RE: Naming/Format conventions for INDEX files

["George D. Greenwade" <[e-mail]>]

On Mon, 27 Jul 92 12:20:21 -0400 George Ferguson
<[e-mail]> posted:
> Just a further commentary to those of you wanting to automate ftp
> retrieval.....
>
> In any event, this subject was hashed out in some other forum a while back
> (comp.archives.admin?). I doubt that any consensus was reached.

Where are comp.archives.admin archives available??  I've looked in all of
the directories shown by archie and, for the most part, they all ask this
question.  I would love to see this previous exchange.

It's interesting and good that the issue has been looked at before.
Moreover, it's also probably good that it is recognized as a problem which
cannot be resolved by an easy consensus -- this is a REAL problem.  This is
my concern -- if the IAFA drafts are submitted and become a standard of
sorts, it's an awful lot better to stop right now and look at this REAL
problem than to merrily move along because some admittedly non-trivial work
has already taken place.

A consensus is not even what I am after since different OS'es are different
OS'es and making one OS look like and feel like any other is, plain and
simple, pretty hard to do.  All I am suggesting is that whatever the IAFA
does in its official capacity be limited in scope to what can be
practically applied right now -- no more, no less.  This may mean nothing
more than filenames are defined to be case insensitive and always contain
one and only one period and the work is largely done.  The changes to the
existing documents could then be:
(a) rename draft.part.I and draft.part.II to something platform-independent
    (that ought to be done anyway!);
(b) include the use of a period in the renaming section of draft.part.I as a
    critical character (if not *the* most critical character); additionally
    *highly* suggest that the files have not more than one period and tell
    why; and
(c) in draft.part.II, quit discussing the example as a Unix-only example as
    it appears that the suggested documents are already pretty portable.

These are easy suggestions.  If you really want to standardize, then the
weakest links among all OS'es in each dimension are going to have to be
used.  If anything other than the weakest link is chosen, it will break on
the systems which possess that link.  I'm willing to give up RMS indexed
files in exchange for the one period rule, as well as giving up filenames
longer than 14 characters in exchange for true case insensitivity.  I
believe this is a fair exchange as an RMS indexed file could easily be NFS
mounted and you could have all sorts of things in the IAFA-proposed
indexes.  Realistically, though, I understand Unix does not support this
concept ("really ought to though; gosh, I guess Unix is just a `dumb' OS"
-- a really stupid remark by a hypothetical VMS user which is of equal
stupidity with a remark by a hypothetical Unix user, "include as many
periods as your heart desires, so long as it's not more than 14"). 

Conceptually, the OS'es are different; conceptually, ftp is ftp;
practically (and here's what matters!), how an ftp site is set up can
largely determine if the concept of "ftp is ftp" rules or if the concept of
"this is my OS, what's yours?" rules.  I sincerely hope that the IAFA moves
to make "ftp is ftp" the rule, rather than having system-centrism.  Every
aspect of the draft and work must be looked at insofar as portability or
system-centrism as the likely result of implementation.

Please note that I am basically impressed with the IAFA work, to date.  To
those involved so far -- congratulations and good job!  It's now, in these
final stages (based on the target November completion date), that I have a
concern.  Instead of wrapping up, it may be time to unwrap and examine the
package, test it, place it on various systems, and see exactly what has
been done.  As well as remove the [VMS] notation from draft.part.I and put
something constructive in there.

Regards,   George
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
George D. Greenwade, Ph.D.                            Bitnet:  BED_GDG@SHSU
Department of Economics and Business Analysis         THEnet: SHSU::BED_GDG
College of Business Administration                    Voice: [phone]
P. O. Box 2118                                        FAX:   [phone]
Sam Houston State University              Internet:        [e-mail]
Huntsville, TX 77341                      [e-mail]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%