walt said:
>   It might be considered broken, I suppose,
>   if someone were to take PG HTML
>   and run their own conversion on it,

so we agree that it is broken
if someone were to do that...


>   It might be considered broken, I suppose,
>   if someone were to take PG HTML
>   and run their own conversion on it,
>   in ignorance of PG's approach.

dear, i guess "in ignorance of p.g.'s approach" means
"if they don't know that p.g. does this thing incorrectly".


>   Or if they were to take PG HTML
>   and use it directly in an ebook reader.

again, again, we agree that it's broken
if someone were to go and do _that_...


>   But if they take our epub output they'll be just fine,
>   because our epub  output won't say "@media handheld"
>   once epubmaker gets done with it.

so because an error introduced early
in the process eventually gets deleted
later in the process, if it goes that way,
and gets that far, the error is not wrong?


>   And if they are running their own
>   conversions of PG books they should do so
>   with some knowledge of what they look like.

since p.g. books -- including d.p. books -- are
not produced with any dependable consistency,
it is simply impossible to accrue such knowledge.

which is one of the points we've been discussing
on this listserve for the last 9 years i've been here.

***

first it was .html proper which boasted that it would
make content "reflowable" to any device out there...

then, when that proved not to be true, the technocrats
gave us c.s.s., which would make good on the promise.

now, when c.s.s. has also proven not to be the answer,
_media_queries_ are now the new "new and improved".

and it all just keeps getting more and more complex.

-bowerbird