On Thu, 13 Dec 2007 12:36:54 -0500, Bob Shannon wrote:
<Ed J.?>>And, why would they? What's wrong with PDS?
>
It is what it is; WAD.
>Refer to the "PDS Pain" white paper for what's wrong with PDSs. I suspect you
>already know. I didn't start this thread, I just provided an opinion.
>
I can find excerpts in:
http://www.redbooks.ibm.com/redbooks/pdfs/sg246524.pdf
PDSE was created as a result of an IBM User Group writing a paper titled
"PDS PAIN". The problems with PDSs eliminated by PDSE include these:
o Periodic compress not needed: Maintains list of reclaimable space
o Fixed size directory eliminated (a source of high connect time):
Indexed based entry
o Need to re-write entire directory when adding members
[Actually, just half, on the average. But deleting
numerous members from a member list approaches the worst
case unless you remember to first sort descending. -- gil]
o Cross system sharing of members and data sets
o Directory cannot be overwritten
o Change in BLKSIZE making older members unusable
o Storage efficiency improved in many cases
(although small members can take up more space)
o 16 extent restriction (PDSE allows 123 extents)
o Need to delete aliases when a member is deleted: Alias deletes when
member is deleted
[From a UNIXcentric POV, that last can be considered a defect,
not a virtue. -- gil]
Is a citation of the original available?
Thanks,
gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
