Prev: MF COBOL - Is it hard to learn?
Next: Mapping (CoBOL) Methodologies to Problem Domains
From: Euro on 22 Jan 2008 13:01
We want to pass an application developed in Fujitsu Cobol (PowerCobol),
using indexed files (ISAM) for a database system (SQL).
Which the best form of doing (the application has tens files), without
having a great deal of the code to alter.
Which the database recommended.
He wanted a straight access without resource to ODBC.

Many Thanks
Joe


From: Alistair on 22 Jan 2008 15:09
On 22 Jan, 18:01, "Euro" <e...(a)euromercante.pt> wrote:
> We want to pass an application developed in Fujitsu Cobol (PowerCobol),
> using indexed files (ISAM) for a database system (SQL).
> Which the best form of doing (the application has tens files), without
> having a great deal of the code to alter.
> Which the database recommended.
> He wanted a straight access without resource to ODBC.
>
> Many Thanks
> Joe

Before anyone answers this they might like to know something about
whether you have the necessary expertise to write complex SQL queries,
how complex a function you may wish to perform and even how much data
you have. Finally, what programming language did you want to use?

If you wish to minimise the code to be changed then you will need a
version of cobol and you will have to use ODBC. But that will still
require a lot of work.
From: Pete Dashwood on 22 Jan 2008 15:15


"Euro" <euo(a)euromercante.pt> wrote in message
news:newscache$pr42vj$79c$1(a)newsfront4.netvisao.pt...
> We want to pass an application developed in Fujitsu Cobol (PowerCobol),
> using indexed files (ISAM) for a database system (SQL).

I have tools that can help with this. They have been used successfully and I
can give you referral sites. These tools are developed specifically for
PowerCOBOL migrations from ISAM to RDBMS. (although they will work fine with
ANY migration from ISAM or VSAM/KSDS to ANY RDBMS).

Data structure conversion to RDBMS is fully automated (resultant DB is in
2NF or 3NF depending on the key dependencies, data load to the new database
from the existing ISAM files is fully automated, and code conversion of
existing programs is partially automated.

I also have a lot of information, gained in practical experience with
migrations, for reference BEFORE you start migration.

Contact me privately.

> Which the best form of doing (the application has tens files), without
> having a great deal of the code to alter.

Data conversion is relatively easy and can be automated. Code conversion of
the existing programs is more difficult. There are pitfalls you need to be
aware of, and things you should know before attempting this.

It ISN'T just a matter of converting every ISAM access to SQL...


> Which the database recommended.

There are no BAD RDBMS. It is all a matter of subjective preference. SQL
Server, DB2, MySQL, ORACLE, and PostgreSQL are all favoured by millions of
users.

> He wanted a straight access without resource to ODBC.
>

Can be done, but may not be wise initially. Closing your options before you
start is seldom a good move...

Pete.
--
"I used to write COBOL...now I can do anything."


From: Robin Lee on 22 Jan 2008 19:16
Euro wrote:
> We want to pass an application developed in Fujitsu Cobol (PowerCobol),
> using indexed files (ISAM) for a database system (SQL).
> Which the best form of doing (the application has tens files), without
> having a great deal of the code to alter.
> Which the database recommended.
> He wanted a straight access without resource to ODBC.
>
> Many Thanks
> Joe
>
>

I recently converted several programs from ISAM to Microsoft SQL Server using
Fujitsu PowerCOBOL. This took the full year of 2006. It required that I learn to
use Microsoft ADO. The I/O was built around some sample code provided here in the
newsgroup by Frederico and posted on www.tek-tips (with a little help from Pete D.)

The resulting code works and seems reliable, but it is not clean, manageable code.

I have concerns about building system code directly into application programs,
knowing that this platform (SQL via ADO) is just another fish in the long stream of
platforms. The original COBOL source code is from the 1970's and I have moved it
across at least five different platforms over the past three decades.

In short, it can be done, but requires some compromises. But the result is that the
data lives on and can co-exist with modern desktop office applications.
From: Anonymous on 22 Jan 2008 21:04
In article <Ke-dndWVJY8HGgvanZ2dnUVZ_jKdnZ2d(a)giganews.com>,
Robin Lee <robinlee(a)news.com> wrote:

[snip]

>The original COBOL source code is from the 1970's and I have
>moved it
>across at least five different platforms over the past three decades.

See what happens when you can't find a reliable 1401 emulator?

DD
 |  Next  |  Last
Pages: 1 2 3 4 5 6
Prev: MF COBOL - Is it hard to learn?
Next: Mapping (CoBOL) Methodologies to Problem Domains