Mainframe, MVS and zOS Discussion

Hi,

I have seen this many times this statement "They are data sets - not files." like the one here too viewtopic.php?f=20&t=2311#p10343

What exaclty you all want to convey with this? I don't understand it. What difference does it make if we call them files or data-sets? Sorry for the question but this has bothered me.

Terminology is critical in IT, where similar terms may mean very different things. A file on a Unix or Windows machine is a sequence of bytes. The file may or may not have line terminators every so often. A data set on a mainframe has a record structure and has no line terminators since the characteristics of the data set are recorded in the catalog and VTOC and within the data set at times. A data set may be sequential, or indexed (if you go back far enough -- indexed doesn't exist on z/OS, though), or relative, or partitioned, or VSAM sequential, or VSAM indexed, or VSAM relative, or VSAM linear. The operating system recognizes and processes these different organizations differently so you can, for example, edit a member of a partitioned data set in ISPF without having to worry about what a PDS really is.

One really good example of the difference is if you create a variable blocked data set on the mainframe, FTP it in binary to a Unix or Windows machine, then FTP it back to the mainframe in binary. Since variable blocked data sets store the record length external to the record, you will lose them by doing this and hence what you get back on the mainframe is not really usable and certainly has no records like the source data set.

Thanks Robert. I did not that difference.