[FoRK] Niklaus Wirth 80th Birthday Symposium
sdw at lig.net
Thu Feb 20 16:50:13 PST 2014
On 2/20/14, 11:11 AM, Dave Long wrote:
> Congratulations to the ETHZ for putting together a great program to
> celebrate Niklaus Wirth (call by name? or call by value?)'s birthday.
But did he apologize for Pascal yet? In particular: lack of
shortcutting of expressions, magic IO (a la Basic) rather than proper
library interface mechanisms like C, and no varargs.
The main good thing I can say about Pascal: It was better than Basic.
Otherwise, countless hours were wasted that could have better been spent
on C (and later C++).
I started my long career of winning technical arguments (sooner or
later) in my first real development job, doing research machine control
development at GE Lighting in 1984. My EE lead simply loved Pascal MT+,
which while it had C-like additions for low-level access, somewhat in
contrast to Turbo Pascal perhaps, was also very buggy. I found at least
7 compiler bugs on the first project. I was arguing for C the whole
time, periodically cursing Wirth. In one of my only 3 undergraduate
classes I ever took, the TA was learning Pascal as he was teaching us.
Kind of underwhelming but at least I narrowly escaped punched card
Fortran which was still a mainstay at universities still but completely
absent at GE Lighting Labs at NELA Park.
I developed for my project on MP/M 8/16 while using a DRDOS PC and
playing with C, sh, etc. on a Perkin Elmer larger mini / mainframe and
another mini, after not much fun on our Vaxes. Emacs, too. It was so
obvious where things would end up even then.
> The day started with memories of the Algol-68 process, followed by
> Vint Cerf musing about bitrot, Carroll Morgan about the promise of
> (in)formal methods (will have to try Dafny sometime; Morgan claims
> that if your implementation fails to meet a spec, current model
> checkers are capable of automatically generating counterexamples and
> are much lighter weight than full on theorem provers), Hans Eberle on
> pushing bits in the simplest way possible (to the point where the
> chip-to-chip interconnect becomes simply capacitance and proximity!),
> Clemens Szyperski on QBE in Excel (for those who still chase the grail
> of end-user programming), Michael Franz (for our UCI contingent!) with
> whom the times have caught up: the old "thin binaries" have turned
> into customized/randomized per-end-user executables (in which Three
> Letter Agencies apparently have great interest), Martin Odersky on the
> future of types in Scala, Bertrand Meyer in his kinder, gentler, (more
> Swiss?) incarnation almost saying nice things about C++ while
> demonstrating how Eiffel has managed to evolve without losing its
> character, Kathleen Jensen on the tribulations, opportunities, and
> above all serendipity of a SCruzian hippie chick coming to 1970's
> Zürich to do computers, and the pièce de résistance, Niklaus Wirth
> showing off his latest hack: a port of his old Oberon system to a
> microprocessor of his own design, instantiated on a cheap FPGA sample
> board (to which he added a CF disk, but didn't see any need of
> expanding the stock 1M onboard memory. Speaking of hippies, on a
> Brandian Computer Bum note, his 1970's hemispherical mouse has
> remained to provide a constant reference size in the pictures as the
> size of his computing systems have shrunk from mini-under-the-desktop
> to behind-the-monitor Raspi'ish format).
> I asked Wirth if he was able to provide library source in Oberon
> instead of assembly in the recent version of the book because the
> hardware had gotten faster faster than he could make it slower with
> software, and he replied, no, mostly he understood the problems and
> could express himself much better now than he could the first time
> around. Truly inspirational.
More information about the FoRK