Hello

being out of office I have started to read all the contributions about
labelling and gave up halfway through. I noticed the mail by Filliatre
to which I subscribe without reservation. 

I AM A HAPPY USER OF THE CLASSIC MODE

and so far did not find it necessary to use labels. I should say that I
have written with collegues quite a lot of code as a part of the
compiler and programming system of sE (that is a language that extends
Java by synchronous programming - combining the Esterel, Lustre, and
Statecharts dialects, generating C for instance. By the way, interest
welcome). I extensively use partial evaluation for a rigorous
denotational style semantics for the synchronous sublanguage. In fact I
suspect that I could not have suceeded without partial evaluation (to
comment on another thread). In order not to loose track, a good naming
discipline is necessary (which is probably equivalent to use labels in
this context without, personal view, cluttering the code: how awful to
write ~f: if I know the argument is a function - by position and since I
tend to use names that mean something).

On the other hand, I have programmed quite a substantial gui using the
most elementary Tk-binding. What I did is to built my own little set of
Ocaml functions for hiding Tk. I admit that labels might have helped for
dealing with optional arguments, though I probably would have done the
same. 

Hence a my vote: 

- If you feel you have to add labelling keep it optional.

Axel

PS. This will be my only contribution to the subject

-- 

Dr.rer.nat. Dipl.Ing. Axel Poigne            http://ais.gmd.de/~ap 
                                             mailto:poigne@gmd.de
GMD - AiS                                    Tel: (+) 2241 142440
Schloss Birlinghoven                         Fax: (+) 2241 142324
D-53754 Sankt Augustin
Germany 

------------------------------------------------------------------
See what happened at the workshop on
"Formal Design of Safety Critical Embedded Systems" (FEmSys'01) 
   http://ais.gmd.de/~ap/femsys
-------------------
To unsubscribe, mail caml-list-request@inria.fr.  Archives: http://caml.inria.fr