HP3000-L Archives

October 1997, Week 1

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jon Cohen <[log in to unmask]>
Reply To:
Jon Cohen <[log in to unmask]>
Date:
Wed, 1 Oct 1997 13:23:56 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (78 lines)
Ron Seybold wrote:
>
> Hello Friends:
>
> Brad Daniels asked if anybody had heard about the timing of MPE/iX release
> 6.0, and Jerry Fochtman replied that "HP's last communication on 6.0
> indicated sometime 1Q98, subject to change as always." In fact, HP's latest
> communications indicate 6.0 is due out in the *first HALF* of 1998.  Given
> the recent delay of Express 3, about which HP's Kriss Rant said "all this
> new functionality on an Express release has created some unique challenges
> for us," it seems likely that 6.0 won't be shipping during the FRONT half
> of 1998's first half. After all, some of the same people who are working on
> 6.0 are still perfecting Express 3 and Express 4 -- the latter of which HP
> still says will be ready by year's end.
>
> You may find less that's brand-new in 6.0 than in prior releases numbered
> dot-zero. HP is starting to use the Express releases as the first draft of
> its newest MPE/iX technology, then consolidating the work from Expresses in
> the next Platform Release.
>
> Perhaps our helpful colleague inside HP, Jon Cohen, can clarify the change
> in philosophy (if any) about the Express releases.
>
> Ron Seybold, Editor In Chief
> The 3000 NewsWire
> Independent Information to Maximize Your HP 3000
> [log in to unmask] http://www.3000newswire.com/newswire
> 512.331.0075

Jon, flattered by the kind mention of his name, responds:

Ron -- your summary of the situation is very accurate.  I'd only add a
minor correction:

   o  In general, there are different people working on Expresses and
      6.0, but we have a natural tendency to give the next thing out the
      door all the attention it needs, which occasionally means that the
      "further in the queue" releases may have to relinquish people
and/or
      resources temporarily.  In general, we don't have problems running
      Expresses & Mainline releases at the same time.

Regarding new functionality in Expresses versus Mainline releases:
I don't think we've made a conscious change in philosophy.  We've
listened to customers, who have told us pretty consistently that
they prefer new functionality to be retrofitted onto their
current releases and that they prefer to avoid major rolls of
the operating system if they can.  While supporting releases for
extended periods of time is difficult for us (why create a patch for
a three-year old OS when the problem is fixed in the one-year old
OS?), we strive to accomodate customers' wishes.  We tend to support
MPE releases for four or five years, which I think is pretty good.  If
a customer so chooses, they can update to every other mainline release
and only have to update once every five or six years or so.

However, retrofitting new functionality onto Expresses does have some
challenges for us.  One primary problem is that not all new
functionality
is "Express-able" -- for example, some low level code (like new I/O
systems or many performance enhancements) is very difficult or
impossible
to submit to an Express; other changes that are pervasive throughout the
OS code also cannot be submitted to Expresses.  Finally, after a few
years of Expresses, our patching processes can get overstrained.  For
these reasons, we plan to continue to release major "mainline" releases,
like 6.0, every 18-30 months or so.  Mainline releases give us a chance
to clean everything up and allow us to lay a foundation for a whole new
set of enhancements (i.e. Expresses).


--
Jon Cohen
[log in to unmask]
Hewlett Packard -- Commercial Systems Division

(The opinions expressed here gotta be mine, 'cause no
one else will own up to 'em.)

ATOM RSS1 RSS2