NICE-ESG-Libs Digest Fri, 12 May 95 Volume 1 : Issue 218
Today's Topics:
Steve Fisher
NICE Eiffel Standards Group -- Library Committee Mailing List
To post to list:
NICE-ESG-Libs@atlanta.twr.com
To send mail to the Chairman of the committee:
NICE-ESG-Libs-chair@atlanta.twr.com
Administrative matters (sign up, unsubscribe, mail problems, etc):
NICE-ESG-Libs-request@atlanta.twr.com
Date: Fri, 12 May 1995 09:46:40 +0100 (BST)
From: S M Fisher
Subject: Steve Fisher
To: NICE Library Committee
Christine has suggested that new members might introduce themselves, so...
I am working with a group from CERN (European Laboratory for Particle
Physics) in Geneva looking at all aspects of OO technologies. URL
http://www.cern.ch/OORD/Home_oord.html starts by saying:
"Moose is a Cern Research and Development project (RD41)
to study the viability of the Object Oriented approach for software
development for High Energy Physics reconstruction and analysis code at
the LHC.
By working on some well defined prototypes experience will be gained in
Object Oriented analysis, design and implementation.
Moreover, requirements specification, risk analysis, project management
and process control will be applied to the development of
the prototypes.
The MOOSE collaboration mainly consists of members of the Atlas and CMS
collaborations from Europe, Japan and the USA.
I am part of the atlas collaboration. Please see
http://atlasinfo.cern.ch/Atlas/GROUPS/ATLASOO/Home_atlasOO.html for
information on what we are doing.
One of the things we have to do is to come up with a set of criteria for
the choice of language(s). So far the atlas group have been using Eiffel
and the CMS group C++. I fear that if Eiffel does not become more
significant to industry (supported by CASE tools, interface builders and
the OMG) then our criteria are likely to exclude Eiffel.
We are very concerned to have multi-vendor support to avoid mistakes we
have made before. I want to see a number of independent vendors with their
own share of the market - and most importantly a PD/shareware Compiler and
Libraries. I do not like to mention C++ again, but the gnu compiler has
had a big effect. For many people this compiler defines the C++ standard.
It might be in the interests of the vendors to promote Eiffel by making
some of the PELKS compliant libraries public domain.
I have 4 Eiffel compilers at the moment (SiG, ISE, Tower and Eon) and see
the problems involved in shifting code between systems. With practice I
have learned to write FORTRAN and C which is highly portable. I want to
see the same portability of Eiffel systems.
This includes:
1) LACE
2) Calling in and out from other languages
3) The language
4) The libraries
I do not expect to change a single line of my code when I switch from one
vendor to another. The code we are developing has to run on about 80
different sites around the world. One lab may prefer to pay more for a
fancy development environment while another may prefer to get a lot of
compilers as cheaply as possible. Yet another lab may want the compiler
which produces the most efficient code because they have a lot of number
crunching to do.
I have been programming in Eiffel for a few years and try to take a fairly
pure approach to OO. You should expect my comments in the future to be
guided by my belief that if A inherits from B, then A is a B. This means
that I do not like classes such as MATH. I also like to keep classes to a
manageable size for the purposes of understanding them. However if a class
needs a feature then it should be there so I am perhaps somewhere between
Paul Johnson (the minimalist) and Bertrand.
I have been watching the progress of this committee for some time and I
am not too happy with the way it conducts its business. In a real meeting
many people are silent but you can gauge from their expressions and grunts
whether there is any level of support for a proposal. With e-mail silence
is hard to interpret. I think it is better to say something which is
wrong and to be corrected than to remain silent until the time of the
vote. If people felt more free to make suggestions and not just very
carefully phrased proposals I think that the process could go quite a bit
faster, consensus would be normal and the vote just a formality.
Regards,
Steve Fisher
|
|
