[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] ratification vote for R7RS-small

Full name (required):	James Wise

Location (optional):	Carmel, NY

Affiliation (optional):	None (private individual in
                        professional practice)

Contact details (optional):  jwise AT draga.com

Statement of interest (not required if you registered for the
R6RS ratification or the 2009 Steering Committee election):

    I've been a Scheme user since 1993, as well as a student of
    programming languages of many families and idioms.  I have watched
    the scheme standard evolve from R4RS through R5RS and R6RS, and have
    watched the R7RS process with great interest.

    Since graduating school in 1995, I have worked as a system
    administrator and software developer, making use of a wide range of
    programming languages (including Scheme and other LISPs) for jobs
    small (glue coding) medium (toolsmithing) and large (development on
    large commercial software suites).

    Throughout this time, I have studied the history and development of
    programming languages.

    My interest is thus two-fold -- to see Scheme keep the features
    which have made it stand out since LAMBDA first became the ultimate
    imperative (and declarative, and GOTO), and to see Scheme develop
    the features it needs to be usable as a programming language for
    real work.  Without the former, there is no reason for there to be a
    Scheme; without the latter, there is no reason for there to be a new
    Scheme standard.

Vote (required):	NO

Rationale (optional): 

    The proposed standard fails on both of the above accounts: it fails
    to make substantive improvements in Scheme as a language, and it
    fails to stay true to the spirit of what makes Scheme what it is.
    In essence, the standard fails to justify its own existence.

    As a language, the proposed standard is a substantial step backward
    from R6RS in usability and fitness for purpose.  In the precision of
    its definition, and in the completeness of the features it does
    provide, it falls short of that standard as well.  (I am very aware
    of the R7RS small language / large language split; I am referring to
    the internal completeness of the features chosen for the small
    language in their own right, and their usefulness as a basis for the
    large language).

    As a Scheme, the proposed language differs relatively little from
    R5RS, and where it does differ, the differences do not "feel" true
    to the history and spirit of Scheme to me in a way that even the
    more sweeping changes of R6RS did.  I suspect both of these
    shortcomings stem from a desire to define R7RS scheme in opposition
    to R6RS, instead of as a natural evolution of the language's

    Last of all, I see the sharp disregard this standard shows for
    compatibility with R6RS Scheme as a move away from unity in how
    Scheme implementations add the parts of a modern programming
    language which R5RS lacked, and I see the sharp disregard this
    standard shows for the effort so many have put into standardizing,
    implementing, and extending R6RS as a blow to the united focus of
    the Scheme community.
    For the aficionado of Scheme in its original crystalline beauty,
    the new standard has little to offer over R5RS.  For the working
    programmer, it has little to offer over R6RS.

    I vote NO.

				Jim Wise

Attachment: pgpHDT75izPcL.pgp
Description: PGP signature

Scheme-reports mailing list