Category Archives: mainframe

My Work With The Mainframe.

I am actually working in a bank, working with a mainframe, which is really awkward. Because it is priviledged and strange at the same time. Many people have misunderstanding on the mainframe system.

Well for one, it is not UNIX, it is not old, it is a different animal. And it is menu driven not much on commandline driven, aka ISPF. In fact it is so different, you really need to use it to see it. The heritage of mainframe is really old. Many old concept still inherited in the modern mainframe.

On linux you can create file, just one line of command, or on windows, just right click and create. Not quite as easy on Z/OS aka mainframe. You need to call tso command, then supply with parameter such as Block Size, Record Length, and even Volume Name. Or you can write a JCL script(I would call it a script). That provide the same information. To create a file, then submit it. But there is a command to create a dynamic datasets.

Also the file is different, flat file is almost the same, except we call it datasets. We have vsam file. Which is something like a files with keys. Which effectively an database file. Except it is not quite there. 

On a linux system, you can write a script in a long line if you want it. Not quite on Z/OS, the limit in a line, is 80 byte. Which is the same length on a punch card. Which is interesting. And limiting at the same time.

JCL is the important part of mainframe. It is use to run batch. Yes, we still run program as batch, like the goold old days. It just that we don’t use punch card anymore. But program is run by batch. A batch run is different that what I remember in programming. It is like running a few program in a flow. And the jcl is used to allocate, files needed to the program. Depending in a situation, it can take a day, or few days to finish running the batch. Depend on the size of the program, maybe system is the more appropriate term.

Actually there is many many difference. So there will be a part 2. And this should be the beginning, of a long multipart article on mainframe. Which is really based on the one I use in my job. Which I need to filter out some sensitive stuff.

playing archeologist @ the job

My job at the bank involves software testing. As a tester. It can be boring at time. And trying to find defect on a huge system can be crazy. And my team testing a software that is huge……real huge and messy.

It just happens I am free(and bored), at the company. So I dig around with the codes. Looking at codes and copybooks(cobol equivalent of a header). Seeing what it does etc. It is like looking at a old writings, writen in a style few would understand. And like digging at the sites.

The hard part is, the system contains hundreds of copybook, and hundreds of working codes, not to mentions JCL script. Some is commented, often not all contain the enough information. And I still try to find the documentations. The full documentations I mean. Piece together a few at a time.

The prelimentary result is, a few of the copybook is found, and it is not documented(not sure how it would help in the work). And still a lot more to dig.

Lesson learnt, file-aid is a useful tools on the mainframe. And cobol is verbose, but readable. And lastly, looking at how different mainframe and other platform software look is interesting.