[net.eunice] Eunice info request

eric@aplvax.UUCP (01/25/84)

	I am interested in hearing people's experiences with
Eunice. Of particular importance is how transportable things
are that are developed under Unix but need to run under Eunice.
Also, Eunice is apparently up to 4.1c BSD, does the networking
work? Thanks in advance.

-- 
					eric
					...!seismo!umcp-cs!aplvax!eric

chris@mddc.UUCP (Chris Maloney) (01/26/84)

We attemped to use EUNICE (from Wollongong) to port aplications
code.  We try for more than 6 months. The problem is that
Wollongong had claim the multiple users could access a file
just as in UNIX.  The truth was they could not.

If 2 processes are working with the same file and one extends
the file (Notifing the other proccess first, of course), the second
would have to close the file to see the extentsion area.
Wollongong did admit this and attemped to fix it.  After several
attemps and 6 months they did seem to get this fix, IF you
define an ENVIROMENTAL VARIBLE to warn the system you wanted
real UNIX type files and DON'T MIND A 12 TIMES slower file
system. THE ENVIROMENTAL VARIBLE is pretty kludgy.

We could not afford this slow a system and I don't think anyone else
can either.  This is all because Wollongong uses VMS/RMS to implement
the IO stuff, instead of the VMS IO system calls.  When I tried
to explain why they should use VMS IO sys calls instead of RMS they
said that the reason they didn't use VMS IO sys calls was because they
didn't have the time to learn them.  I even gave them the code to do the
unix style IO in sys calls finally.  They didn't even have have time
to look at it because they are still to busy trying to make RMS work.
Good luck, VMS/RMS doesn't even get close to haveing the facilities
to unix style IO work fast.

This my opinion and nessarily not that of MDDC.

Please call me if you want more details.

Chris Maloney
Management Decisions Development Corp.
Fairfield, Ohio   45014
(513)874-6464

...{ucbvax,decvax,inhp4,mhuxi}!cbosgd!qusavx!mddc!chris		(uucp)
cbosgd!qusavx!mddc!chris@BERKELEY				(arpa)