> If you can recognise what constitues a 'character' (trivial
> on paper tape
> or 9 track magtape, less obvious on bit-serial media like
> floppy disks),
> and if you assemy there will still be enough ASCII-encoded
> text around
> that some of it is in a normal human language, then the
> problem seems to
> reduce to solving a monoalphabetic substitution cipher, which is
> generally not too hard....
I don't doubt that it's doable given the motivation, a
modicum of knowledge and some extant text. My point was
that it is hardly what I would call self-describing.
Right now ASCII is so deeply engrained into us all that
the idea that we might forget how it works seems utterly
ludicrous. A few hundred years from now that may change.
That's probably true of ordinary written language too
(every time I see examples of medieval text they look
quite "foreign" to me - difficul to make out the
individual letters because of the unfamiliar style).
So I guess I'm saying that PDF is not much worse than
ASCII in the absence of a specification. Given a
speciication then it really is no worse (admittedly the
initial PDF reader implementation is likely to take
longer to create than the initial ASCII reader ...
but that's a one-off cost).
Antonio
--
---------------
Antonio Carlini arcarlini_at_iee.org
Received on Sun Feb 06 2005 - 11:22:14 GMT