[Chicken-hackers] Re: Serialization of tinyclos objects in Chicken Scheme

[Tony Sidaway]

I wrote

"Thus I suggest that, of the 256 possible byte-sized data type tags
(only 20 of which
are in use), 64 should be set aside for user-implemented compact
encoding of tinyclos objects. "

Actually the word "tinyclos" is unnecessary there. There's no reason
why the user (that is, the application writer) should feel constrained
to use tinyclos at all.  The important point is that a good proportion
of the data type tag namespace should be set aside for use by
individual applications, so as to encourage the compact encoding of
application-level serialized data.