Good books in computer science?

Phil Runciman philr at aspexconsulting.co.nz
Sun Jun 21 22:40:08 CEST 2009


->From: Bob Martin [mailto:bob.martin at excite.com] 
-.Sent: Thursday, 18 June 2009 6:07 p.m.
-Subject: Re: RE: Good books in computer science?

-in 117815 20090617 221804 Phil Runciman <philr at aspexconsulting.co.nz> wrote:
->Because it reminds me of when things went badly wrong. IBM360, Von Neumann =
->architecture, no hardware stacks ...
->
->IMHO Burroughs and ICL had better approaches to OS design back then but had=
->less resources to develop their ideas.=20
->
->However, mainly this period marked a transition from the excitement and dis=
->covery phase of computing to commercial power plays and take-overs. The bes=
->t ideas in a field tend to get lost in the melee of competition. Early comp=
->uters were rooted in academia and there was a lot of cross fertilisation of=
->ideas and approaches. IMHO commerce affected layers of the stack where it =
->had no useful contribution to make. Vertical integration warred against sou=
->nd architecture.
->
->The book has an important message and I recommend that people read it. The =
->book is to me, and possibly only me, an icon representing when things went =
->wrong.


-Well, it's an opinion, but certainly not one I would agree with!
-AFAIAC the IBM 360 got everything right, which is why the instruction set is still
-going strong 45 years later (I've used it for every one of those 45 years). 

Yes, I was afraid someone would use that sort of argument. Sadly, having the best
instruction set does not lead to commercial success. If it did then Interdata would
still be with us. They used IBM360 instructions.

How many instruction sets have you used? I have used at least 9.(I nearly missed
the DG Nova). KDF9 had the best set for general computing that I had the privilege
of using but that is not to say it was the best. The Burroughs B series or PDP11 may
have been better... and doubtless there are many more candidates.

What I can say is that for scientific/engineering calculations the RPN of KDF9 was
Great because assembler was no harder than using algol60 for the calculations part of
the problems I worked on.

Oh yes, I even used assembler on the IBM360 series. It was a 360/50. The experience
Did impact on the force of my observations! FWIW I learned it using the training material
For the ICL System 4 which was superior to IBM's. The ICL System 4 was not a success...
despite its instruction set. ;-)

-AFAIAC the IBM 360 got everything right
How many known bugs did the OS end up with? I know it hit 50,000+ and counting. LOL

Suffice to say we are on a journey and Python is part of the scenery.


Phil



More information about the Python-list mailing list