At 5:57 PM 11/18/94, lee(_at_)sq(_dot_)com wrote:
I think this is reasonable, actually, but I am not sure how to implement it
in a UI for the MUA -- how do you decide if this document is readable?
There are SGML documents that tag literally every word, marking part of
speech, for example. There are also SGML documents that have only the
occasional <Head> or <p> in them.
It seems to me that you've hit on a very decent blunt-cut tool; a markup
vs. content metric.
This could be expressed in terms of (bytes of markup)/(bytes of content),
(instances of markup)/(words of content), (instances of markup)/(lines of
content), or all three.
rules could include something like:
If more than 10% of the bytes in the document are markup commands, it
belongs in application.
I'm not saying this would be the *only* guideline, but it's reasonably easy
to compute and *might* be able to do a reasonable job at assessing files.
The other angle is that tools could be written that deliberately produce
"readable" output, and then you wouldn't have to decide after the fact.
--
Steve Dorner, Qualcomm Incorporated. "Oog make mission statement."