Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site bu-cs.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!whuxlm!harpo!decvax!genrad!mit-eddie!think!harvard!bu-cs!root
From: root@bu-cs.UUCP (Barry Shein)
Newsgroups: net.news,net.news.group
Subject: Re: Removing net.flame (attempt at a fix)
Message-ID: <451@bu-cs.UUCP>
Date: Sun, 30-Jun-85 20:37:56 EDT
Article-I.D.: bu-cs.451
Posted: Sun Jun 30 20:37:56 1985
Date-Received: Fri, 5-Jul-85 02:43:27 EDT
References: <3892@alice.UUCP> <1913@ut-ngp.UTEXAS> <476@umcp-cs.UUCP>, <1624@qubix.UUCP>
Organization: Boston Univ Comp. Sci.
Lines: 83
Xref: watmath net.news:3568 net.news.group:3219

[note: this got a little long trying to anticipate some design issues]

1. This is not original, I believe something like it was used in a
system called Xanadu.

2. Suspend your immediate ideas that all this could be circumvented by a
hacker, assume peer pressure and other checks by other sites.

3. This focuses on quantity more than quality but should improve both.
Ok...

The primary way to prevent abuse of resources is to put a cost on those
resources, even an artificial one (eg. computer funny money.)  This is a
summary of an idea I heard second hand about another system that made
some sense to me. Maybe rather than flaming flames people could work out
the necessary details and software if they are interested.

Assume you are given an account to send some number of messages and some
amount of message text per unit time (say, month.)  Now, everyone starts
out with some reasonable amount which limits them. The positive feedback
in the system is that if people find your article useful (this part is
hard!) they return to you, at no cost to them, one 'unit' of I like your
stuff which then ups your allotment. If they don't, your account just
drains.

The analogy is publishing. You pay for the cost of printing etc and if
you did something useful then people buy your book and you are
encouraged to do more. If they don't, well, you are out of business.

Problems with implementation:

1. How do people 'pay' you? First, I would hope the system would not
limit this to only if they agree with you, just that they found it
interesting. An overly simple idea would be just that they read (printed
out) your article tho that's the only way to often discover that it's
just garbage and maybe not what the Subject: line claimed. A purely
voluntary system (please enter value: ) would not work either I don't
think, although maybe if the answer was just 'y' or 'n' (Pay? ) that
would suffice.  Yes, people would abuse it somewhat but hopefully it
would average out to the truth and everyone is subject to the same
silliness anyhow.

2. How do we enforce? What is to stop people from upping their accounts
on unregulated machines? I would just assume that a filter would keep
accounting at every site that wanted to and major abusers would be
spotted with typical consequences (warn, threaten, shut them off.)
Counterfeiting is a serious crime.

3. What about things like net.sources which are voluminous, wouldn't
that discourage some of the most useful stuff?  Not if it were really
useful, as I said, you get paid back.  We have to assume your initial
account is large enough to allow such postings or perhaps some groups
like net.sources could be unregulated, lower rates or post-accounted.

4. Isn't this gonna cause a lot of errors due to added complexities?
What if my '$$' don't find their way back to me due to some transmission
problem? Them's the breaks I guess, again, would probably affect
everyone the same if that's any comfort and it would be sad if such a
concern would kill an otherwise good idea, no?

5. Won't all this accounting info cause even more overhead? If it causes
more traffic than it reduces than it is a failure of an idea. I suspect
though we aren't talking about an enormous amount of data though the
whole assocative data base could be problematic for sites with limited
disk space. A back of the envelope calculation:  Say you kept each site
in a table (I dunno, 1,000 sites?) which took 20 bytes each, that's 20K.
Add to that 20 users for each site (ie. non-zero this month only) at 8
bytes for a userid plus 8 bytes for an account amount (4 for usage, 4
for payage), (16 bytes) which is another 320K or 340K total. Double
that and you're still not up to very much disk space.

6. These data bases of accounting will *never* be in synch!  Not
generally important unless you suspect someone of being an abuser which
is probably the only time you would look at another site's usage/payage.
In this case you could always double check with some other sites.  The
main hope is that your local base for your users is kept up which should
not be as hard. Another possibility is a small number of sites try to
act as official record keepers sending out summaries at the end of the
month for everyone or on demand (tho in any case you keep your local
users accounting.)

	-Barry Shein, Boston University
	Free associating in public, as usual