On Wednesday 25 August 2004 01:06 pm, Jason Diamond wrote:
> By the way, Clark, Oren, and Brian: if you want me to cease discussing
> SSYN on your list, let me know and I will. I can start a project on
> SourceForge and use a mailing list there. (I'd just hope that some of
> you would like to come and play for a while.)
> -- Jason
May I ask that the list not at least in some form. I am feeling a little
intimidated writting this, but I will give it a shot. Personally I feel
Jason has touched on an important milestone for YAML. I can say that in my
projects case I forced the perl YAML emitter to emit only blocks, because the
python yaml parser did some crazy things with implicit types. Now I know the
implicit thing is behind the project now, but it does reflect on the
dificulties around parsing yaml, and interpritation by the developer. The
patches I made to PyYaml were somewhat time consuming and difficult to fix.
I am not trying to blame the author, or the spec, just pointing out an
experience with that difficulty. I have found emitting yaml easy as pie, but
parsing it, well not so easy unless you reduce down to only folding blocks,
etc like I did.
I guess what I am trying to say is I am all for a simplification, etc to ease
the building of parsers, because good versions of these tools I needed
yesterday and don't exist today. Python, Perl, PHP, Ruby, shell should all
have parsers and emitters that implement the spec fully. One advantage of
xml currently is that this is true, so yaml looses out just because a proper
parser/emitter does not exist, but yaml would be a much better solution. I
believe this is a hurdle that yaml must overcome before wider adoption.
Which is something I would like since I don't want xml forced upon me
everywhere I go.
A users perspective.
Frustrating Hanging Crashing
Blue Screen of Death