[TYPES] Grand Challenge Problems?

[KS Sreeram]

Here is a grand challenge:

Make a clean, expressive, high-level programming language which
has performance that _equals_ C++.

Many programming languages have features that make it safer/cleaner/more
expressive than C++. But every one of those languages sacrifice performance
in the process of achieving those features.

The following line of research doesn't seem to have been explored much:

How far can we go in terms of safety and expressiveness without sacrificing
even an iota of performance?

---
KS Sreeram
Tachyon Technologies


On Sun, Jan 17, 2010 at 8:44 AM, Purush Iyer <purush.iyer at gmail.com> wrote:

> [ The Types Forum, http://lists.seas.upenn.edu/mailman/listinfo/types-list]
>
> Derek
>
>  Thanks for the pointer.  I do remember looking at these discussions after
> 09 POPL.  But they did not, unfortunately, solve my quandry
>
>  By "scientific" I am referring to problems that need either new
> mathematics
> to be invented or perhaps a new method.  It is hard to define what is
> scientific, though I am sure all of us would be able to say when presented
> with an example whether it is a scientific problem or not.
>
>  I consider problems posed by Alan Mycroft's 1980 paper (on call-by-need
> and
> value) and his 1984 paper (on Recursive type schemes) as posing scientific
> questions.   The former led to ten years of papers in POPL and other places
> (with nifty ideas by Hudak, Mishra, Jensen and others).  The latter led to
> the notion of semi-unification (by Henglein, Kfoury and others).   At the
> risk of sounding  vain, I consider my 1997 paper in TAPSOFT on quantitive
> reasoning in Probablistic Lossy Channel Systems as raising a nice
> scientific
> question in Formal Methods; it took the discovery of notion of attractors
> in
> infinite markov chains (by Alex Rabinovitch, Parosh Abdulla and others) to
> solve the problem effectively.
>
>  My question is simple:  Show me problems that are scientific, that need
> new
> mathematics or blindingly new insights to solve.   I am stumped while
> explaining what the deep problems in our field are to Physicists and
> Control
> Theoreticians.   In all of this I realize that work in PLS is synthetic in
> nature -- part science, part engineering -- and that, may be, we typically
> make up a problem and solve it right away.   Perhaps there are no
> scientific
> grand challenges; but, I dared to ask anyway.  Thanks for hearing me out.
>
> sincerely
> Purush Iyer
>
>
>
>
>
>
> On Sat, Jan 16, 2010 at 5:58 PM, Derek Dreyer <dreyer at mpi-sws.org> wrote:
>
> > [ The Types Forum,
> http://lists.seas.upenn.edu/mailman/listinfo/types-list]
> >
> > A note from the moderator:
> >
> > I agree with Matthias that Purush's question is a vague and extremely
> > broad one, and it's worth asking if there's a more "scientific"
> > definition of "grand challenge".
> >
> > However, given that at last year's POPL there was a panel session
> > devoted precisely to the topic of "PL Grand Challenges", I believe the
> > question (even in its vague form) is not out of bounds for a
> > scientific forum like the Types-list.
> >
> > FWIW, to address Purush's question, some notes from that POPL'09 panel
> > session on PL Grand Challenges are available here:
> >
> > http://plgrand.blogspot.com/
> >
> > Thanks,
> > Derek
> >
> > On Sat, Jan 16, 2010 at 11:25 PM, Matthias Felleisen
> > <matthias at ccs.neu.edu> wrote:
> > > [ The Types Forum,
> > http://lists.seas.upenn.edu/mailman/listinfo/types-list ]
> > >
> > >
> > > Dear Purush Iyer,
> > >
> > > what is a "grand challenge"? I naturally know the funding agency
> > > answer:
> > >
> > >  a tool for extracting a huge amount of money from politicians
> > >  who need something graphic to convince voters that their spending
> > >  is acceptable and produces 'good' things
> > >
> > > but I'd like to know whether there's a "scientific" definition
> > > that would make it worthwhile discussing the idea on a science
> > > mailing list.
> > >
> > > Thanks -- Matthias
> > >
> > >
> > >
> > >
> > >
> > > On Jan 16, 2010, at 5:04 PM, Purush Iyer wrote:
> > >
> > >> [ The Types Forum,
> > http://lists.seas.upenn.edu/mailman/listinfo/types-list ]
> > >>
> > >> Colleagues
> > >>
> > >> Now that the lek (er, POPL community)   has decided on vexing
> questions
> > >> about papers at POPL, I thought it best to consult this group on a
> > harder
> > >> problem.  In my current job at a funding agency I am confronted by the
> > >> question, from Physicists and Control Theoreticians, "What are the
> main
> > >> *scientific* grand challenge problems in Programming Languages and
> > >> Systems?"   I have not too successful to date; hence this e-mail.
> > >>
> > >> If y'all can come up with a list of five to ten challenges I would
> > >> appreciate it very much.  I don't take this question lightly; I had a
> > long
> > >> chat with Jens Palsberg over the summer and I don't think we were able
> > to
> > >> make much headway.
> > >>
> > >> Sincerely
> > >> Purush Iyer
> > >
> > >
> >
>