CLIM mail archive
[Prev][Next][Index][Thread]

Re: format directives

  Date: Thu, 3 Sep 92 14:47:52 PDT
  From: Bill York <york%oakland-hills@lucid.com>
  To: TYSON@ai.sri.com
  Cc: smh@franz.com, SWM@stony-brook.scrc.symbolics.com, 
      jcma@reagan.ai.mit.edu, clim@BBN.COM
  Subject: format directives
  Reply-To: York@lucid.com
  
     Date: Thu, 3 Sep 1992 12:21-0700
     From: Mabry Tyson <TYSON@ai.sri.com>
  
     As far as portability is concerned, all that should be hidden in
     CLIM:FORMAT.  (I'm guessing that all implementations of CLIM require
     you to use CLIM:FORMAT since the first arg might not be a CL:STREAM.)
  
  Actually, the ultimate goal is for there to be no CLIM:FORMAT.  This
  can happen once all the Lisps follow the same CLOS-based stream
  prototocl, and all the FORMATs are purified to use only the standard
  stream protocol to do their output.  This may not ever happen, but its
  a good theory, and I'm loath to invent a new reason to keep
  CLIM:FORMAT around.

I agree.  For performance reasons, when i'm not dealing directly with a
CLIM stream, i use (gross) LISP:FLOORMAT.  Personally, i'm not as opposed
to the "liturgical"

 (with-mumbling (mumble)
   (do-mumbly-something))

style, as i am with the

 (draw-mumble* :do :amazing :things :at :runtime :with :an :extra
  :gazillion :keywords :you :dont-need)

style.  Sure, a sufficiently smart compiler can compile them both away, but
only your LISP knows for sure.

On the other hand, i'm not opposed to a few good variants of FLOORMAT that
many people find useful for font bashing, or earcon mumbling.

k



0,,
References:
Main Index | Thread Index