|
<<given a seq. file with fixed-length records, can i, based on the RFA of any
<<given record, safely calculate the number of this particular record? Of
<<course i'd have to take into account if i have even/odd-sized record sizes,
<<and take care of the block-span attribute. But if my algorithm does the right
<<things, do i have to take care of anything else? Of course a deleted
<<(=nulled-out) record would count as a valid record, but i'm aware of that fact.
If your algorithm does the right thing (and your caveats appear to have
covered all the bases), of course you should be able to generate the correct
record number.
<<Would such an algorithm be supported?
I presume you are asking that if your algorithm was added as functionality
in some user's application, would OpenVMS engineering support it if it didn't work
at some future time. If that is your question, then the answer is no.
<<This would depend on the right interpretation of the RFA for a seq. file as
<<holding a 4-byte vbn and a 2-byte byte offset. I couldn't find that documented
<<anywhere, so is this a supported interpretation of an RFA value?
Its internal representation is not documented (supported) for the very reason
that its internal representation may change at some future time. Now while
that is not likely to happen, there is no 100% guarantee.
-- Elinor
|