> --- Eric Dittman <dittman_at_dittman.net> appears to have written:
> > What I'd like to see is a community effort to make a new OS that
> > does what needs to be done without worrying about being compatible
> > with something that currently exists.... Sure, it may not have a
> > lot of software available...
Hmmm... my only concern about compatibility would be that files could
be transferred to and from other systems, and that ASCII could be used
to make the exchange of e-mail, Usenet news, etc. possible. I won't
mention a desire for TCP/IP, as, perhaps someone might come up with an
idea (as someone may already have done in the past) for superior
networking.
Regarding file transfers, perhaps a UFUTS filename universal
translation system (UFUTS) could be used so that u don't spend too
much time futzing with file names after copying them; that is, so no
one would have to spend time STUFUing over filenames after copying
them.
While kermit often does a reasonable job, there can be problems when
multiple files have to be translated, and when copied back to the
original system from another system that, let's say, supports fewer
filename characters, etc., the name will not be what it originally was.
So, perhaps each system could contain a translated filenames database
that also contains a special datafile for each system that it exchanges
files with that uses different naming standards.
I guess that's still not too clear of an explanation, so, please allow
me to expand upon that. Let's say that system A allows only the
ASCII alphanumeric characters in filenames, but allows 1024 characters in
a filename. System B allows no more than 10 characters in a filename,
yet, it allows many different graphical images to be used as special
characters in filenames. Now them, let's suppose that system A knows
about the special characters used with system B, and system B knows about
the lengthy filenames used with system A, for the following reason: When
the file transfer first began, systems A and B checked to see if the
UFUTS database contained a special datafile for the others system/OS type.
If that file didn't exist on either system, then it would be transfered
to the system, or systems, needing it, prior to the transfer of the file,
or files, that need to be transfered.
The above is, perhaps, a great oversimplification of my idea, however,
using this basic sort of technique, files could hypothetically be
transferred from system to system, and, when transferred back to the
system of the type on which they originated, the file names would
automagically become the original filenames that they started out life
as.
This may be of particular use to users of classic systems who want to
archive and preserve many different types of files on a live
filesystem for safe-keeping over the years, as well as of interest to
those using emulators.
After someone mentioned PERQs, this idea came to me after I began
thinking about problems that I had when attempting to transfer large
numbers of PERQ files to a UNIX system via kermit. Also, if this file
translation and transfer mechanism could also recurse directories, and
transfer the knowledge of the directory structures between the
systems, as it would with the filename format information, it could be
even more useful, allowing for entire hard disks to be archived on
other "non-compatible" systems.
What are the thoughts of others regarding this idea and UFUTS?
--
Copyright (C) 2001 R. D. Davis "Golf course" properly defined: a horrid
All Rights Reserved misuse of good pasture and farm land often
rdd_at_rddavis.net 410-744-4900 perpetuated by those who are capable of
http://www.rddavis.net neither farming nor riding.
Received on Thu Jul 26 2001 - 23:36:22 BST