Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: (ASCEND) The MAX TNT

At 06:21 PM 10/20/97 -0400, Jim Howard wrote:
>At 12:53 1997/10/20 -0700, Joseph Hickman wrote:
>>On Mon, 20 Oct 1997, Josh Beck wrote:
>>> This is a distinctly poor idea. Integrating "new features" into patch
>>> releases makes them... um, let's see... unstable incremental releases. The
>>> patch releases are already recieving too many new features as it is and
>>> aren't getting more and more stable, as they should be. I say stop adding
>>> stuff to patch releases and let things stabilize again for once!
>
>that's all good and well, however for anyone who needs these
>"features" that exist only in an incremental release,
>the NEWEST release they can run is 5.0ai13, and 
>as the FTP server shows:
>>   it was last modified on Wed Jun  4 09:54:41 1997 - 138 days ago
>that was four and a half months ago...

Watch out for details of our new software release strategy to be posted
soon. Things will change (some for the better, some with the appropriate
cons).

We did agree many months ago with those customers who said that the patch
release
branches can become too unstable and/or risky for production networks with
the continual addition of new features at a high rate turn-around. The
releases
were not fully tested, so that we could get the turn-around that was being
demanded. The request was to separate a branch from the current base (like
5.0A)
for patches *only* to provide a more secure/stable load for production
networks
with the existing feature set.

Problems that arise now are that with the addition of numerous new features in
the incremental branch, the sanity tests fail more often - so we don't release
the binaries until that's fixed (the fixes being made at the same time as
ongoing new features in this branch - that's the nature of the incremental
branch, until a features-freeze).

>I, like others, cannot run any release except recent patches
>which included the new rockwell code, and has the latest SNMP MIB.
>
>for anyone using K56flex modem cards you must use 5.0ap{24,27}
>for anyone who wants to be able to manage their MAX with NavisAccess 
>or be able to get stats on users, you can't use 56k modems (5.0ai13).

The positive thing about the patch branch is that there are no new
features that could introduce new unknowns into the code. To change that
approach now would put us back to where we were 18 months ago.


Kevin


++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>
Follow-Ups: References: