On Tue, 11 Sep 2001, j h woodyatt wrote:
> This "phantom types" design pattern is one I have never seen before.  It
> doesn't seem to be used in the standard library anywhere I can see.  It
> looks like it might be useful in presenting a safer network programming
> interface than the low-level wrappers around BSD sockets (which I've
> never liked).

We used phantom types extensively  for exactly this purpose -- a typed
wrapper around a low-level interface. It seems that the trick has been in
use
already in the 80's for ML libraries that accessed sockets (allthough it
didn't
had a sexy name in that time). We tried to describe the general use pattern
in detail in the paper "domain specific embedded compilers" which can be
found at:

http://www.cs.uu.nl/~daan/pubs.html

We also used phantom types to encode single type inheritance for the
H/direct frame work (described in "calling hell from heaven and heaven from
hell")

All the best,
    Daan Leijen.


-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr