Full name : Emmanuel Medernach
Location : France
Statement of interest :
I first learned about Scheme because I was (and still are)
dissatisfied with mainstream languages. Many of them feel
to me as convoluted hack, making ideas laborious or verbose
to express and making people think coding is mainly about
tinker instead of thinker. On the contrary Scheme
immediately strucked me because of both its elegance and
expressiveness. I then carry on reading the famous Lambda
papers (and some other papers from the readscheme website)
which were enligthening and great readings.
So Scheme is my language of choice for expressing my ideas
and fast prototyping them. And even if in some project I
cannot use it as it deserved to be (my job is to work as a
computer research engineer in physics lab), I try to follow
its philosophy each time I could.
Vote: yes
Rationale:
- Does it satisfy the requirements of the WG1 charter ?
I believe the followings are true:
"Extensible"
"Facilitate sharing of Scheme code"
"Intended for use in education, programming-language research, etc."
"Easy to learn and understand"
"Mostly compatible, and comparable in size, with R5RS."
About the library decisions, I think that removing
versioning and phasing from WG1 was a "necessary" step to
have a simpler module system. The module system was chosen
to be static which make it a good base for language
extension.
About the error system decision IMHO the main R6RS issue was
to write in stone an arbitrary condition hierarchy, which
were incompatible with everything existing at that time and
thus too early to standardize.
About the record system decision, IMHO the R6RS one feels
like a "least common multiple" approach was taken instead of
a "orthogonal concept decomposition". I am convinced that
Scheme needs diversity (of record system for instance)
instead of unification.
- My view about the work of Working Group 1
We had to deal with a sensitive subjet as to find a balance
between existing implementation practices and
standardization effort. Concessions were inevitables. As
such I think the voting process has been a success as we all
had the opportunity to express opinions and to read others
point of view, which is interesting and enriching.
Of course I have a list of issues: I am not quite happy with
the interaction between exception system and dynamic
wind/continuations (somewhere something orthogonal must be
missing). I strived for removal of the dreaded "unspecified
value" in favor of allowing no value to be returned, but
without success for now. I view making SRFI-9 "The One True
Scheme Record" as a bad decision, I admit it is widespread
and should have been standardized but as a module instead.
I dislike having parameters inside the core language:
however convenient there are I think promoting them is a bad
idea.
But all in all I am very satisfied with the outcome and want
to give thanks to all participants for their work and
patience.
--
Emmanuel Medernach