[Default] On 15 Dec 2016 07:44:37 -0800, in bit.listserv.ibm-main [email protected] (Pommier, Rex) wrote:
>Do they care about efficiency? Chances are that the vendor supplied routines >are going to be more efficient than home-written routines. By the time you take into account CALL overhead, I would not be surprised if I could write a routine that is faster in COBOL 85 or newer. I know that I can substantially beat the LE routines for date conversions. Clark Morris > >-----Original Message----- >From: IBM Mainframe Discussion List [mailto:[email protected]] On >Behalf Of John McKown >Sent: Thursday, December 15, 2016 9:25 AM >To: [email protected] >Subject: how to: convince programmer something else is better. > >I'm having a bit of a problem convincing a COBOL programmer to "not do" >something because, IMO, there is a better way. However, it almost seems like >he is not hearing me. > >Some background. We have an in-house written RLE blank compression routine >which we use for some selected data sets. These data sets are known to have >large sequences of blanks in the records. We have used this program for over >25 years (it was here when I came here). So this predates SMS compression by a >couple of decades. It was also used because, back in that day, VSAM KSDS data >sets were limited to 2 GiB. Oh, the program is written in HLASM. > >Well, we are getting off of z/OS to Wintel. Regardless of cost and time. >"So let it be written. So let it be done!" As part of this, we will need to >uncompress the data and transfer it to Windows. There are no problems with >that (other than screams from the SAN people about the size of the resultant >files). Part of this is using MicroFocus COBOL to do some of the work (i..e >port the COBOL programs to Windows). The aforementioned programmer wants a >COBOL version of the HLASM compression and decompression routines. In fact, he >has written the decompression routine. He wants me to write a COBOL version of >the compression routine. Which I am working on. >This is being done "just in case" they are needed (i.e. the SAN people scream >long & hard enough about space issues). However, MicroFocus supplies >3 different compression routines which can be called from COBOL. What I'm >trying to convince the programmer to do is write the Wintel COBOL versions of >the compression routines to take the same parameters as they have >historically, but to call the MIcroFocus routines within them to do the actual >compression and decompression. But, again, it is like he is just not hearing >me. > >Do any of the COBOL programmers out there have any idea how I can phrase this >information so that it makes better sense to him? I.e. convince him that using >the vendor compression routines, via a compatible interface routine, is the >way to go? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
