I attended, either.
>
Indeed, but it certainly WAS the type of school that my colleague in NYC 
had attended.


-- 
martin@   | Martin Gregorie
gregorie. | Essex, UK
org       |

[toc] | [prev] | [next] | [standalone]

#33079

On 11/10/2015 8:01 AM, Martin Gregorie wrote:
> On Mon, 09 Nov 2015 22:12:40 -0500, Jerry Stuckle wrote:
> 
>> On 11/9/2015 9:10 PM, Martin Gregorie wrote:
>>> Yes, but its what a *University* should be doing - and apparently a
>>> number of US State universities weren't doing it in the 70s and haven't
>>> changed their ways 20 years later, as confirmed by ED. The question is:
>>> have they changed since then.
>>>
>>>
>> It's hard for a university to teach students something that won't even
>> exist for 20 years.  
> Yes, of course, but what I was trying to say is that a University should 
> be teaching their students to learn, not just pounding facts into their 
> heads with rote learning.
>

You can teach a person facts, but you cannot teach them to learn.  That
can only come from one's self desire.

>> For instance, when I was in college, there was no
>> C, much less C++.  So how was my university supposed to teach it?
>>
> They weren't, but if they did turned out properly educated graduates, 
> they'd be equipped to learn new languages by themselves.
> 

Yes and no.  I used to teach in the corporate world (contract
instructor).  Among the classes I taught were Java and C++.  And, of
course, within those courses were OO techniques.  The syntax of the
language was easy for most students to pick up.  OO concepts were much
harder for most students.

> For instance, nobody ever formally taught me to write TAL, the language 
> Tandem wrote the Guardian OS and its utilities in, but I quickly spotted 
> its similarity to PL/9 (a derivative of PL/M) which I'd taught myself and 
> was fluent in, so I quickly picked up enough TAL to write a module that 
> transferred messages between SNA LU.2 sessions and TCP/IP sessions. We 
> needed to make an AS/400 financial application talk to banking package on 
> an HP UNIX box via a Tandem NonStop box. 
>  

But you already had a similar language, so it wasn't much of a jump for you.

>>> In NZ, where I got my degree, and in the UK computer science degrees
>>> have certainly taught programming languages, but as tools that let
>>> students get hands-on experience of the non language-specific skills
>>> that the degree is meant to teach them. Things such as the elements of
>>> hardware and software architecture, systems analysis principles (not
>>> how to use SSADM, Bootch or Rational Rose), modular program
>>> decomposition (not how to look up design patterns). IOW, they're
>>> teaching general principles and problem solving - skills which *can*
>>> serve you well throughout a career.
>>>
>>>
>> We learned that, also.  But we didn't learn object oriented programming,
>> for instance - another concept which didn't exist at the time.  The same
>> with relational databases, SQL and any number of concepts we use
>> regularly today.
>>
> Same thing - you evidently went to a pukka university which equipped you 
> to learn. 
> 

No, *I* equipped myself to learn.  The university just taught me the
sciences of the day.

>>> I'd call any school that simply taught coding in a programming language
>>> a polytech: its certainly not a University, at least not the sort of
>>> University I attended.
>>>  
>>>  
>> It wasn't the type I attended, either.
>>
> Indeed, but it certainly WAS the type of school that my colleague in NYC 
> had attended.
> 
> 

I don't know what college your friend attended, but even in the 70's I
don't know of any universities which only taught COBOL as a CS major.
In my university, engineering students were required to learn FORTRAN
(and not much else CS-wise), and I think (but am not positive) some
business-related studies required COBOL.  But CS required much more.

-- 
==================
Remove the "x" from my email address
Jerry Stuckle
jstucklex@attglobal.net
==================

[toc] | [prev] | [next] | [standalone]

#33080

On Tuesday, November 10, 2015 at 12:23:26 PM UTC-5, Jerry Stuckle wrote:
> I don't know what college your friend attended, but even in the 70's I
> don't know of any universities which only taught COBOL as a CS major.
> In my university, engineering students were required to learn FORTRAN
> (and not much else CS-wise), and I think (but am not positive) some
> business-related studies required COBOL.  But CS required much more.

My brother was a CS major.  He does AI development.  He learned FORTRAN, Pascal, C, all that fun stuff, no COBOL.  They ought to teach COBOL to CS majors, considering how many with CS degrees seek jobs in the CIS field.

[toc] | [prev] | [next] | [standalone]

#33093

On 11/10/2015 1:10 PM, Eric Douglas wrote:
> On Tuesday, November 10, 2015 at 12:23:26 PM UTC-5, Jerry Stuckle
> wrote:
>> I don't know what college your friend attended, but even in the
>> 70's I don't know of any universities which only taught COBOL as a
>> CS major. In my university, engineering students were required to
>> learn FORTRAN (and not much else CS-wise), and I think (but am not
>> positive) some business-related studies required COBOL.  But CS
>> required much more.
>
> My brother was a CS major.  He does AI development.  He learned
> FORTRAN, Pascal, C, all that fun stuff, no COBOL.  They ought to
> teach COBOL to CS majors, considering how many with CS degrees seek
> jobs in the CIS field.

Learning COBOL will just learn you to program in COBOL.

You should learn to program.

Languages like Algol and Pascal is very good at teaching
basic imperative logic and algorithms.

C++ would have taught you OO concepts.

Arne

[toc] | [prev] | [next] | [standalone]

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

Back to top | Article view | comp.lang.java.programmer

csiph-web