In <oiag5uotf1nnq4rho9psvcq9d8m8etf...@4ax.com> on Wed, 30 Jan
2002 17:32:24 +0000, {R} Richard Ashton <'{...@semolina.org>
wrote:

I despise top-posting and overquoting with passionate venom, so
I'm totally sympathetic to the intent behind this suggestion but I
honestly can't see how it can be made to work in practise.

The BI is dealing with something much more precisely calculable
and which is invariable. There is *never* any valid reason for
multiposting to the extent of exceeding the BI. Contrariwise,
there are actually quite frequent circumstances can arise when
there can be a legitimate reason for quotes>newtext. I've seen you
do it on the odd occasion. (insert smiley of choice).

It will also be argued, perhaps over-pedantically (but this is
news) that, in spite of your disclaimer, it *would* be
content-based cancelling in that it would be cancelling on the
basis of a calculation made on what the post did in fact actually
contain. That argument could run and run.

Also, it would be ridiculously easy to work round the bot by not
using recognisable quote-marks or marking quotes in any way. I'm
not a for-real programmer (pace Midnight Cowboy) so I could be
entirely wrong but I can't imagine an algorithm which would
accurately and consistantly determine what is quoted and what
isn't except by checking the References: and doing a comparison.
And that would be a doddle to subvert.

And, as someone else has pointed out, it's also veering
dangerously close to retro-moderation, which opens up another can
of worms.

Unfortunately, I think this is a case of, "Nice try but no cigar
yet".

--
DG