Stupid reposting service (was Re: SLIP/getty/printers on Terminal Servers)

[Edward Vielmetti]

I believe my article <EMV.89Dec7021427 at noether.math.lsa.umich.edu> in
alt.config and alt.sources.d should address any "fair warning" issue.
These were clearly marked as experimental, and should be considered as
such.  (Have we all forgotten what alt is for?)  I am generating new
information: where sources are, some classification information, and
the actual contents of the original articles for the convenience of
those who didn't see it the first time around.  I think I've also
increased the amount of other people's postings to alt.sources so the
side effects seem to be OK, at least within alt.

After a few days worth of repostings I think I have a better handle on
what constitutes good things to repost, bad things to repost, and
questionable stuff.


Bad: "very long sources" (1000 lines?  500 lines?  I dunno.  Maybe 0.)
     patches in comp.sources.bugs
     one-liners (even clever one-liners)
     announcements of where to find sources (they're not sources)
     Archive-name: without a valid SysV pathname immediately following
     Anything that's already in comp.sources.*

Good: postings of up to about 8K (300 lines)
     general interest stuff posted to obscure groups
     (some say nothing would be good)

Questionable: postings from 8K to 25K (300-1000 lines)
     special interest things (i.e. elisp, REXX, DCL, perl :-)
     patches, esp. patches not by the author (alt.sources.patches ?)

On the matter of headers.  I think it's reasonable to tack on headers,
especially if it makes it easier for someone to archive this stuff;
the cost isn't too great, and the added likelihood that someone else
will have what you once saw in alt.sources and will be able to find
it makes the difference.  I've tried to come up with a minimal 
subset to see to it that not much information is lost; the original
article is currently in the References: line so if you want the
real McCoy it's very easy to get it.

If I can get cooperation from archiving software writers, I'd like to
propose a change for my treatment of jumbo stuff.  Instead of sending
out the full text of the reposting, I'll tack on appropriate headers
that have the message-id and original newsgroups of the thing in
question.  Your software will have to parse this information, grab the
article in question, and set it aside.  This can be problematic if the
repost is from a particularly obscure group, if you've expired it
already, or if it hasn't arrived yet; that would be up to you to
cope with it.

--Ed