loess	released by AT&T, Bellcore
From trevor Mon Jul 11 15:27:25 EDT 1988
To: ehg
Subject: GAIM

It is in reasonable running order, although 4 years old now and slightly out of date with how we would do it today. Nevertheless, we do still send it
out to brave applicants.
Do you want me to email you a copy?

From trevor Tue Jul 26 13:18:15 EDT 1988
To: ehg
Subject: GAIM

It is not released, since I wrote it before coming to the labs. I also 
included a disclosure in my contract, so I dont think there is a problem.

The list of files (on gauss) that are needed are in
gauss/usr/trevor/gaim/package/flist 

If you cannot get into gauss, I can send you a mail file with them bundled up.
Trevor

From wsc Tue Oct 11 00:00:40 EDT 1988
To: ehg
Subject: stl

Yes, it's all set to go and /n/gauss/wsc/software/stl has the official
source.
From wsc Fri Dec  9 15:25:48 EST 1988
To: research!ehg
Subject: files

The file and 3 directories are in /usr/wsc/software and are README, stl, loess
and lowess.

From wsc Fri Dec  9 17:51:12 EST 1988
To: research!ehg
Subject: loess

What message does one send now to get loess from netlib. Are you distributing
both the single and double. Could you replace the old stl stuff with the new
when you pick it up for the floppy. I assume the request is still 
send stl from a. I began to wonder if you want an s for statistics. Perhaps
you do not want to pollute netlib.

From nls Thu Jul  6 21:51:52 EDT 1989
sure, go ahead.
>From ehg Thu Jul  6 20:47:57 EDT 1989
Is it ok to put the SSAF menagerie in netlib without the
formality of a release?  We're only talking about a
91 line Fortran function.


.