On Tue, Jun 26, 2007 at 10:38:17AM -0700, Jason Thorpe wrote:
> 
> On Jun 26, 2007, at 10:16 AM, Thor Lancelot Simon wrote:
> 
> > if the encodings
> >are carefully chosen
> 
> The problem I see here is that the encodings *aren't* being carefully  
> chosen.  "SCN" is not compatible with any other data serialization  
> format used for this purpose, and that forms part of the basis for my  
> objection to it.

Reading what you wrote, I can't help thinking that you take the latter
("not compatible") to be evidence for the former ("not carefully chosen")
but one should note that you offer no argument establishing that that
is in fact the case.

When you say "any other data serialization format used for this purpose"
I'm not sure how tightly you want to constrain "this purpose".  If you
mean "for externalization of property dictionaries", it seems to me that
the observation establishes essentially nothing: after all, the original
library that supported these externalizations was not really intended for
direct interaction with humans, and so, of course, no format designed with
human usability as one of the principal desiderata is going to, in fact,
be shared with it.

Restricting the set of acceptable encodings that one will rather arbitrarily
label as "carefully chosen" to the set supported by some other version of
the NeXT/Apple code seems to me to be simply a backdoor way to ensure that
no additional encodings will ever be anointed as acceptable.  But this is
really just semantic quibbling about "carefully chosen" and establishes
precisely no conclusion as true.

If, on the other hand, you mean that SCN isn't quite like any of the other
C-like configuration file formats out there, I note that:

1) That's an inevitable consequence of being able to externalize arbitrary
   property dictionaries

2) I, at least, would far rather see one new, reasonably sane C-like
   syntax be established as a stake in the ground now, than ever have to
   learn five more "almost the same" variants in the future -- ala the
   so-called "ISC configuration file format" which as we both know (and
   hate) is actually at least 3 different formats because of lousy
   specification of its syntax and independent implementation.

-- 
Thor Lancelot Simon	                               tls@rek.tjls.com
  "All of my opinions are consistent, but I cannot present them all
   at once."	-Jean-Jacques Rousseau, On The Social Contract