It is July, 1999.  Scheme does not have records.  I have given up
   waiting for a consensus.  I thought that the whole point of the SRFIs
   was to make it possible to move forward without one.

That's *a* point of SRFIs. I think it is better to make some effort to get a
design that has consensus support. In this respect, SRFIs allow you to
*factor* the process of getting a standard together. And they allow people to
adopt standards on a finer-grained level, feature by feature. But that is my
opinion -- I have dragged out the SRFI-1 process by months in order to hash
out consensus, but I think it has been a worthwhile thing to do in order to
get a SRFI that (many) people (mostly) like.

We digress.

   You want an extensible syntax.  Matthias wants initialized fields.
   There is no end to what a record definition can be made to do.  Look
   at Common Lisp's DEFSTRUCT.  I do not believe that we can come to
   a consensus right now.

Fine. But simply by adding *a pair of parens* -- simple syntax, no extra
semantics -- you make it possible for later record proposals to *extend*
SRFI-9 in a fashion that allows SRFI-9-compliant code to run unaltered
in the newer systems.

Note that I am not pushing you to work in any extra semantics, such as
methods or initialisation fields.

OK, I've had my say; further comment would merely be repetitious. Note
that just adding a pair of parens would allow future record forms to
have backwards compatibility with SRFI-9.
    -Olin