HTML "transition" (was Re: [ih] spam...)

[Perry E. Metzger]

Dave Crocker <dhc2 at dcrocker.net> writes:
> However, the point was made that the heuristic of filtering HTML does
> raise the count of false positives.  This is to be expected, since the
> correlation between HTML and spam is merely a transition effect, as
> HTML becomes more commonly used.
> 
> Imagine the transition from uppercase only ASCII to upper-lower case
> ASCII.  No doubt there were interesting usage correlations during that
> transition, too, but it would be a mistake to believe the correlation
> would hold true longer term.

An interesting question, away from the spam discussion, is whether or
not we are "transitioning" to HTML as the preferred email carrier.

Now, it is true that a lot of HTML email is going around out there,
but so are a lot of microsoft .doc files and I doubt that .doc will be
anything but a memory in 20 or 25 years. Experience says that no one
will have tools that will read .doc from now with anything like
original fidelity. Therefore, mere volume isn't entirely germane.

A lot of the HTML email going around these days isn't particularly
good or clean HTML, and often has things like external links -- which
are likely quite transient -- added in, frequently for nefarious
purposes like detecting the reading of the mail. For the most part,
sophisticated HTML email is unpleasant to compose, and simple HTML
email is usually conveyed as well as ASCII.

One might therefore ask whether or not ASCII email is, as is often the
case, an improvement over its successors as well as its predecessors.

-- 
Perry E. Metzger		perry at piermont.com