[Oberon] Some ideas on ETH-oberon & Linux/C relationship

easlab at absamail.co.za easlab at absamail.co.za
Sat Jul 8 13:01:08 CEST 2006

I've  read the criticism/defence by the C originator[s] re. 'criticism 
from Pascal'.  An important aspect was about the 'unsafe arrays'.

I could believe that 'wiki has been there all the time' except that
the articles are dated [latest edit] and I'm astounded at what's
available via http://en.wikipedia.org/wiki/  these days.

Wikipedia has exeeded the expectation of the idea which I had been 
promoting: contributions/effort which instead of aging and breaking,
gets refined & better by usage.

Recently I've been led to a 'tree' of wikipedia articles which deal with 
the various types of problems [virus, worms etc.] which are possible 
because of the unsafe arrays of C.   It's strange to me that no one 
raises the question of 'languages that avoid the unsafe arrays
problem'. OTHOH wikipedia will tend to be a [well formatted &
easilyn accessible] repository of main-stream knowledge.
Ie. NOT original/contraversial ideas.   It could even tend towards 
more dumbing-down towards the herd-opinion ?

OTOH I think that ETH-oberon should continue to track linux, to
tap into and benefit from the big contributor knowledge base.

I remember that Turbo Pascal had a substantial library of procs/funcs.

Would  Linux/C's  library be partly a good guide to a suitable
similar Oberon library ?  Perhaps OOC has already solved this ?

I think linux/C could be used somewhat as one has references to
ancient Greek & Roman. Ie. to avoid repeating the same research
and mistakes. 

A tool which assisted to eg. transform a linux/C device driver to 
ETH-oberon syntax shouldn't be very difficult to build, and would 
be of great value.  It need not be automagically-deterministic.
Even if it's partly like Bablefish.   What happened to Bablefish ?

Again a tool that can be used soon, but can continue to evolve
with increased usage  -- probably 'data driven programming' --
is the way to go ?

Feedback on ETH-oberon via linux:
Using the linux ppp, sure solved the N-O ppp problem.
I've still  found no good reason why ISPs don't feed back an engaged
signal, to avoid the client the cost of a call which can't get/give data 
because of system overload. 

Amazingly ETH-oberon fetches http also when connecting via linux/ppp.
And well, for simple/reasonable web-pages.
IMO most web-pages which use elaborate facilities do so as a 
substitute for valuable content.     Gmail eg. is a disaster !

Apparently ETH-oberon communicates with the linux ppp socket ?
It would be nice if email & nntp sockets of linux could also 
communicate with ETH-oberon.   Is that difficult to do ?


== Chris Glur.

More information about the Oberon mailing list