Re: [mmox] XML serialization

[Jon Watte <[e-mail]>]

Hurliman, John wrote:
> length integer encoding (such as the base-128 varints used in Google Protocol Buffers: http://code.google.com/apis/protocolbuffers/docs/encoding.html) would be a nice fix for the binary encoding.
>
>   

That needs a fix for signed-ness, though. My proposal is to encode the 
sign bit in the 6th bit of the first byte IF the value is less than 0 or 
greater than 127 (keeping the encoding of small positive integers to a 
single byte). For negative numbers, encode the one's complement and the 
sign bit, and negate when done. This nicely sidesteps problems you'd get 
when trying to encode two's complement in a variable-width container. 
But that might be a slightly premature technical detail ;-)

> -1 on this. The LLIDL syntax is very easy to parse, and I posted a simple grammar file to generate a lexer/parser. Introducing a new XML schema creates just as much implementation 

My point is that XML for a schema language is even easier to parse:

xd = new System.Xml.XmlDocument();
xd.LoadXml(...);

Done. And this allows various code bases to easily adopt various levels 
of knowledge of the serialized data using existing tools, instead of 
having to add yet another kind of language parser with yet another set 
of mutually ambiguous or buggy implementations.

Sincerely,

jw