osh

[Steven M. Haflich]

Curtis Jackson hits the nail right on the thumb with his remarks about
naming old versions of programs.  His example was that
	nroff1, nroff2.4, etc. should be used very infrequently
	anyway, since everyone should be converting over to the
	new nroff.
Really?  Which "new nroff?"  The original one (the one that replaced
roff)?  The one that came out around the same time as Unix 6.5-7?
The slightly revised version which updates that one?  Or is there a
new one about which I have heard nothing?

I find it tremendously difficult to keep track of versions of the
various large system programs on Un*x -- system programs here intended
to include compilers, *roff, uucp, etc. -- and similar difficulty
establishing where a particular version was written, where its support
(such as it is) might be centralized, or where distributions can be
obtained.  (By the way, I am a Unix wizard of 7 years standing, although
I have been reading the net only 8 months or so.)  For example, recently
I became interested in getting ditroff, if possible, but there is no
obvious mechanism to determine its source or availability other than
nuisance broadcasts to the net.

We have a (monthly?) net posting of the list-of-lists summarizing
the currently valid newsgroups.  This costs the moderator some time, but
is of great service to the community.  (If only more new users could
could be directed to examine it.)  How about something similar for
"systems" sources, called "net.versions" or whatever?  It could
summarize the state of the art for each of the several dozen "systems"
and include:
	- The name of the system.
	- A brief description of what it does, limited to a line or two.
	- What it runs under.
	- The (original) authorship/institution.
	- Derivation from earlier versions.  (e.g. nroff9.2 is an enhanced
	  version of nroff9.1 with bug fixes but no new features).
	- Availability (and licenseing).  Most often this would be a simple
	  entry like "distributed with standard 4.1".  Otherwise, the site
	  or sites which are willing to distribute could be named, or if
	  an item is already *widely* distributed, one would know to
	  check adjacent sites first.
	- If appropriate, a central repository for bug fixes.  Look at the
	  current mess with uucp and the news system.  Who knows how
	  find all the bug fixes for a given version?  I can no longer
	  even keep the versions straight.

The reason this will never happen, like so many other things, is that
it would take too much work for someone.  (Probably a good deal more
than the list-of-lists mailing.)  I am certainly not volunteering, but
maybe some other fool will do so, start a list, and solicit updates
and corrections.