COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	histor[e84,jmc]		Opinions on the history of computing
C00006 ENDMK
C⊗;
histor[e84,jmc]		Opinions on the history of computing

hacking:	The incremental development of a computer program or other
		object.  One starts by building the part one understands
		or immediately wants, and builds more with this as a base.
It has its limitations compared with a top down design starting with a
theoretical understanding of the function of the object and its possibilities.
It works better with computer programs than with anything else.  Building
an airplane this way wouldn't work.  One story buildings can be hacked, i.e.
one begins with some rooms and adds on, but multi-story buildings cannot.
Complexes of buildings, e.g. Stanford University, start with a planned
structure and then are hacked.

	What von Neumann thought he had to do to justify a computer.
	John von Neumann wrote a report in 1946 proposing to build a
stored program computer.  The report proposes specific problems (partial
differential equations) to be solved, argues that there will be enough
work to keep a computer busy, and gives fragments of program.

	IBM 701, 704, 709, 7090, Stretch, 702, 705, 650, 610, 795, 7100,
360, 370, pc.  It is especially interesting to consider the computers
that not even IBM could sell - to account for the errors.

pdp1,pdp3,pdp4,6,10,2060,7,9,15,16,vax.

besm and other Soviet computers.

360 good things, combine commercial and scientific on binary
mistakes, half-duplex, no relocate, 24bit, displacement

generic mistakes, short address,

bad ideas, directly executable higher order language, decimal, firmware
paging and segmentation,

nonsense about nth generation

Well, if all this is true, what current ventures are likely to succeed
and fail.

Good and bad ideas in programming languages and programming generally.

The importance of hand holding.

Putting the thesis in hardware.

Cray's successs.

The vices of engineers, the vices of business men, the vices of computer
scientists.

Academic log rolling.

The vices of engineers:

1. Like any other specialists, engineers often over-magnify their
part of an enterprise.  In computing this has taken several forms.
First the set of operations of a computer tends to be inflated.
Historically this has been at the cost of address space.  Lack of
address space has been the biggest historic reason for the obsolescence
of computer architectures.  Examples include the IBM 704-709-7090-7094
sequence and the D.E.C. PDP-11 sequence and more recently the
DECSystem-20.