OPENMPE Archives

December 2002

OPENMPE@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:
Tracy Pierce <[log in to unmask]>
Reply To:
Tracy Pierce <[log in to unmask]>
Date:
Fri, 13 Dec 2002 08:09:35 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (77 lines)
the 3rd party stuff was listed by Jeff Vance as being


  - MKS for their POSIX shell and Utilities,
  - Mentat for their MPS Streams code needed for pipes and some
    other IPC functionality. MPS code is used by: 12H Raid Mgr,
    XP Raid Mgr, POSIX pipes [pipe(), HPPIPE()], Predictive, CSTM,
    and Patch/iX.
  - MB Foster for their 32 bit ODBC driver.

which yes, includes some important stuff.

POSIX shell and Utilities: Linux.
pipes, IPC: Linux.
12H Raid Mgr, XP Raid Mgr: integrated to MPE, or added on?  (treat the disk
as disk; raid is a way to make disk more reliable, and it's available for
Linux systems.)
Predictive?  Who's going to support Predictive?  sell the pieces with
Predictive ;-)
Patch/iX?  there ARE non-MPE OS update procedures out there.
ok, I'm skipping CSTM.

isn't all this stuff (except CSTM and Predictive) already all there in
Linux?

Tracy Pierce

> -----Original Message-----
> From: Mark Klein [mailto:[log in to unmask]]
> Sent: Friday, December 13, 2002 7:49 AM
> To: Tracy Pierce; [log in to unmask]
> Subject: Re: MPE Licensing
>
>
> On 13 Dec 2002 at 7:28, Tracy Pierce wrote:
>
> > re the 1st two, IF the emulator were software designed to run atop
> > linux, all ux stuff could be just left out of OpenMPE, and ux calls
> > could be as simply invoked as is MPE's HPCICOMMAND.  Voila, no 3rd
> > party fees.
>
> That's an over simplication, because this stuff isn't just bolted on -
>  it is integral to MPE at this point. Further, one piece of the 3rd
> party code handles some of the inter process communication internal
> to the OS. Can't simply "leave it out".
>
MPE worked before integration of Posix was even a dream; my biggest hurdle
re that integration was always "Who's on top, Posix or MPE?" and the
answer's of course "Neither".  But why?  Absolutely necessary?

> > Granted that's a huge IF, and quite contrary to the hardware emulation
> > touted by the OS experts here, but if MPE emulation amounted to
> > intrinsics written to run atop Linux, that OS's continuously improved
> > features would be pretty durn near.
>
> Gotta keep in mind again that there'e "emulation" and "simulation".
> What you are describing is simulation. There are many companies
> already out there today with MPE simulation tools. Since you state
> that you don't care about recompiling your application, the
> simulation is probably a better fit for you. There are those where
> that isn't possible, or at least cost justifiable where emulation
> works out better for them.

If somebody has a comprehensive simulator, who where what and how much?  If
such a product exists, the only reason I'd have not to jump to it is the
fact that e3k hardware is now so dirt-cheap and getting cheaper on a nearly
daily basis.  As long as I can transfer licenses (which will be as long as
HP hasn't got hold of and destroyed the boxes to which those licenses are
attached), I can wait to use the simulator.

I'm sure there are examples of sourceless object code out there, but these
are fires just waiting for nearby oxygen.  And again, cost justifiable?
Who's going to fund hardware emulation for an OS that's incapable of being
upgraded?

Tracy

ATOM RSS1 RSS2