Steve claims:
...
> Fine; but Adager/DBGENERAL/whatever do not in any way constitute the
> entire universe of available methods to repack a data set. A perfectly
> legitimate, albeit naive, way to get the set back "in order" is to read
> the set, write it out to a file, empty the set, then read it back in from
> the file.

The same argument could be applied to a clerk reading a QUERY REPORT out
loud, over a phone, to another clerk who enters the data onto another
machine ... but that doesn't mean it's a method of repacking that would
have been considered in this discussion, *any more than the read/erase/load
method would be considered a reasonable method of repacking*.

In short, repacking simply does not lose chronology.

Since it apparently didn't go without saying before:
   "repacking" means repacking a dataset with a real-word usable & reasonable
   tool like Adager (or a one or two others),
and
   "chronology" refers to the relative logical location of entries on a chain,
   not the relationship between the physical locations.

Or, in other words, the usual empirical definitions of the two.

> Presumably the read would be done in some sorted order that
> makes sense to the user, such as to group records with the same values in
> a particular field together. What happens to the chronological sequence
> of non-sorted paths under this transformation?

It's a nearly irrelevant question, as it's being applied to an
exceedingly unlikely repack method.  If you're doing that kind of
repacking, dial 1-800-LEE-REGO and ask for help now! :)

> If there is a
> properly-constructed sort item on each chain that the user wants to
> remain in chronological sequence, not a thing. If there is no such sort
> item, I think it's pretty safe to say that the integrity of the
> chronological sequencing of those paths will be negatively impacted.

Also, a sort item doesn't guarantee anything in the unload/load scenario
either, unless the sort values are guaranteed unique and correct.

> Remind me again; which of my statements were "incorrect"?

Sorry...just re-read your posts, and compare/contrast with mine and Denys.
Also, factor in the "reasonable interpretation" of repack (see above).

--
Stan Sieler                                          [log in to unmask]
                                     http://www.allegro.com/sieler.html