Pandering to Lisp

[Dave Cuthbert]

Drat.  I wish I had discovered this a year ago.  :-)

Anyway, I've developed a separate, proprietary protocol (Wombeyan) for 
our products here, but its spec has a lot of similarities to D-Bus.  I'd 
like to investigate moving over to D-Bus instead -- while Wombeyan has 
been fun to work on and is useful, it's really not our core competency...

The main difference in the Wombeyan design is that it panders 
extensively to SKILL (Lisp derivative) -- the intent being to integrate 
with the Cadence products.  Notably, it omits a distinct boolean type, 
unsigned types, 64-bit ints, and dicts.  On the flip side, it includes a 
distinct symbol type as well as (heterogeneous) lists.

None of these is a show-stopper by any means.  However, I'd like to hear 
what others might suggest for mapping symbols and lists to D-Bus.  My 
immediate reaction is to use the custom type, but I don't want to 
(ab)use it if there's a better solution.

Mapping to booleans is, sadly, problematic.  While the 't symbol is 
obviously equivalent to true, nil is both false and the empty list.  In 
practice, this hasn't been an issue, even in Java.

Anyway, looking forward to seeing how D-Bus plays out.