<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>
Date   : Thu, 07 Nov 2002 15:28:26 -0800
From   : Angus Duggan <angus.duggan@...>
Subject: Re: Which format do you want BBC manuals in? RTF/HTML

Colin writes:
>As a few of us have said, although Acrobat is the usual free reader, there
>are a few PDF writers available for free on the web - Ghostscript and PDF995
>which have both been mentioned here. Nobody seems to have commented on
>them - is there a particular reason that this free stuff is no good?

As somebody who has working exclusively on PostScript and PDF RIPs and
technology for the last 9 years, I feel qualified to answer this. It's not
that they are "no good", it's that they don't do as much of the job, or as
well, or as fast, as the commercial offerings. They work fine for some jobs,
but are not as sophisticated or complete.

The core of the PostScript language and PDF format are quite simple, but
there are a huge number of additional specifications and ambiguities that
need to be dealt with. (For instance, I recently tracked down and read over
40 documents providing significant information about the font formats used by
PostScript and PDF.) Also, in the real-world, complying with a specification
doesn't mean your PostScript interpreter/PDF reader will work. There are many
applications and drivers that have produced incorrect output that worked on a
particular device, and these require compatibilty hacks. In commercial
software, there are challenges with high data rates, perceptual problems
(such as colour matching and smoothing), trapping, and extra requirements for
customers with specific workflow problems.

PDF is a suitable format for archival of visible representations, because it
is compact, scalable, data is extractable (not as easily as source form,
though), and the specification is openly published. Explicit permission is
given to use some of Adobe's patents in writing compliant software.

The latest versions of PDF have structures for metadata, so you can map the
multiple contents streams spread across the pages to articles, you can
search, re-flow, and extract information more easily.

a.
<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>