On Thu, 29 Jul 1999 12:24:51 -0400, "Stigers, Greg [And]"
<[log in to unmask]> wrote:

>X-no-Archive:yes
>Wow.
>
>We had our share of issues and problems with ftp and telnet, but I'm not
>sure that I would lay those at the feet of HP. Wirt's work with telnet, the
>RFC and its adoption for MPE, are highlights, an evidence of something good
>and robust.

Agreed ... I haven't really experienced or heard of any problems with
telnet.


>The only thing I would fault HP for is expecting other network
>components to work as advertised, and not quite outsmarting their ability to
>fail badly (the lost gateway problem seems to me to be an example of this -
>gateways are supposed to work, but not stop and then start working). We had
>to put a :PAUSE in our ftp command files, to allow our NT ftp server to
>finish flushing its buffers to disk... Whose fault would that be? Others can
>answer the question of who did the work if appropriate.

Well I would call that an MPE problem IF you don't have to do that
:PAUSE trick when ftping from other operating systems. I know I don't
have to do that if I'm sending from another NT system, a Win95 system,
or an HP-UX system. Perhaps your experience differs from mine in that
regard.

I was asking who did the work purely for informational/historical
reaosns, not to lay blame. In any case, poor software quality can
almost always be laid at the feet of poor lab testing, which most
often comes as a result of business pressures from the organization
and has nothing to do with the technical competence of the coders.

>
>Is it the experience of others that MPE networking "*sucks*"? Can someone
>from HP state how prevalent problems appear to be?

OK maybe "sucks" was too harsh and general a term. However, compared
to the rest of MPE, and compared to the implementation present in
other less capable OS's, the networking portion of MPE is substandard.
It fails to live up to my expectations that it work except in the most
pathological of situations, and that if it does fail it does so
gracefully and in such a way as to allow a ready diagnosis of the
problem.

I guess I should have phrased it that way the first time. :)