>>>>>> about 200 lines <<<<<<
As per previous message I posted a few days ago,
Dick Kranz has graciously agreed to publish the latest
version of the "C++ on MPE" survey I have been working
on in the next issue of InterexPRESS.
I've gotten almost as many suggestions for changes and
additions to the survey as I have responses so far, so the
latest version is considerably changed and expanded
(again) from what I last posted. Thanks to all who
responded; the survey got a lot better with your help.
But now I have *got* to do some other stuff, so no more
changes unless somebody else does them. Note that
HP has informally blessed the contents of this version.
Summary of all responses to original post of earlier
version of the survey coming soon......... really......
Last suggestion: The C++ survey Interex will be
publishing (which follows verbatim) has major changes
and additions. So if those of you who were good enough
to respond to the original wanted to "vote" again by email
to Interex and answer the additional questions, I can only
say: In the best traditions of Chicago politics, "vote early
and often". But be assured your original responses will
also be sent forward..............
:-) ,
Ken Sletten
================================
email: [log in to unmask]
smail: Ken Sletten
NUWC Division Keyport
Code 3311, Building 894
Keyport, WA 98345-7610
tel: 360-396-2525 fax: 360-396-7861
================================
>>>> From here down is final of what we requested
>>>> appear in the next issue of InterexPRESS.......
=================================
The need for a good C++ compiler on the HP3000 was
one of the hot topics at the IPROF-95 HP management
roundtable. The two options which generated the most
discussion were:
(a) Port the HP-UX C++ compiler to MPE/iX.
(b) Port the Free Software Foundation (FSF) GNU C++.
HP said they have "no current plans" to port the HP-UX
C++ compiler to MPE/iX. But they do plan to facilitate a
port of the FSF GNU C++ product. In addition, HP has
said that users will be able to purchase HP 3000
software support for GNU C++ directly from HP.
If you need C++ on the HP3000; or if you want to
see the software development community be able
to create/port C++ MPE/iX applications that you
would be able to buy; then please fill out and return
the following survey to Interex . Results will be
forwarded to HP. You can:
fax to: 408-747-0947,
attention: Interex Advocacy Operations Group
or email to: [log in to unmask]
or mail to: Interex
Advocacy Operations Group
1192 Borregas Ave
Sunnyvale, CA 94088-3439
-------------------------------------------------------------------
**** C++ on MPE/iX: Requirements and Options ****
1) Where would an MPE/iX C++ compiler be the
most useful to your site (circle one or two choices):
a) Porting applications from HP-UX.
b) Porting applications from other UNIX platforms.
c) Developing new C++ applications from scratch.
2) When do you expect to need a C++ compiler on
MPE/iX (circle one):
NOW 3 months 6 months 1 year 2 years
3) What is most important to your site (circle one):
a) Getting C++ on the 3000 as quickly as possible,
even if not all the desired MPE-unique features
are incorporated in the initial release.
b) Fully integrating C++ with MPE before releasing
the product (i.e.: direct support of MPE-unique
features, maximizing compiler optimization,
perhaps HP Loader support, etc.).
4) What is your current perception of the HP-UX
C++ compiler, as compared to the GNU C++ product:
a) Porting HP-UX C++ to MPE/iX is the best choice.
b) One has no real advantages over the other.
c) Porting GNU C++ to MPE/iX is the best choice.
d) Don't know enough to have an informed opinion.
5) If a C++ compiler were enhanced to support MPE-
unique features, what capabilities would it need ??
Circle "NS" (not sure), "DC" (don't care), "US" (useful),
"NE" (need eventually), or "PIP" (must do as *part of
initial port*) for each:
(a) Direct support for intrinsics and missing parameters
(example: #pragma intrinsic HPFOPEN). Note that
intrinsics could still be accessed without direct C++
support through a separate set of interface routines:
NS DC US NE PIP
(b) If direct intrinsic support was not provided initially,
would access to Image/SQL through an independent
class library change your answer to 5)-(a) ?
YES NO
(c) If direct intrinsic support was not provided initially,
would access to KSAM through an independent
class library change your answer to 5)-(a) ?
YES NO
(d) Direct support for long pointers. This would require
C++ syntax extensions to support long versus short
pointers (similar to C/iX using ^ and *). Note long
pointer support could also be handled through
canned class libraries or external 'C' routines:
NS DC US NE PIP
(e) Support for standard MPE-escaped namespace
semantics (live with POSIX namespace semantics
if do not get this):
NS DC US NE PIP
(f) Other needed MPE-unique features: __________
_______________________________________
6) If a C++ compiler were ported to the 3000 using the
POSIX features of MPE/iX, and did not directly support
MPE-unique features, would you use it?
YES NO MAYBE
7) Would you trust a third-party compiler that was
supported by HP ?
8) What would you be willing to pay HP on a long-
term basis, for annual support of C++ on MPE/iX ?
9) If the cost of acquiring C++ on the 3000 was low
(a few hundred dollars to cover cost of distribution)
or non-existent, would you be willing to pay higher
support fees over a guaranteed term (say, two years),
where the fees over that period would be similar to the
total cost of buying and maintaining a comparable HP
compiler (Pascal/iX or C/iX) over the same term ?
YES NO MAYBE
10) Additional comments/concerns on this subject:
__________________________________________
__________________________________________
11) To give us the opportunity to contact you, please
enter:
name:_________________________________
phone:_________________________________
email:__________________________________
address:________________________________
city:____________________________________
state/prov-zip:____________________________
---------------------------------------- END of Survey.
|
