Moving from a Symbolics to an Explorer.

[cstacy@ai.mit.edu (Christopher C. Stacy)]

   Date: Tue, 14 Aug 90 10:29 PDT
   From: RDP@alan.kahuna.decnet.lockheed.com (Robert D. Pfeiffer)

   Many thanks for all of the useful replies on the topic of moving from a
   Symbolics to a microExplorer!

   There are a few discrepencies that I would like to clear up if anyone
   cares to comment further.

It sounds like your best bet is indeed to dump your files to some kind of
industry-standard file server (ie., a SUN) and then use tar tapes between the
the Unix boxes, as you suggest.  If you care about the file properties, you
will have some minor issues to deal with when you move from LMFS to Unix.
Depending on what file system you finally wind up keeping the files on,
you have to write a program to somehow record and restore the original
file properties.

   On the question "Will Zmail mail files be readable?":

BABYL format (which is what XMAIL files are) was designed for a mail reader
named BABYL (on ITS in TECO) that predated ZMAIL.  The format is fairly
standard, although ZMAIL introduced minor changes in the file header format.
That's the stuff at the very beginning of the file where it says BABYL
OPTIONS.  Things like Version and Append and Owner should be OK, but you may
have to change things like Mail and Summary-Window-Format.  I think that the
"Labels" field should not have to be changed, but my experience with TI's
version of ZMAIL is limited.  I don't think you will have to change the stuff
associated with the individual messages. I would just make a safety copy of
your mail BABYL file, fire up TI's ZMAIL on it, and see how it goes.
Most likely it will just ignore fields it doesn't understand.
Then save it back out, and see what it did.

       Date: Fri, 10 Aug 90 13:12 PDT
       From: snicoud@atc.boeing.com (Stephen L. Nicoud)

      ....Local variables in a frame are only accessible by

   This surprises me as I thought that this functionality existed at MIT.

The debugger having the frame's variable context was definitely a feature
added by Symbolics to the original MIT version, and I don't know if TI ever
implemented this or not.

       7) Multiple file versions weren't supported about a year ago.  I don't know if
       they are now.

   This also surprises me and seems like it may cause me all kinds of
   problems.  Anyone know if this is still the case?

Huh? I thought I saw this working on a uExplorer prototype, years back.
The original Explorer certainly had a file system, maybe not as fancy as LMFS,
which supported multiple versions.  I don't know what kind of interface they
support to the losing Unix file system, however.