Checksum: 53300
Path: utzoo!utgpu!adam
From: adam@gpu.utcs.toronto.edu (Adam R. Iles)
Date: Mon, 21-Mar-88 17:51:20 EST
Message-ID: <1988Mar21.175120.4618@gpu.utcs.toronto.edu>
Organization: University of Toronto Computing Services
Newsgroups: comp.sys.ibm.pc
Subject: Re: Moderation...
References: <1424@polyslo.UUCP> <149@tscs.UUCP> <125@aeras.UUCP> <275@syntron.UUCP> <345@c10sd1.StPaul.NCR.COM>
Reply-To: adam@gpu.utcs.UUCP (Adam R. Iles)

In article <345@c10sd1.StPaul.NCR.COM> johnson@ncrcce.StPaul.NCR.COM (Wayne D. T. Johnson) writes:
>Would it not make more sense to find 1 (or more) nodes that are willing
>to archive the binaries and have the authors (or their representatives)
>just distribute a synopsis of the program for all to see.  These archive

<stuff about a server>

>This would greatly cut down on network traffic because only requested
>software would be transmitted. 

NO NO NO !!!!!!

This would GREATLY INCREASE the trafic on the net.  The present method
only involves sending a file between any two nodes once for everyone on
the net.  Your scheme would see it being transmitted for everyone who
requests a copy (almost everyone :-)  And don't forget that this includes
EVERYONE who is downstream from the server.  In the worst case (assuming
all subscribers request all the files only once) this would MULTIPLY the
load on the system by about 1/2 (assuming an even distribution between
all nodes) of the number of subscribers!

>				 Not everyone wishes a copy of every
>program sent out.  In my case I think I only keep 2-3% of the binaries
>that pass by.

The load isn't caused by someone keeping the file it's caused by someone
requesting the file.

>		The amount of time spent re-transmitting portions of a 
>program (and the requests to retransmit it) could also be eliminated.

What would be eliminated is the transfering from a node to a PC.  This
could be done by being careful about what you download.

>I have seen references to various archives around the network, since I
>am new here, some of what I have suggested may be already in place.  But
>the basic idea of only brodcasting small descriptions of a program and
>allowing those interested to order it instead of mass distribution makes
>sense to me.

Not to me!  It's VERY easy to listen to see smaller numbers and think that
it means that it's cheaper, but this plan wouldn't work.  The idea of
including a synopsis of the program along with it has been around for a
long time, but that's ut to the person posting/moderating.

I really would like to thank Wayne for posting his article to
to Comp.sys.ibm.pc rather than the binaries group.

Thank-you for reading this.  I'll just put away my soap box now.
-- 
    
        Any opinions stated above may, or may not, refect those
        of any sane person living, dead, or just sleeping.

       Adam R. Iles:	adam@utgpu