home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.c++,comp.lang.eiffel,comp.lang.c,comp.object,comp.software-eng
- Path: assip.csasyd!news
- From: donh@syd.csa.com.au (Don Harrison)
- Subject: Re: Portability of code & skills (Beware of "C" Hackers etc)
- Message-ID: <Dot383.45s@assip.csasyd.oz>
- Sender: news@assip.csasyd.oz
- Reply-To: donh@syd.csa.com.au
- Organization: CSC Australia
- References: <31517E6F.5930@dmu.ac.uk>
- Date: Mon, 25 Mar 1996 04:31:14 GMT
-
- Graham Perkins wrote:
-
- :Don Harrison wrote:
- :> :adoption of C and Unix.
- :>
- :> Aside from the historical argument, could another reason be that people
- :> love power and permissive languages such as C give such power?
- :
- :Permissive, yes. Power? I don't think so. Perhaps too many people
- :confuse the two. For example, local functions give a certain power,
- :expecially if you can pass them as parameter as in Pascal. This power
- :is unavailable in C. Simple manipulation of structured objects is
- :quite awkward, and has been severely limited and error prone for most
- :of the lifetime of the C language. Cobol's pattern matched input and
- :pattern editing output is extremely powerful, but C does not even check
- :you pass the same number of print items as you specify, never mind give
- :you the power to control currency sign, leading zero suppression, decimal
- :format, credit/debit formats... And what about that file handling?
- :Read an arbitrary number of bytes into an un-typed buffer .. that's power?
- :
- :No curried functions, function constructors, lazy evaluation, or generic
- :data and function types. No assertions, exceptions, concurrency, or
- :parallelism. No built in string, vector, matrix, or complex operations.
- :No garbage collection. No embedded SQL. No opaque types or language
- :construct for naming a group of related operations. Think of any powerful
- :construct you like from any language and you will find it NOT in C (or
- :at least not during the period C was catching on).
-
- I was thinking of 'power' in terms of having the freedom to do what you want.
- I should have said 'freedom' or 'permissiveness'. If you define 'power' as 'the
- ability to say little and acheive much', C's nature, as you indicated, is the
- antithesis of it.
-
- :> :..universal assembler ..does not seem to have happened, and I don't understand
- :> :why.
- :>
- :> Someone else already pointed out that Eiffel uses it as such.
- :
- :Well it hasn't really happened in the way I meant. Those who demand their
- :systems in C are resistant to Eiffel because they perceive it as a language
- :that is not C, in spite of the C back-end.
-
- If they perceive it that way, then they are right because Eiffel is very much
- a language in it's own right. The fact that is compiles into C is incidental.
- It could (and, IMO, *should*) compile into a better 'universal assembly language'
- than C. What would the characteristics of such a language be?
-
- - maps as directly as possible onto the hardware (for efficiency, portabilty)
- - inheritance/polymorphism (can a new generation of OO hardware remove/reduce
- the performance penalty of dispatching?)
- - ??? (fill in the gaps).
-
- [...]
-
- :It seems to be a matter of flavour. If the Eiffel browsers worked almost
- :entirely visually (eg, class tool contains rename table, redefinitions
- :colour coded, export control by dropping classes into or pulling out of
- :a viewpoint icon) then much of the syntax could disappear or be hidden.
- :Then people would be more inclined to view it as an OO development system
- :for C rather than as a different language.
-
- These sound like good ideas. However, IMO, Eiffel is not a tool that would be
- readily marketed to the C community because their focus is different. Eiffel has
- too high an emphasis on quality and reliability to appeal to most C users ;-).
- It would be more attractive to those who actually practice real software
- engineering (rather than paying lip service to it).
-
- :--
- :person: Graham Perkins paper: School of Computing
- :voice: +44 (0)1908 834936 De Montfort University
- :dots: +44 (0)1908 834948 Milton Keynes MK7 6HP
- :bits: grp@dmu.ac.uk United Kingdom
-
- Don.
-
-
-
-
-
-