On Mon, 26 Dec 2011 08:00:17 -0600, John McKown wrote: >On Sun, 2011-12-25 at 23:17 -0500, Robert A. Rosenberg wrote: >> At 19:20 -0600 on 12/24/2011, Paul Gilmartin wrote about Re: >> Eight-character TSO Userid Support: >> >> >Hmmm. How many upper-case alphanumeric characters would be needed >> >to provide unique identifiers for all the files an enterprise would ever >> >need? >> >By that metric, 44(8) is more than sufficient. >> >> A dataset name is up to 44 characters long but that is based on a set >> of 36 characters (26 letters and 10 numbers) CURRENTLY broken up >> into 1-8 character blocks separated by periods. The 8 character >> blocks are composed, if I remember correctly of 1 Alpha (A-Z) >> followed by 0-7 Alpha (A-Z) and Numeric (0-9) characters. There may >> also be some special characters in addition to the 36 (I forget). > There's a PARMLIB option that allows the use of far more than 36 (39?) (40?) and removes the 1-8 character blocking requirement.
>Each "node" is 1 to 8 characters. First character is alphabetic or a >national character, of which there are three: @#$ in U.S. The other >seven characters may be any of those plus the digits 0 through 9 and a >dash (-). The dash threw me, when I first saw it. I don't know when that >became legal in a DSN. > And it's bizarre. When last I tried it, dash was allowed in the DSN= parameter, but not in the DCB= parameter of the DD statement. Go figger. Conway's law? Has IBM any good reason not to make this specification uniform? It shouldn't take a dammed SHARE requirement. I hate JCL! >> For GDG Datasets, the limit is 39(8) since the last 9 are required >> for the ".GxxxxVyy" suffix to the GDG File Base Name. What do you >> mean by 44(8) anyway? >> By analogy with the Windows FAT file system limit, often called "8.3". -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: INFO IBM-MAIN