D-day

[Tyson Richard DOWD]

Zoltan Somogyi <zs at cs.mu.OZ.AU> writes:


> > 	It'd be very interesting if you could give something on Mecury
> > that you have been heavily involved for the D-day.

>We have not done so in previous years, because we couldn't come up with
>anything that the average person in the street would relate to. There are
>at least two major difficulties to be overcome: that nothing about Mercury
>is visual, and that Mercury is a tool for programmers, not users.

Well, we're not _entirely_ trying to appeal to the "person in the
street", there will likely be some inexperienced programmers (sorry,
coder doodz ;-), and potential computer-science students. Also, you'll be
likely to have some IT professionals (and their children) looking at the
course.

>I suppose we could put up a poster listing Mercury's benefits and possibly
>some sample Mercury code alongside equivalent C code. We might also mention
>that several students working on the project had published papers while
>still undergraduates. However, I suspect you were asking for something
>snazzier than that.

>Anyone got any bright ideas?

Well, if you assume there will be some people who have some C or
assembler knowledge:

There's *some* potential for simplistic posters showing 
	- differences between declarative languages and imperative langs.
	- how complilation to C works (roughly)
	- features and benefits of Mercury (for large-scale applications
	  in particular).
	- examples of program transformations 
	- code generation techniques (register allocation)
	- code optimization techniques (before->after for simple
	  things, eg peephole)

However, the question is, if we spend time on these things, are people
even going to look at or read them. And what are they going to see as
relevant to the undergraduate course, since Mercury isn't offer as
anything other than a segment of a second year subject.

I think if we are careful to keep it accessible to people with
rudimentary knowledge about programming, we could do something
along these lines.

The other opportunity is to show how Mercury can be *used* - for example
Tom's 3-d space thingie, Tom's observatory program, Tom's maze program
(don't suppose you have anything else Tom?) - Fergus's bridge player? 
Development time and code length might be what we want to emphasize as
well.  Doing these demos depend upon what machine resources we can use.

One more possibility is to write a _very_ good checkers player, and hook
it up to the vision checkers player (not saying they don't have a good
one already). I can just see the cheesy "Beat the logic program and win a
prize" sign now.... ;-)