Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 exptools 1/6/84; site ihuxp.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!harpo!ihnp4!ihuxp!wbpesch
From: wbpesch@ihuxp.UUCP (Walt Pesch)
Newsgroups: net.news.group
Subject: Re: net.cs proposal - (nf)
Message-ID: <664@ihuxp.UUCP>
Date: Mon, 5-Mar-84 15:17:01 EST
Article-I.D.: ihuxp.664
Posted: Mon Mar  5 15:17:01 1984
Date-Received: Tue, 6-Mar-84 03:20:27 EST
References: <971@inmet.UUCP>
Organization: AT&T Bell Labs, Naperville, IL
Lines: 79

It's time for the old foot in the mouth routine:  a recent posting
(inmet!tower) listed a few of the CS discussions that have occured in
the path, and where they occured.  From the tone of his article, I
think that he meant the listing as evidence against the creation.  Let
us look at the examples he poses:

>>>  RISC has been discussed in:
>>>  net.arch		Computer architecture.

The discussion (and I have read old net.arch) dealt in the
majority with the theory of RISC, not the RISC I chip, which would
properly belong there.  There was mostly posting on references to
journals.  But I will agree that this is appropiate for atleast a
subset of the discussion.

>>>  Number theory has been discussed in:
>>>  net.math		Mathematical discussions and puzzles

But what about the Comp Sci applications... numerical analyses.  Are
we to cross-reference between net.math and net.arch if we are to
discuss new hardware for on-board floating point error and the
calculus behind such?

>>>  database design theory has been discussed in:
>>>  net.arch		Computer architecture.

Boy, this is a great one.  I suppose you want to discuss the proper
ways of mixing concrete in net.arch too.

>>>  generic data structures has been discussed in:
>>>  net.lang		Different Computer Languages.

See above.  What if I want to discuss tree structures (something all
netnews administrators should want to).  Do I have to go to
net.lang.ada?

>>>  Hamming Codes has been discussed in:
>>>  net.dcom		Data Communication Hardware and Software

Good match here, but how many people read this group?  By the way, I
could make a good arguement as to putting this in the all-encompassing
net.arch, for architectures entail Hamming Codes in memory (or atleast
parity bits.)

>>>  Robotics has been discussed in:
>>>  net.ai			Artifical Intellligence

From the application side?  I really feel it is a misnomer to put AI
and robotics under the same roof.  (Though they may be on the same
block in some cases.)

>>>  Queuing Theory has been discussed in:
>>>  net.math		Mathematical discussions and puzzles

Do they really want me to come in their and discuss the Mean Time to
Failure of various computers systems and their integral parts?

                          -------------

Well, where does this take us and what does this show.  I see that we
have here a variety of topics (pardon me but "Orphanned Topics") that
do not comfortably fit in a given newsgroup, for they can also be
considered either only a related topic or else a superset of the
stated purpose of the newsgroup or else a gross misuse of the
newsgroup.  (Database theory in net.arch?  PAH!)  Once again, the
critics give more evidence.  

The list above is nt comprehensive of what can be discussed in a
generic computer science newsgroup.  But it is representative.  I
still do feel that it also shows the necessity.  A newsgroup today may
save nine.  (A generic newsgroup may stop the creation of many small
limited and quickly-dying discussions.)

Once again, thanks for your time.


                                          Walt Pesch
                                      AT&T Technologies
                                     ihnp4!ihuxp!wbpesch