It will depend on the SMS setup, but the settings are typically just the defaults, so if you were to code with no space, you would get whatever space the site has coded as a default. Personally, I try to keep the default allocation pretty small because people will tend to waste a lot. You can code RLSE, but that can cause issues for some jobs.
Basically, you can code without the Space, and without a DCB, but you need to remember to keep it as defaults. I have been to sites where they override the SPACE and DCB, and it works for small subsets, but almost always gets them in trouble when they try to expand it to cover more things. Many sites set up SMS to be based on the dataset name parts, but I think that's taking the assumption that the users can spell a bit too far. :) JES3 was a little better at this than just SMS alone, but now even with JES3+ things are pretty close and SMS can do it alone. Whether you 'should' keep people from using space is debatable. Brian On Sun, 3 Aug 2025 00:13:01 +1000, Clement Clarke <[email protected]> wrote: >I am writing a simple EXEC command so that running Z/OS programs can be >executed in much the same way that Linux and Windows programs are >executed. For example: > > EXEC PAYROLL1 Payroll.Input.dets Payroll.validated.dets; > >where the program Payroll1 inputs and outputs a file after validating the >transactions. > >---------------- >I asked Google's AI (gemini.Google.com) the following question: > >"These days, with SMS what does a typical JCL statement look like for >allocating new data sets? Does one code Space, Unit and Recfm, or leave it >off the JCL and rely on SMS?" > >Google gave a long answer (try it yourself?) but in essence it said: > >"With SMS, the JCL becomes much cleaner. The system handles the allocation >details based on its policy. A typical DD statement for a new dataset now >looks more like this: > >"Code snippet > >//NEWDS DD DSN=MY.TEST.FILE, >// DISP=(NEW,CATLG,DELETE) > >---------------------------- > >Is this correct, or does most JCL still have SPACE etc coded for new data >sets? > >Many thanks, > >Clem Clarke > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
