[comp.std.unix] why standard tape format needed

dan@wilma.bbn.com (Dan Franklin) (08/18/87)

From: dan@wilma.bbn.com (Dan Franklin)

In Vol. 12, No. 13, Doug Gwyn makes several good comments about
cpio vs. tar, but then goes on to say:

> I never did understand what inter-system archive interchange formats
> had to do with specification of a portable environment for
> applications...

The most obvious way they relate is that most applications need to be
installed somehow.  How should the developer of the application
package it so it can be read by all POSIX systems?  If a standard
format, such as tar, is specified, then the developer can produce the
software in that format and have it copied onto the appropriate medium
for distribution to different machines.

Some applications may also want to provide a facility to save their
file state, carry it onto another machine, and bring it back.  A
standard format would be useful here too.

> You probably couldn't read my 1/4" tape cartridge no
> matter what archive format I used on it.

Irrelevant, on two counts.  First, there are a lot of machines out
there with half-inch tape drives, and they *could* interchange tapes if
the issues being discussed now were resolved.  Second, there are (as I
understand it) only two major 1/4" tape formats, and work is underway
to make one of them standard, so in the future I *will* be able to read
your 1/4" tape cartridge--if the POSIX format is standardized.

Regarding 1003.1 vs. 1003.2: although an argument could be made for
pushing this discussion off into that group, I think it would be a bad
idea.  The tape format is not the user interface, as this discussion
has emphasized.  Keeping the tape format in 1003.1 helps everyone to
remember that fact.  Not only may other programs want to use the same
format, but more importantly, for 1003.1 we can ignore the user
interface altogether.  Since both the user interface and the tape
format are incendiary subjects, in this way we have some hope of
actually settling at least part of the issue.

Also, if we postpone this discussion to 1003.2 the whole context of our
current discussions will be lost.  We will go through the same cycle
all over again.  (Does anyone really want to have to keep repeating, when
the 1003.2 discussions begin, "that point was raised in our 1003.1
discussion and we decided that..."?)  The only difference will be that
the discussion will get even more muddled than it is now, because
people will jump back and forth between format issues, which are
fundamental and hard to change, and user interface issues, which are
surface issues that are much easier to change over time.

Please, let's try to decide the issue now.  We seem to be getting
closer to a resolution; let's not lose our momentum.

	Dan Franklin

Volume-Number: Volume 12, Number 16