Wirt Atmar <[log in to unmask]> wrote:


> Ken writes:
> When I wrote "organization," I meant vendor organizations, not individual
> users :-).
>
> While there has been a gold rush mentality of late to port all of the UNIX
> protocols over onto the HP3000 as quickly as possible, this phase is just
> about over. People aren't generating new protocols very quickly nowadays
and
> the HP3000 has, or is the process of obtaining, all of them. If anyone
really
> wanted to make the HP3000 completely compatible with the modern networked
> world and safe at the same time, taking the time to rewrite these
protocols
> from scratch, using only the specifications, would nearly guarantee its
> safety.

I realized that's what you meant but it's still not going to happen.  Even
an organization as big as Microsoft does not write its protocol
implementations from scratch.  For TCP/IP, they used Berkeley code.
Everybody uses some kind of reference implemenation as a base.

People ARE generating new protocols all the time, at all levels of the
protocol stack, and there are security issues involved with each level.
There are new low levels such as SCTP (on the same level as TCP and UDP),
new mid-levels such as BEEP, and dozens or hundreds of higher levels.  I may
well use BEEP in some projects, but *only* because I can use a portable
implementation.  The HP3000 only has most of its protocols because portable
implemenations exist.

If you write your own implementations of protcols, there will, of course, be
more bugs tha the standard implemenation.  The only hope is that your bugs
will remain obscure because your operating system is unpopular.  Is that the
future you want for the HP3000?

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *