This  documentation  file  accompanying  PA0GRI  NOS  version  910531v1.6h 
(N1BEE-910608v0.1)   is  prepared  by  Mike  Bilow,  N1BEE,  who  takes   sole 
responsibility for its contents.

    I. Things to watch out for

    The  default  startup file is AUTOEXEC.NOS, which is  different  from  the 
usual  naming convention used by KA9Q and G1EMM.  This can be overridden  from 
the command line. 

    PA0GRI  NOS  requires  a  period  following  the  domain  suffix  argument 
("ampr.org.") in the AUTOEXEC.NOS file.  Again, this is different from G1EMM.

    The  "mail"  command  invokes BM.EXE by default;  the  MAILER  environment 
variable  is tested and will override this with an alternative mailer name  if 
you want to do that.

    This NOS is very large.  Part of that is just because I elected to compile 
in  most  of  the standard options so that those interested  in  putting  this 
software  through its paces could do so.  I strongly recommend including  "mem 
eff  on" as the first command in the AUTOEXEC.NOS file; otherwise the  program 
will probably crash.

    II. Known bugs

    The  bug  which  prevented the "param" command from  working  at  all  was 
repaired  by  me,  using  exactly the code suggested  by  WD5HLS  in  the  NOS 
developers conference on the Internet yesterday.

    The  DRSI drivers work as badly in PA0GRI as they do everywhere else.   In 
particular,  all  outgoing frames containing data show a checksum  error  when 
traced.   The frames are being transmitted correctly, however.  This error  is 
not  the fault of PA0GRI, who is using the DRSI drivers as supplied.  I  don't 
expect the asy drivers to be as buggy, but I have no way of testing them at my 
station, which runs exclusively DRSI hardware.

    III. General warnings

    This  is brand new code.  Look upon it as an experiment.  To make  matters 
worse, I had to modify some things in order to get it to compile using Borland 
C++  2.0  (in standard C mode, of course).  Some of these  things  are  fairly 
important,  such as storage classes and function calling conventions.  It is a
great deal cleaner than the G1EMM source, and I have a lot more confidence  in
it for that reason.

    Although  this is something of the "deluxe" model with everything  stuffed 
into  it, including NETROM, POP, RIP, RSPF, NNTP, and who knows what  else,  I 
can't really attest to any of it actually working correctly.  I have had it on 
the  air and it succeeded in sending and receiving mail, but the first  useful 
thing I am going to be doing with it will probably be making it FTP itself  to 
the WB6NIL switch!

    IV. Source code

    I  plan  to  post  the  source  code,  both  the  original  and  with   my 
modifications,  to  the  ChowdaNet landline BBS,  (401)331-0334.   The  ZIPped 
source  is  about 600K.  I do not think it is appropriate to  allow  something 
this large to go out via FTP.

    V. Feedback

    I would be interested in hearing any reports from users of this NOS,  even 
on a trial basis.  Bug reports and so forth are expected.

-- Mike Bilow, 08 June 91, 1000z

   n1bee%n1bee.ampr.org@w1cg.ampr.org  (SMTP)
   N1BEE @ KA1RCI.RI.USA.NA  (AX.25)
   mikebw@idsvax.ids.com  (Internet)