On Tue, Mar 27, 2018 at 7:53 PM, Arthur A. Gleckler <xxxxxx@speechcode.com> wrote:
>> In any case I think the best "key" for a particular definition should
>> be as simple as `srfi-x--<identifier>` (note the double `-` separating
>> the SRFI identifier and the definition identifier.
>>
>> Or any other key that can be automatically derived given the library
>> and definition identifiers.  (This process also has to be reversible,
>> i.e. having the key, deriving the library and definition identifier
>> without any additional lookup.)
>
>
> How about an s-expression-based key, like ((srfi x) identifier)?

Indeed my `srfi-x--identifier` is directly derived from your proposed
s-expression syntax:

  (string-concatenate
      (string-join "-" (map symbol->string (first key)))
      "--" (symbol->string (second key)))

The reverse involves splitting strings, first on the double hiphen,
then the first part on hiphens.

BTW would the following library identifier s-expression be acceptable:

* for actual SRFI's `(srfi n)`, or `(srfi n submodule ...)` in case a
SRFI has multiple "submodules";  (I think some SRFI's, like the ones
exposing both procedural and syntactic layers, have something
similar);

* for RxRS `(scheme n)` or `(scheme n submodule ...)` for partial
implementations like `(scheme n lazy)`;  (the R7RS uses such a
notation, except that I introduced an `n` between `scheme` and `lazy`,
to be able to note different versions of the standard;)

Or perhaps for submodules we could use a third component in the key
like `((srfi n) (submodule ...) identifier)` and `((rrs 7) (scheme
base) cons)`, thus allowing the "module" part to be any valid object
as specified by that particular SRFI or standard.  (I think I'm in
favour of this third component, as it decouples the first component
which we can standardise upon, and the second component which is up to
the author of the specification.)

Ciprian.