home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.misc
- Path: sparky!uunet!think.com!ames!nsisrv!author.gsfc.nasa.gov!rkoehler
- From: rkoehler@author.gsfc.nasa.gov (Bob Koehler)
- Subject: Re: What makes Unix Special?
- Message-ID: <4JAN199309464458@author.gsfc.nasa.gov>
- News-Software: VAX/VMS VNEWS 1.4-b1
- Sender: usenet@nsisrv.gsfc.nasa.gov (Usenet)
- Nntp-Posting-Host: author.gsfc.nasa.gov
- Organization: CSC System Sciences Division
- References: <1992Dec31.062544.5838@news.columbia.edu> <31DEC199210495450@author.gsfc.nasa.gov> <C04zI3.EuK@mtholyoke.edu>
- Date: Mon, 4 Jan 1993 14:46:00 GMT
- Lines: 83
-
- In article <C04zI3.EuK@mtholyoke.edu>, jbotz@mtholyoke.edu (Jurgen Botz) writes...
- >Unix wasn't really "written in 1969". It was begun in 1969. There's a big
- >difference... the common Unix variants (SV, BSD4) are products of the late
- >70s.
- >
-
- Of course, is fairly obvious from the advances of the last 5 years that the
- writing of UNIX isn't done. My claim, however, is that the UNIX kernel shows
- the result of decisions made early on, back in the days when we rebooted our
- PDPs fairly often, and 1 MIPS was not available on the typical budget. I will
- site only one example. The used and free space in the file system buffer in a
- typical UNIX is tracked through a series of fixed size arrays. When the buffer
- becomes more fragmented than can be tracked due to one of the arrays being
- full, the next to last entry is dropped (the last entry is likely to point to a
- large space). Several resources within the kernel are tracked in this manner,
- thus leading to the gradual loss of such resources until the system is
- rebooted. This algorythm is fast, simple, and require little overhead in
- memory, but the extra instructions needed to use something a little more
- complex, like a doubly-linked list, are insignificant for todays CPU speeds and
- memory sizes.
-
- People with access to sources have told me that BSD actually contains improved
- algorythms in several (not all) areas simmilar to this in the kernel, but
- System V does not contain all of even these improvements.
-
- >
- >Arguments can be made for anything. Sound ones... well, make some! ;-)
- >
-
- On comp.unix.misc? You must think me suicidal! :-)
-
- Yep, I could argue the point, and I suppose you could argue the counter-point,
- or in fact I could argue the counter point, or that DOS is better than MVS, or
- (fill in you favorite 2 OSs)...
-
- Most of this would boil down to anecdotal evidence, a large body of which
- surely exists for almost any point of veiw. Beyond the example sited above
- I will not fill bandwidth with what is not likely to be a full, satisfying
- answer.
-
- >>One of my pet peeves in UNIX land is its closed nature. To me, open means you
- >>can look inside. I cannot afford the current UNIX source license prices, but I
- >>have found source licensing arangements for proprietary systems often an order
- >>of magnitude lower.
- >
- >Really? I don't know otherwise, but I find this /very/ surprising. How
- >much for a VMS source license? Can anyone even get it?
- >
-
- I have had the VMS source-listing microfiche, which used to ship with VMS ($0,
- suprise, ah for the good old days), and I last bought for around $1500 (its
- probably a bit higher now). The sources have always been available to me
- whenever I needed them. A machine readable copy would be much more expensive,
- and it seems right now that the only option available from my UNIX vendors.
-
- >Besides, a BSD-NET2
- >"source license" is $0. Ditto for Linux. Ditto for the Mach microkernel (not
- >Unix, but a good foundation for a Unix-workalike). Ditto for GNU.
-
- I'm not sure this does me any good when I'm tracking down a problem within an
- AIX, HP-UX, or ULTRIX based application. The last nice round figure I got
- was "oh, about $50000" (ouch).
-
- >
- >>Personally I hope POSIX will catch on [...]
- >
- >I agree.
-
- Now I new there was something we could agree on, along with UNIX being better
- than (at least) most.
-
- >*grin*... I don't consider this response a flame, and I'm also not writing
- >it to goad a response out of you, so go ahead and ignore it.
-
- Well if I had thought it was a flame, you probably wouldn't have seen this.
-
- ------------------------------------------------------------------------------
- Bob Koehler | Any illusion to these opinions being other
- rkoehler@author.gsfc.nasa.gov | than just mine alone is just that.
-
- " Life is life, and fun is fun, but it's all so quiet when the goldfish die. "
- - Blixie
-
-