On Thu, 15 Jan 2015 12:19:46 -0500, Shmuel Metz (Seymour J.)  wrote:
>
>>It's more complicated than that.
>
>To begin wth, well coded assembler routines will use DYNALLOC
>directly.
> 
Correctness is in the eye of the beholder.  Which do you consider
correct, to DALPERM or not to DALPERM, or should interfaces such
as ALLOCATE and BPXWDYN provide an option.

I hate MVS.  Its surfeit of features is not offset by the usefulness
they provide.  I suspect much of the complexity of DYNALLOC
arises from an attempt to contend with a storage constraint now
to be regarded as ridiculously small.

Allocate DDNAMEs; free them when you're done with them.
What if you exhaust TIOT entries?  Well, DYNALLOC will pick
some to reuse.  What if you really needed them?  Well,
DYNALLOC lets you set a flag on those which are not to
be reused.  And so on.

UNIX has a limit, OPEN_MAX, vaguely similar to DYNAMNBR.
If I attempt to more files than that, open() fails.  There is no
option allowing the kernel to pick some files to close and
proceed.

Simpler Is Better.  How seriously do DOS and UNIX users miss
the ability/need to specify RECFM, LRECL, BLKSIZE, etc. on
their files?

--gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to