Non Sucking YAML parser

[Robin Berjon]

On Sep 17, 2006, at 14:27, Phil Pennock wrote:
> On 2006-09-13 at 13:49 +0200, Robin Berjon wrote:
>> If that's a problem for you, you're better off shipping C structs
>> around than even using YAML or JSON. One of the advantages of XML is
>> that you could swap in a binary XML format, and (assuming the means
>> for detection, in the same way that libxml2 will automatically read
>> from a gzip'd document) it'll just work.
>
> On this basis, you're better off going for ASN.1 specification;
> completely language neutral, very mature, has XML bindings so that you
> can use your ASN.1 schema for programming and use XML on the wire.

I have extensive experience with ASN.1. Out of the fact that I am  
good friends with some of the people behind it, and also because  
until the end of the month by chairship of the EXI WG requires of me  
total neutrality, I shan't describe this experience :)

But what I can say is that ASN.1 simply doesn't fly IMHO as a binary  
XML format because a) it requires a schema and b) it doesn't handle  
deviance from schemata at all. Both these facts make it inapplicable  
to at the very conservative least 80% of real world XML usage. It  
also has a few other issues depending on the encoding rules you use,  
e.g. BER is not sufficiently compact, but PER isn't streamable.

If you're interested in the requirements we're after:

   http://www.w3.org/TR/xbc-characterization/#N102EC

-- 
Robin Berjon
    Senior Research Scientist
    Expway, http://expway.com/