Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!sdd.hp.com!uakari.primate.wisc.edu!dali.cs.montana.edu!milton!uw-beaver!zephyr.ens.tek.com!tektronix!percy!data!kend From: kend@data.UUCP (Ken Dickey) Newsgroups: comp.edu Subject: Re: sys & lang standardization? Message-ID: <429@data.UUCP> Date: 16 Nov 90 19:36:15 GMT References: Organization: Microcosm, Beaverton, OR Lines: 40 hp0p+@andrew.cmu.edu (Hokkun Pang) writes: >Will we ever get to the stage that operating system and programming language >are well stablelized(am I dreaming?). It seems to me that every month or so, >there's a new (or major upgrade to an existing) operating system or programming >language. So much energy are wasted (can I say this?) in learning new features >and rewiting old codes. I don't think that you will see standardization as long as hardware and software implementation technologies keep changing. The basic reason is that both the requirements and the cost equations change. If you look at personal computers, we have gone from 4-8K -> 64K -> 640K -> 6+MB and, by extrapolation, -> 640MB. During this time we have gone from ttys and "glass ttys" to windowed displays. Major changes have taken place to use the additional resource (ram, blitters, etc.) in terms of bitmaps, network buffers, etc. Soon the challanges will be interactive video and voice recognition, etc. What has happened is that new capabilities are taking up the added storage, communication, and processing power. This has meant the approach of computers used by mere mortals (as opposed to us CS nerds, who are merely mortal 8^) as user interfaces evolve. The rays of hope I see is that interoperability of computers over heterogenuous networks will bring some commonality. Also, it still seems that hardware technologies are changing much faster than software technologies. This means that to protect investment in developing software, software must be written in a way which is abstract enough to allow it to outlive many hardware technologies. In addition, it is adviseable to use software languages which have trivial syntax and formal foundations as we can then use source->source transformations to migrate working software to newer software technologies as our models of computation change and clarify. All of this is driven by basic economics. My suggestion is to watch sw implementation technologies that have the above properties, e.g. Scheme & ML, etc. -Ken Dickey kend@data.uucp [Yes there will always be MVS/COBOL/FORTRAN, and Unix/C, but do you really want to work in these environments? 1/2 8^]