Jungshik Shin wrote on 2002-10-14 06:37 UTC:
> When I made a patch, I wrote to the maintainer of cdrtools,
> but his response was not so positive. At first, he asked me whether
> iconv(3) is available on any platform other than Solaris. After I replied
> that iconv(3) is a standard API specified in Single Unix Spec and that
> Glibc 2.2.x has had it for a few years and Bruno's implementation
> of iconv(3) in libiconv is widely available and had been ported to
> virtually all platforms, he didn't reply.
Try again, he's just busy. (I interacted on this with him as well)
> He eventually wanted to move
> onto a more generic format (for DVD and similar media) whose name is
> currently escaping me.
He means ISO 13346 and its profile UDF 2.01.
Background:
ISO and ECMA have already published the successor of ISO 9660 called
ISO 13346:1999, which supports all the things currently missing in
ISO 9660, namely long file names and being designed for random access
write:
Volume and File Structure for Write-Once and Rewritable Media
using Non-Sequential Recording for Information Interchange,
ECMA-167, 1997-06
ftp://ftp.ecma.ch/ecma-st/Ecma-167.pdf
ISO 13346 = ECMA-167 is designed such that the same CD or DVD can
comply to both ISO 9660 and ISO 13346 simultaneously if needed, so there
is a smooth migration path between the standards.
ISO 13346 is the official file system to be used on DVDs, using the
following profile (= parameter subset):
Universal Disk Format Specification, Revision 2.01, 2000-03-15.
http://www.osta.org/specs/
http://www.osta.org/specs/pdf/udf201.pdf
However, as apart from DVD players, operating systems do not support it
yet widely, common practice of users of DVD-R drives is still to
continue using the so-much-hated MS-DOS era ISO 9660 standard.
I think ISO 13346 is in principle a good standard. One header field
contains the ISO 2022 sequence that defines for all filenames on the
disk what the encoding is. So the obvious thing to do is to write
ESC % G into that header field and then use UTF-8 everywhere.
Unfortunately, the authors of the UDF spec seemed to have been
unfamiliar with UTF-8, so they invented their own two ESC sequences in
order to switch between ISO 8859-1 and UCS-2 in the file names. :-(
If someone out there has lots of spare time, it would be a neat project
to build an iso13346mkfs (which could include all of mkisofs), as well
as a kernel driver.
In case you do a kernel driver, please support in addition to the file
name encodings required by UDF also UTF-8 with ESC % G. Then ISO 13346
could be the ideal DVD-R and CD-ROM file system standard for Linux.
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/linux-utf8/
