On Tue, Sep 29, 2015 at 9:47 PM, Per Bothner <xxxxxx@bothner.com> wrote:

On 09/29/2015 05:07 AM, John Cowan wrote:
Per Bothner scripsit:

Using SRFI-10 has the known problems - and looks ugly IMP.

*All* unfamiliar lexical syntax looks ugly: this is a psycholinguistic fact.

No doubt.  Which is to be preferred:

#,(hash (re . 3) (im . 6))

or:

#hash((re . 4) (im . 4))

Not a huge difference, agreed.
I think it would be good to standardize *something*, preferably
something already in use.  Co-existing with quasi-quoatation
in some sane way would also be desirable: in a pinch alist->hash-table
could be used, but it is mildly verbose.

I'd like to see something that:

  1. is simple and flexible
  2. allows easy definition of the "obvious" read/write for records
  3. handles record uuid's properly (see gambit and chibi)

Beyond that I'm not too concerned about syntax.

-- 
Alex