Thomas Chan wrote on 2001-02-03 19:19 UTC:
> > [ISO 10646-1:2000] Available from
> > http://www.iso.ch/cate/d29819.html
> > in PDF on CD-ROM for just 80 CHF (~45 USD).
>
> I thought there were problems opening some of those PDF files; c.f.
> http://anubis.dkuug.dk/jtc1/sc2/wg2/docs/n2296r.htm .
I bought the ISO 10646-1:2000 PDF CD-ROM and I had no problems with
opening any of the PDF files on it.
The problem described on the URL above is just related to the fact that
the filenames are not necessarily most convenient for users of systems
such as MacOS and MS-DOS 4.0 that don't support Microsoft's Joliet
extension of ISO 9660. They look as garbled as a long Win32 VFAT
filename looks with an old FAT operating system such as MS-DOS 4.0.
There is no problem under Windows95/etc. or Linux and any technically
skilled user can surely circumvent the problem trivially under MacOS and
MS-DOS.
I agree however with Michael Everson's statement "It is curious that the
CD-ROM published by ITTF uses a non-ISO-9660 file system for
distribution of its files." Well, ITTF uses actually ISO 9660, but with
a backwards compatible Microsoft extension that makes the filenames
inconvenient to read on pure ISO 9660 drivers. I agree that ISO/ITTF
should definitely be more enthusiastic and knowledgeable about actually
making good use of ISO's own standards in their own products. ISO 9660
level 2 provides for up to 32 characters long filenames and ISO 13346
provides for arbitrary length UCS filenames that can coexist with
backwards compatible ISO 13346 filenames. ISO should really give the
people in charge of preparing CD-ROMs free copies of the ISO standards
relevant to CD-ROMs!
If I had been in charge of the final production of this CD-ROM, then
a) I would have used meaningful filenames that fitted into the
ISO 9660 level 1 8.3 [A-Z0-9_] format for maximum portability.
b) There might have been additional UDF/ISO 13346 (and perhaps even
Rockridge and Joliet) sectors to provide nicer filenames on systems
that support it.
c) I would actually have placed the entire ISO 10646-1 document into
one single PDF file. PDF (unlike PostScript) was designed to handle huge
files efficiently, PDF merging and splitting tools such as
Acrobat Exchange are widely available, and splitting up sections
into several PDF files has neither a usability nor a performance
advantage IMHO. In PDF, pages are compressed individually and can
be accessed directly via a stored table-of-contents data structure.
A single file named "10646_1.PDF" under ISO 9660 and named at the same
time "ISO 10646-1:2000(E).pdf" under the parallel ISO 13346 directory
structure would have seemed to be the obvious and appropriate thing to
do here instead.
But perhaps ISO's central secretariat follows just the common industry
practice pioneered by Microsoft: "We will get it right in the third
release."
Markus
--
Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at acm.org, WWW: <http://www.cl.cam.ac.uk/~mgk25/>
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/lists/
