srradia@watmath.UUCP (sanjay Radia) (03/15/85)
In article <171@dmsd.UUCP> bass@dmsd.UUCP (John Bass) writes: >From Mr Tevanian: > "Would you care to say *why* the file system is the only logical > place for global names to reside? .. In fact, THERE IS NO REASON > that the kernel need know what a file is! Of course, many of the > readers of this newsgroup (hopefully not all) are still living in > the primitive world of "the Unix Way." > >More Ivory Tower snobbery we can do with out ... every CSc department has >its pet students and operating system research programs to produce mindless >trash to be ignored by the rest of the sane world (how this for an attention >getter living in the primative world). OS research is the fabric of learning >what and what not do do in real life -- more the latter I feel. I will grant >him that an OS NEED NO KNOW about a file as a fundamental resource for some >applications -- on the other hand the file is the basis of most commercial >applications environments. > The kernel does not need not know about the file system; this can be provided outside using processes. Such file systems are very flexible and lend themselves very easily to being distributed over networks. Just look at the complications and mess that people have created in trying to get distributed file systems in UNIX. At waterloo the the Thoth project produced a full fledged operating system. Its current version, called Port, has a nice distributed file system which resides outside the kernel. Port is now being commercially marketed. It is much more suited to the network environment than Unix. The V-Kernel at Stanford is also based on Thoth. Operating systems research programs can produce worthwhile products. The reason why many such operating systems do not get commercial popularity is because it is very difficult to switch to new operating systems. You have to port a ton of languages and applications. All I am saying is that many superb OSs produced at universities and in the research labs of companies just never get out. I am sure that 15 years ago people would have reacted in the same way to putting the command processor (the shell) outside the kernel. Also, remember, Unix was the product of little pet project of KT & DR at bell labs and it sure produced a lot more than "mindless trash". -- sanjay UUCP: ...!{ utzoo,decvax,ihnp4,allegra}!watmath!srradia ARPA: srradia%watmath%waterloo.csnet@csnet-relay.arpa CSNET: srradia%watmath@waterloo.CSNET