HP3000-L Archives

June 1995, Week 2

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:
Jim Wowchuk <[log in to unmask]>
Reply To:
Jim Wowchuk <[log in to unmask]>
Date:
Tue, 13 Jun 1995 10:20:39 +1000
Content-Type:
text/plain
Parts/Attachments:
text/plain (53 lines)
At 01:58 PM 12/6/95 PST8PDT, Bob Walker wrote:
>On our second day on 5.0 the system crashed when I ran a NMPRG from
>the shell(with a Compatiblity Mode suddendeath!).
[modus mortis snipped]
 
I've been surprised by the appearances of these problems when running under
the shell.  My own limited experience has shown problems with picking up
command-line parameters in the main() function occurring when run directly
from the shell, but not when from callci.
 
Beyond POSIX specifications, there obviously must be a set of programming
constraints, (dare I say 'paradigm'?), that must be known in order to
develop well-behaved systems.
 
Where abouts can this information be found?
 
I'm not a fan of the POSIX shell on HP3Ks, but do like the api set.  I'm
sure of any who know me, most will not be surprised at that.  But I do hope
that some of the really useful features of the P-shell, like piped commands,
will soon make their way to the MPE CI.
 
After years of exposure to the POSIX shell, starting with MPE 4.5, I liken
it to Windows 95 (and not simply because of the time elapsed since first
announcement).  Everything appears to run slower; pre-existing programs
become flaky; Sudden Death becomes a close aquaintence; and many hoped for
(hyped?) benefits just aren't there.
 
I would expect that companies porting to HP3000 will live only in the shell.
But what of the rest - the MPE vendors, MPE in-house developers and
especially HP?  Are we creating a dichotemy in the operating system, where
developers will need to choose whether they live in the P-shell or live in
CI?  Is the plan from HP designers for shell independance?  Or does the
shell, in effect, become the operating system providing a necessary base
which the programs are built on?
 
I expect to see more and more problems as people move to 5.0.  The reasons
though, I believe are *NOT* the fault of the Lab or the designers, but more
a lack of communications.  I've seen precious little on changes to
application design - such as the actions of TurboIMAGE across fork(), to
assist in recognizing the problems *BEFORE* they happen.  Is there a "Posix
Shell Migration Guide", as there was for MPE/V to MPE/XL?
 
As always, these comments are my own personal opinion.  If I've got it
completely wrong, it wouldn't be the first time, so tell please tell me. :)
 
----
Jim "seMPEr" Wowchuk           Internet:    [log in to unmask]
Vanguard Computer Services     Compu$erve:  100036,106
 _--_|\                        Post:        PO Box 18, North Ryde, NSW 2113
/      \                       Phone:       +61 (2) 888-9688
\.--.__/ <---Sydney NSW        Fax:         +61 (2) 888-3056
      v      Australia

ATOM RSS1 RSS2