Harold Fuchs wrote:
But this is exactly my point. It shouldn't be up to the program. The
program shouldn't have to "be able to handle it". The program shouldn't
know. The program should just trap the error generated by the OS and
tell the user what the OS said.
Moving stuff from one OS to another is an issue but any decent migration
guide will bring it to the implementer's attention.
But the implementer is likely to be a user who has never read any decent
migration guide, or has read it years ago, and doesn’t now recall
anything about illegal characters.
I agree that your belief that the program “shouldn’t know” is one way of
looking at this problem.
But the OpenOffice.org programming staff has chosen instead to permit
only the minimal set of characters allowed in filenames on all systems
on which OpenOffice.org runs, so that the user can be sure that whatever
file they output should run on any system to which the file is transferred.
The OpenOffice philosophy appears to be that OpenOffice ought to run
almost exactly the same on every system, which means that of necessity
that it must limit the characters allowed in filenames, which in effect
means allowing only those characters allowed on the MS-Dos/Windows
systems and perhaps even fewer characters.
Many Linux users complaint about the Unicode BOM character, for example,
though I don’t recall anyone ever complaining about this in a file name.
Also some characters, while technically allowed on a Linux system, will
break shell expansions, pipes, etc, etc. Samba has more limitations than
does Linux which in most systems only disallows the “/” character and
the NUL character.
Many Linux users complaint about the Unicode BOM character, for example,
though I don’t recall anyone ever complaining about this in a file name.
I think it more helpful to a user to disallow characters that ought not
be to be used because they sometimes cause problems, than to allow
everything, as the numerous talks about this on the web indicates that
applications producing files with illegal or problemical characters is
something that comes up again and again. The problem especially occurs
with networks that have mixed operating systems in their servers.
And somehow these files with characters that are illegal on a particular
system do get onto the drives, presumably through applications that
follow your philosophy that applications ought not to know about such
things.
At least on most systems all file names are now in the Unicode character
set which avoids many previous problems.
Jim Allan
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]