--- On Tue, 3/17/09, Elardus Engelbrecht <[email protected]> wrote:
> From: Elardus Engelbrecht <[email protected]> > Subject: Re: COBOL and dynamic allocation (my two cents) > To: [email protected] > Date: Tuesday, March 17, 2009, 2:00 AM > Ed Gould wrote: > > >Personally I find it abhorrent to hard code anything > like destination in any > programs. JCL is the ONLY way to go. It can be searched > with multiple utilities > (FILEAID and others) rather easily. This issue is that when > (I did not say IF) a > destination changes you must remember to recompile the > program. > > > What about writing the COBOL (or any other language) > program so that > destination(s) and other variable changes can be read from > a dataset instead > of hardcoding it or writing such things in JCL? > > There is thus only one compile/test action and many changes > in > the 'parameter' dataset on as needed base. > > Any better ideas? > Actually no. once you hide in another dataset it doesn't get scanned like proclibs do. I suppose if you comment the hell out of it is semi acceptable but be aware people will miss the link somehow. And do not forget people leave their jobs and take knowledge with them. We had a similar case about 15 years ago (or so). We had one person maintaining the DMS exception table. I got to look at it and it was worse than a pig sty no comments no nothing. That we the case to use DFHSM instead and it actually worked better. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
