Subject: | |
From: | |
Reply To: | |
Date: | Wed, 1 Jul 1998 08:23:36 -0500 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
At 06:12 PM 6/30/98 -0700, Jeff "CI am He" Vance wrote:
>...
>That is, the exiting output is unchanged for LDEVs up to 3 digits and
>then shifts right by up to 2 more digits to accomodate the largest
>anticipated LDEV number.
Oooohhh!! NO!
Not only is this a more complicated solution to code, but it is much harder
to parse (most people use fixed column parsing, I'd bet). Why not just
accept a new fixed field width (5 digits seems reasonable)?
>This leaves 9 bytes for a disk model number
>before we have to shift the output again. Today we have 8 byte model numbers
>(although the output field only shows 7 bytes) but we expect to need to
>support up to 24 byte model numbers.
Several people have commented on the "tremendous" need for a 24 byte model
number. How about floating a proposal to limit model numbers to 12
characters. This does, after all, allow for 36^12 (4.738 * 10**18) model
numbers, which allows each of the inhabitants of the earth since life
appeared to have 2 personal model numbers and then some over!
>
>We could:
>4) apply rule 3) but make the field width 12 bytes rather than 9. This
> would look like:
>
> LDEV-TYPE STATUS VOLUME (VOLUME SET - GEN)
> ---------- -------- --------------------------
>12345-C23456789 MASTER MEMBER1 (MPEXL_SYSTEM_VOLUME_SET-0)
I prefer #4, always assuming, of course, that the real data is available
both from an intrinsic and also from a command wrapper!
Tony
|
|
|