home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.object:3368 comp.lang.smalltalk:1750 comp.lang.eiffel:1102
- Newsgroups: comp.object,comp.lang.smalltalk,comp.lang.eiffel
- Path: sparky!uunet!stanford.edu!CSD-NewsHost.Stanford.EDU!Xenon.Stanford.EDU!hoelzle
- From: hoelzle@Xenon.Stanford.EDU (Urs Hoelzle)
- Subject: Re: Future Issues of Object Orientation
- Message-ID: <hoelzle.715392239@Xenon.Stanford.EDU>
- Sender: news@CSD-NewsHost.Stanford.EDU
- Organization: CS Department, Stanford University, California, USA
- References: <923@ast.dsd.northrop.com> <17tiuqINNn7i@network.ucsd.edu> <8ecdopy00awJ4HOlo0@andrew.cmu.edu>
- Distribution: usa
- Date: 2 Sep 92 00:03:59 GMT
- Lines: 32
-
- sm86+@andrew.cmu.edu (Stefan Monnier) writes:
-
- >Self (which is since recently available at self.stanford.edu in version 2.0)
- >claims to approach the C speed although it is FULLY dynamic
- >(even dynamic inheritance) by special 'run-time' compilation (which
- >writes different versions for special cases) !
-
- This is a little bit of an overstatement: not all Self programs
- approach C speed (yet :-). However, on the (C-like) Stanford integer
- benchmarks Self does very well.
-
- Speed discussions somewhat remind me of the 70s "C (or Pascal or..)
- versus Assembler". At that time, people had to decide between higher
- productivity (i.e. a high-level language) or a factor of around 2-3
- performance gain (assembly). Most people chose the "slow" high-level
- languages for normal applications, and I don't think anyone regrets
- it. Furthermore, as compiler technology advanced, the performance gap
- narrowed considerably.
-
- To me, today's "should I use {Eiffel, Smalltalk, Self} or C" sounds
- like a deja vu. If OO really helps productivity, it's worth a certain
- performance hit (the factor depends on the application, of course).
- If you can design and implement your new fancy application in half the
- time, you're better off than an application that is twice as fast but
- comes to market much later.
-
- -Urs
-
- ----------------------------------------------------------------------------
- Urs Hoelzle urs@cs.stanford.EDU
- Computer Systems Laboratory, CIS 57, Stanford University, Stanford, CA 94305
-
-