<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 10-May-15 21:35, Clem Cole wrote:<br>
<blockquote
cite="mid:CAC20D2NzMwWe0On890AcCXX=3f8E=jTTgjxELdBf6d_beHhtxg@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra"><br>
<div class="gmail_quote">On Sun, May 10, 2015 at 5:47 PM,
Timothe Litt <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:litt@ieee.org" target="_blank">litt@ieee.org</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex"> It
might be Ultrix CDFS - anyone know that format/a linux
tool?</blockquote>
</div>
<br>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif">This question
is causing some random bits in my memory FS to slowly be
retrieved but have not been active for 20+ years I fear.
;-)</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif"><br>
</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif">My memory is
that Ultrix CDFS was strange for a UNIX FS or CDs and some
of what you describe sounds about right. IIRC Ultrix used
512 bytes for the block size, when "normal" CDs used 1024.
We had a funky Unix system in ZK which we used to write the
CD masters for ultrix and VMS - I want to stay it named
Young something (youngbloods so some such - I've
forgotten). Again, Armando may remember - but I think Fed
Cantor did hte interfacing,</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif"><br>
</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif">I remember,
you had to have very special media to write them, in those
days since the CD writers were not forgiving. It took a
source media created a CD image on a raw disk first with
whatever parameters you were using for the CD like block
size, then did a bit map copy from the disk to the CD and
IIRC correctly we used an funky ANSI tape program to write
the masters.</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif"><br>
</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif">Standard ISO
CD were's 1024 bytes, but could not handle Unix or VMS style
file attributes. So Ultrix did not use the ISO format, as
folk has support for for full case, extended file attribute
etc of UFS etc, i.e it was before rockridge extensions were
standardized for ISO. By the time of Tru64, Rockridge was
standardized for the Unix community (and ISO) and I think
that was some how used, but I pretty sure Ultrix did not. </div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif"><br>
</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif">But I do
think the DEC tools of time to write most CD's were based on
Ultrix which interfaced to the box in ZK. I looked for
some docs on the ultrix format, I had at one time, but can
not find them. I'll keep looking, but I fear I lost much of
that in the flood in my old condo about 25 years ago. I'm
going to have dinner with another pack rat tomorrow night
and I will ask him if he thinks still has any of the Ultrix
docs stuffed away. If I can find it, I'll try get it
scanned and pass it on.</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif"><br>
</div>
<div class="gmail_default"
style="font-family:arial,helvetica,sans-serif">Clem</div>
<br>
</div>
</div>
</blockquote>
Thanks.<br>
<br>
That all sounds about right. Except that the standard sector size
for CDROM is 2048 (usable, after EDC).<br>
There's also mode 2, which is 2336 usable, with no EDC. DEC wouldn't
have used that.<br>
<br>
VMS read the 2048 byte sectors, and the device driver mapped the 512
byte LBNs to the right fractional sector. <br>
VMS also could put its (FILES-11/ODS) file structure on a CD -
actually many Pathworks & DOC CDs were dual-format; <br>
Both the ISO9660 & FILES-11 structures were present, and because
the mount processes look for the filesystem <br>
root in different places, everything just works. Perhaps Ultrix
does the same thing to present 512-byte LBNs.<br>
I suppose it might have written true 512 byte sectors - but I'm not
sure that works with the readers.<br>
<br>
We (my 36-bit hat) probably just wrote the ANSI tapes & sent
them to ZK.<br>
<br>
I wonder if what I have is a text header (perhaps for the label?) +
data in "compress" format. LZW would explain the<br>
occasional few bytes of cleartext. So maybe one dd's the header off
to one file, then dd's the rest of the file thru<br>
compress? The ASCII hex "4b9c4" is 309700.; perhaps the size of the
CD data? If in 2K sectors, that would be<br>
~634MB, or about the size of a CD. The extracted binary (assuming I
got that right) is 171M, which would be ~3.5:1 <br>
compression.<br>
<br>
Hmmm, the last cleartext is at 7f6, then at 800 the binary
starts.... If I dd skip=4, the resulting file isn't<br>
a recognizable format. (and so not compress.) But it this feels
like the right track. <br>
<br>
In any case, if all this is correct, the format of the "funky ANSI
tape" is also interesting, as I guess<br>
that's what I have.<br>
<br>
Not that archeology isn't fun.<br>
<br>
Starting at 0x800, the first few bytes are (non-ascii):<br>
<br>
0000000 25c10d04 120057c0 01140462 40430009<br>
0000020 54041011 75500401 040800dc 005c0196<br>
0000040 010a0020 00807005 150c4240 001404c6<br>
0000060 6001709f 1d000287 27c07cc0 102a00dc<br>
<br>
<pre class="moz-signature" cols="72">This communication may not represent my employer's views,
if any, on the matters discussed.
</pre>
<br>
</body>
</html>