| View previous topic :: View next topic |
| Author |
Message |
Himesh
CICS Forum Moderator
Joined: 20 Dec 2002
Posts: 80
Topics: 21
Location: Chicago
|
|
| Back to top |
|
 |
zatlas
Beginner
Joined: 17 Dec 2002
Posts: 43
Topics: 4
|
 Posted: Thu Jan 30, 2003 10:23 pm Post subject: |
 |
|
There are only two problems with the scheme described in the scond article:
a) There is no ISPF for linux and you have to resort to one of these useless Unix editors like Emacs or Vi. 
b) There is no decent COBOL compiler and you have to resort to C or even worse C++ 
Oh, and you do not know the length of text record in advance, you have to look for end of line.
OTH, Perl is available freely on this platform 
ZA |
|
| Back to top |
|
|
DaveyC
Moderator
Joined: 02 Dec 2002
Posts: 151
Topics: 3
Location: Perth, Western Australia
|
| Posted: Fri Jan 31, 2003 5:02 am Post subject: |
|
|
You can use SlickEdit which has a superb ISPF emulator. Best editor on the market...
ZAtlas, for somebody who is a big fan of Perl, you seem to be anti C. Last time I looked Perl had very similar syntax to C/C++. Also, can you give me a good reason why anybody would want to write COBOL code for Linux ?
_________________
Dave Crayford |
|
| Back to top |
|
|
zatlas
Beginner
Joined: 17 Dec 2002
Posts: 43
Topics: 4
|
| Posted: Fri Jan 31, 2003 9:58 am Post subject: |
|
|
Hi Dave
1. Actually, I love C, even though it does not show in my postings. I am from the generation that used to code whole business systems in Assembler (the stupid decision to do so was my bosses' not mine.) In any case, C is probably the best Assembler replacement around and if both are available I would use C over Assembler any time.
2. I do not take back my comment about Emacs and Vi, so thank you for the hint. I will have to have a good look at this Slickedit.
3. Here is where I differ from the C camp: While C (and I guess C++ and Java and all the other object oriented objects out there) make great devices for nerdy system programmers (probably like both of us,) there is no justification to code business logic in those cryptic and read-resisting languages. I actually have even problems with any non-static storage allocation for business logic. It could be done behind the scene, Albright - but should not show in the business code. That's why I would like to have COBOL - and you know what - the none object oriented, the straightforward one - in any business logic development environment.
You know what else, we do demand from our compilers to take care of the optimization and push the efficient usage of the CPU. Why don't we demand from our compilers to hide all this s--- about whether we allocate memory from the heap or from the stack or from God only knows where. Why do we care whether we point to an object or just reference it? Why do we have to overload the poor shift operator ('<<') to mean iostream when we may simply issue a read command?
That's why I want COBOL!
ZA |
|
| Back to top |
|
|
DaveyC
Moderator
Joined: 02 Dec 2002
Posts: 151
Topics: 3
Location: Perth, Western Australia
|
| Posted: Mon Feb 03, 2003 6:24 am Post subject: |
|
|
| Quote: | | Why don't we demand from our compilers to hide all this s--- about whether we allocate memory from the heap or from the stack or from God only knows where |
Nowhere in the C/C++ standards will you find any reference to a Stack. Some implementations don't have program stacks. That's why the C alloca() function is not portable.
The big reason for using Linux on the mainframe is for web applications. IBMs language of choice for doing that is Java. Java doesn't have the dangerous features of C++ (pointers, memory leaks etc) so is pretty safe to write code with.
Also, reading a Linux directory/file is not well suited to COBOL, which is designed for processing records, not streams or HFS structures.
I personally don't see why COBOL should be considered better for writing business logic, and I used to write COBOL programs. In the last 20 odd years a big emphasis has been put on information hiding (all variables in a COBOL program are global to the file) and loose coupling (all procedures in a COBOL program are coupled). OO COBOL has not taken off (programmers had no interest) and is now only supported by IBM for Java interoperability. IMHO, PL/1 was always a better language anyway. Unfortunately, PL/1 was only a partial success.
If you look at the geneology of COBOL, it's FLOW-MATIC (1957) -> COBOL (1960) and that's where it stops, apart from some features incorporated into PL/1. FORTRAN, the old beast, is still very much alive in many of the features of the most modern programming languages.
_________________
Dave Crayford |
|
| Back to top |
|
|
Dibakar
Advanced
Joined: 02 Dec 2002
Posts: 700
Topics: 63
Location: USA
|
| Posted: Wed Feb 05, 2003 8:03 am Post subject: |
|
|
Dave,
Could you explain what is 'HFS structure'?
Diba |
|
| Back to top |
|
|
kolusu
Site Admin
Joined: 26 Nov 2002
Posts: 12380
Topics: 75
Location: San Jose
|
|
| Back to top |
|
|
Dibakar
Advanced
Joined: 02 Dec 2002
Posts: 700
Topics: 63
Location: USA
|
| Posted: Fri Feb 07, 2003 4:36 am Post subject: |
|
|
Thanks kolusu,
Diba |
|
| Back to top |
|
|
|
|
|
FAQ
Search
Quick Manuals
Register
Profile
Log in to check your private messages
Log in
