On 2013-11-29, at 12:16, John Gilmore wrote: > Under OS/360 the notional, antetypical 'longest' index had the syntax > > <level1>.<level2>.<level3>.<level4>.<level5> > > Then, since <leveli> values could be at most 8 characters in length, 5 > x 8 + 4 yielded the maximal character count of 44. The 44-character > and 8-character maxima remain; the 5-level maximum does not. > It should be easy to relax the 8-character maximum (except, perhaps on the HLQ) without incompatibility with existing data areas, even was done for the antetypical 5-level maximum.
In an article here about a week ago that I haven't the stamina to find, the writer argued that the customary syntactic limits should be retained in new or upgraded components for reasons of some sort of compatibility. Hmmm. Called from Assembler, allocation, BLDL, STOW, and I suspect (haven't tried) ATTACH, LINK, LOAD, and XCTL support far beyond the customary majuscule alphameric plus a handful of special characters. Similarly for Binder with CASE(MIXED) and for catalog with DISABLE(DSNCHECK) in effect. I disagree with that writer. In an environment where facility A accepts a more tolerant syntax and facility B accepts only a more restrictive syntax, if I were crafting an application that might interact with either A or B, what should I do? I might choose to support the syntax of A, aware of the risk that I might create objects that would be inaccessible to B, or I might choose to support only the syntax of B, aware of the risk that I would be unable to access objects created by A. I would unhesitatingly choose A's syntax; in fact the union of all the grammars I might need to deal with. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
