At 01:57 PM 7/31/2001 -0400, Clark C . Evans wrote:
Third, notation is everything.  Indeed it has been
said that our current set of mathematical notation
is reponsible for the bulk of the growth in the
sciences in the last 200 years. 

Well, I'm slowly warming up to it.

By comparison XML has... [...]
And we have just gotten started... these certainly
have not hurt the acceptance of XML.

I think XML succeeded for reasons that have little to do with technology and that probably won't apply to YAML.  For YAML, the technology really has to sell itself.

I don't think that we need to grow the language
all that much more (infact it would be harmful).
The indicators of XML are pretty darn constant
and yet everyone things XML is extensible.

I think XML's only means for extension is reserved attributes.  The W3C can and has extended the language by adding more reserved attributes, without changing the fundamental syntax.

But in retrospect I agree that XML would have been better off if that extension mechanism didn't exist, provided that its absence would have altogether prevented like extensions.

Without a means for extension, though, at some point you seal your fate.

Anchors and references are needed to serialize a graph,
which is the YAML information model.   A tree is not
sufficient.  How many XML specs have had to invent
their own notion of anchors and references (id/href)
to layer a graph structure on top of the XML tree
structure and how many of these are slightly different
enough to make it confusing when switching between
vocabularies and/or APIs?  This is one lesson I don't
think we have to re-learn.  *evil grin*

I just spent time preparing a meticulous counter-argument and ended up solidly convincing myself that you are right.  <embarassed-flush>  ID/href provides only benefits and no drawbacks.

| So there you go.  I realize that opinions on these matters will
  vary.  But still, because I find the good very good and the bad very
  bad, I find myself in a love/hate relationship with YAML.

Understood.  Which things bug you the most?

Wow, looks like you've mostly knocked them down.  My gripes are much more specific now:

(G1) Inflexible formatting/forced indentation
(G2) Type and format support in the core (not convinced yet)
(G3) Overloaded operators (due to type and format support)
(G4) There are two different list notations: ---- and @
(G5) @ and % do not suggest list and map to me

(G1) and (G2) bug me the most.  (G2) will probably later keep us from layering YAML-Schema strictly on top of YAML-Core, which could introduce problems akin to those of DTDs.

I think I'm starting to love YAML more than I hate it!