Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ihnp4!ptsfa!ames!rutgers!cbmvax!vu-vlsi!UUCP
From: UUCP@vu-vlsi.UUCP
Newsgroups: comp.sys.tandy
Subject: Tree Directories For The TRS-80
Message-ID: <867@vu-vlsi.UUCP>
Date: Wed, 3-Jun-87 20:08:08 EDT
Article-I.D.: vu-vlsi.867
Posted: Wed Jun  3 20:08:08 1987
Date-Received: Sat, 6-Jun-87 10:47:08 EDT
Sender: UUCP@vu-vlsi.UUCP
Lines: 39

This discussion on the idea of implementing a tree-type directory
structure on the TRS-80 has kind of gotten under my skin.  I'm going to
talk about this as working under LDOS 5.3/TRSDOS 6.3, since that's the
existing directory structure I know best.  Note, however, that I am making
**NO** claim to the ability to IMPLEMENT this system; just speculating on
how such a system would work.

A directory would be allowed a 7-character directory name.  It would hold
a slot in the root directory with the name convention that the LAST
character of the directory name will be "D", and the extension will be
"/SYS".  These files will also have the system attribute bit set.

The filename convention will be as follows:
filename/ext.pathname:drive.  The segments of the pathname will be
separated by periods, eg.: LETTER/TXT.LETTERS.PRIVATE.MOM:1.

The protection level for ALL files will be set at FULL, with the exception
of system files in the root directory.  The owner password area will be
used for a pointer to the directory level to which the file belongs
(probably the Directory Entry Code in the HIT table).  So, the file
mentioned above, "LETTER/TXT", would have a pointer to the directory
"MOMD/SYS", which would point to "PRIVATED/SYS", which would point to
"LETTERSD/SYS", which would, as a convention, point to "DIR/SYS" to
terminate the chain.

The advantage to this system, as I see it, is that it allows the resulting
disk to be read under an un-modified DOS, while allowing a directory tree
structure that is very extensible.

The only problem I see with this idea is the fact that it consumes
directory entries for the subdirectories.  However, I personally find it
RARE that a floppy disk ever uses anything approaching all of its
directory entries.  On a hard disk, this might become a more pressing
problem, but that can be dealt with by using as many partitions as
possible, and/or the diskDISK program available from Misosys.


I invite commentary on this idea; I realize that it's pretty preliminary,
but I thought it might get the ball rolling.