home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.acorn.tech
- Path: sparky!uunet!pipex!warwick!pavo.csi.cam.ac.uk!idg10
- From: idg10@cl.cam.ac.uk (I.D. Griffiths)
- Subject: Re: EPS files
- Message-ID: <1993Jan21.115246.17905@infodev.cam.ac.uk>
- Sender: news@infodev.cam.ac.uk (USENET news)
- Nntp-Posting-Host: ufford.cl.cam.ac.uk
- Reply-To: idg10@phx.cam.ac.uk
- Organization: U of Cambridge Computer Lab, UK
- References: <ARM200-921214124010-52326667*@MHS <1992Dec23.155806.22223@gec-rl-hrc.co.uk> <1993Jan12.111800.15139@infodev.cam.ac.uk> <1993Jan18.102504.18587@gate.esat.kuleuven.ac.be>
- Date: Thu, 21 Jan 1993 11:52:46 GMT
- Lines: 47
-
- In article <1993Jan18.102504.18587@gate.esat.kuleuven.ac.be>, tytgat@esat.kuleuven.ac.be (John Tytgat) writes:
- |> >also be rather hard, since you would need to have a postscript interpretter,
- |> >which you don't need to read EPS files (since they're not really programs, but
- |> >a postscript representation of a data structure), and then you would have
- |> >problems with programs possibly not terminating, and it all gets a bit
- |> >horrid.
- |>
- |> Tsj, tsj, tsj. PS is a real graphics language, so PS and EPS files can be
- |> considered as programs (if all the procedures are defined). And what do you
- |> mean with "possibly not terminating programs" ???
- |>
-
-
- It's possible to prove that one cannot generally prove that programs will
- terminate, but that was done by Alan Turing a while back so I won't bother
- with details.
-
- To give a useful example though, there is a raytracing package written in
- Postscript, popular amongst students for annoying departments' laser
- printing services, since they typically take 3 hours to render a page, and
- are rather hard to stop neatly since flushing the queue doesn't work -
- the page has already reached the printer by the time you work out
- something has gone wrong, and turning the printer off and on doesn't
- go down well with queuing software. (OK, so you can get round this, but
- that's not really relevant to the topic.)
-
- An importer to convert this into, say, an Artworks file would take quite
- some time to execute, and it's not impossible to conceive of a scene that
- would never finish rendering (just have zero loss reflections) so what
- does the import do?
-
- It is precisely BECAUSE Postscript is a real language that you get problems.
-
- On the other hand you'll only get these termination problems with stuff that
- wouldn't print anyway.
-
- However I am dubious as to the extent to which one can usefully interpret it
- into a data structure, since it is a real language. How does one deduce
- that a function is, say, a graduated fill?
-
-
- As for your comment that CC have the in house technology to do this, do
- they? Oh. If you refer to the code in Showpage, that's rather different,
- it's just interpretting postscript, and producing a bitmap, and the importers
- on Artworks rely on known prologs to do conversions.
-
- Ian
-