7

The famous The UNIX-HATERS Handbook claims this mailing list had been inspired by TWENEX-HATERS(1) and other *-LOVERS mailing lists, a long tradition of MIT.

Moreover, the quote below implies there were long-time complaints against mainframe platforms by hacker-hated IBM and wry-regarded DEC.

Users said that they wanted Unix because it was better than the "stone knives and bear skins" FORTRAN and COBOL development environments that they had been using for three decades.

But, in choosing Unix, they unknowingly ignored years of research on operating systems that would have done a far better job of solving their problems. It didn’t really matter, they thought: Unix was better than what they had.

By 1984, according to DEC’s own figures, one quarter of the VAX installations in the United States were running Unix, even though DEC wouldn’t support it.

There also seems to have been disputes between the mainframe community and minicomputer community:

So, were there any equivalents of The UNIX-HATERS Handbook for mainframes like the IBM S/370, or minis like the DEC PDP-10 or DEC VAX, covering aspects including, but not limited to, hardware architecture, programming difficulties, user experience, and so on?

Other links:

(1) TWENEX is the nick name of TOPS-20, a mainframe operating system by DEC.

(2) A quote from ctrl​‑alt​‑delor, in the chat room for that question, states:

Unix is an old system, that should have been replaced years ago. Unfortunately it is the most modern system we have, and still years ahead of MS-Windows.It seems to have been at least 10 years ahead of MS for at least the last 30 years (probably longer).

paxdiablo
  • 4,742
  • 20
  • 47
Schezuk
  • 3,752
  • 1
  • 17
  • 40
  • 1
    Hmm, interesting, but why should someone write such a 'book' when mainframes were already years ahead, as the quote supports, carrying way less deficit to complain about? (All beside the fact that I wouldn't call any DEC system a mainframe. They are minis). – Raffzahn Jan 28 '23 at 13:49
  • 2
    VAX wasn't a mainframe by any standards (though the ill-fated VAX 9000 aspired to that label). – dave Jan 28 '23 at 14:25
  • 3
    Your question sees oddly unbalanced - Unix Haters is about software (which runs on several hardware bases) and the question is asking about hating hardware (each instance of which can run several different operating systems). – dave Jan 28 '23 at 14:36
  • 4
    Sign me up for getting a copy of the x86-Haters Handbook though. – dave Jan 28 '23 at 14:47
  • @another-dave Oh, that'll be interesting, as I consider myself an 8086 lover :)) – Raffzahn Jan 28 '23 at 15:09
  • @Raffzahn Do we include the Xeon Phi in the family of 8086 products that we love? :) – doneal24 Jan 28 '23 at 21:39
  • The ** footnote appears to have no references to it, making its intent unclear. – paxdiablo Jan 28 '23 at 21:47
  • @doneal24 Let me give you an analogy: I love the /360. The /370 did renew that love. Sure, here were some oddities but it was easy to ignore them. Then came /390. What a great offering 31bit was - but it burred all beauty under inches of make up and odd applications. Everything after that stays better behind behind a curtain. Better to look at old paintings what e Beauty it once was --- or in other words, there is a reason I worte 8086 in a reply mentioning x86 :)) – Raffzahn Jan 28 '23 at 23:27
  • 1
    I feel similarly about extended (multiple section) addressing on the PDP-10. It's a wart on an otherwise-handsome face. – dave Jan 28 '23 at 23:47
  • @paxdiablo Source. I should have made it clearer but I thought it belongs to How should we interpret Dave Cutler's criticism of Unix? – Schezuk Jan 29 '23 at 00:58
  • @another-dave Should this question be rephrased as programming and administrating on OSes of something-larger-than-average-minis or non-Unix-minis? – Schezuk Jan 29 '23 at 01:13
  • 1
    Schezuk, thatnks for clearing up the footnote issue. I have edited the question to hopefully make that clearer, incorporated some comments (mostly to do with referring to PDP/VAX as a mainframe), and fixed up some minor grammatical bits that my CDO railed at :-) Please check to make sure I haven't changed the essence of the question. – paxdiablo Jan 29 '23 at 02:47
  • @another-dave: if you haven't yet read Tracy Kidder's "Soul of a new machine", you may find it interesting. In there, Ed De Castro (I think) from DG specifically called a mode bit (selecting 16/32-bit mode) a "wart" and wasn't that impressed with the way VAX had been done :-) – paxdiablo Jan 29 '23 at 02:50
  • 1
    Yes, I read Kidder, and at the time thought it a fine depiction of life in programming (I was at DEC). I sent a copy to my Dad, hoping he might understand sort-of what my job was about. But as for mode bits, I think the VAX approach was better - it allowed the new ISA to be clean, and eventually the old ISA was dropped completely. That is, PDP-11 mode was seen as a transition aid. – dave Jan 29 '23 at 03:20
  • 1
    @paxdiablo Please accept my thanks for your kindness. – Schezuk Jan 29 '23 at 03:42
  • 1
    THe DEC-20 was a timesharing machine. Timesharing machines have tended to be lumped in with either mainframes or minicomputers, but they really should have had a category all their own. – Walter Mitty Jan 29 '23 at 12:35
  • I do see that some mainframes were only used for batch processing and some only for time-sharing but how to classify ones that were used for both? The IBM 709 through 7094 is used for FMS batch and for CTSS time-sharing. – Knickers Brown Feb 01 '23 at 19:37
  • The quote from the sourced materials "By 1984, according to DEC’s own figures, one quarter of the VAX installations in the United States were running Unix, even though DEC wouldn’t support it." is incorrect in that DEC had a supported UNIX, v7m (sometimes v7m-11) which was at version 2.1 in September 1981, A PDF file of the documentation is in this directory: http://bitsavers.org/pdf/dec/pdp11/ultrix-11/ – Knickers Brown Feb 01 '23 at 20:19

1 Answers1

9

"The Unix-Haters Handbook" is a product of a particular time and place that made an entire book possible, the rise of RISC workstations and the Internet.

People had been grumbling about mainframes and then minicomputers before then. For an example that is very much in the spirit of the Unix-Haters Handbook, search for "Darryl Rubin A Problem in the Making". This is a very funny comparison of HAL 9000 and IBM mainframes that was published in a major professional IT magazine. It's so good that the most common version now circulating is a copy with IBM mainframes replaced by Microsoft.

However there were a lot less mainframes and minicomputers around, so a lot less people who read this kind of writing. By the 1970s there weren't any real prospects for introducing a new kind of mainframe because it would have cost too much, and by the 1980s it was getting almost as expensive to introduce a new kind of minicomputer. So complaints were mostly aimed at improving existing products, not replacing them with new ones. A "Mainframe Haters Handbook" would have had a very small potential audience, and no Internet to spread by recommendation rather than an expensive advertising campaign.

The Unix-Haters mailing list starts in the late 1980s and the book is published in 1994. This is the time when RISC is the hot new thing, a revolutionary new CPU architecture making high performance computing much more available. Personal computers existed, but at that time they were very much considered toys, not powerful enough for real work. Personal computers do mean that there is a much larger audience reading about computer technology: magazines such as Byte would cover advanced research as something expected to arrive on PCs in the future.

RISC builds on the latest decade or so of research in computer architecture. There's also been a decade or more of research into operating systems and user environments going on, so the OS on your brand new RISC box will be equally advanced, right? The sort of advanced environment from Xerox PARC or MIT will now be generally available?

Nope. These super duper RISC workstations are all running some variant of Unix from the early 1970s. From the viewpoint of the people who worked in OS and related research, it's as if the Concorde or Boeing 777 were being built as biplanes.

So The Unix-Haters Handbook is not just a rant against Unix, but also pointing out all the better alternatives that were being ignored. The Internet meant that these people could easily find each other and generate informal publicity, personal computing had created a wider potential audience, RISC computers had introduced a time of transition with a lot of people unsure about what might come next and looking for information. Enough to persuade a publisher to make an actual physical book.

More modern versions of the theme expressed in The Unix-Haters Handbook are found on YouTube. For example, Rob Pike "Systems Software Research is Dead" and Timothy Roscoe "It's Time for Operating Systems to Rediscover Hardware".

Hugh Fisher
  • 311
  • 1
  • 4