Re: [XCON] On encodings

["Pete Cordell" <pete at tech-know-ware.com>]

----- Original Message From: "Henning Schulzrinne" <hgs at cs.columbia.edu>

> Thanks for the summary. I'll just add that having the IETF, in a  small 
> niche of its portfolio, go out and invent a new text encoding  that nobody 
> else uses and is likely to use (sorry...) seems just  plain hubris. We're 
> no longer in the world of one-off systems where  we build every system 
> from scratch as if the only tool in our tool  kit was a C compiler (I 
> suspect some people here would actually  prefer assembler...).
>
> Every new encoding means new libraries, means new possibilities for 
> buffer overflows, new porting for every language from C to C# to PHP  to 
> Java to Perl to Ruby, new training, new debugging tools and new 
> interoperability testing. I can choose among dozens of XML and SOAP 
> libraries, maintained by the largest IT corporations on the planet if 
> that's my choice, and a corresponding number of tools, from open  source 
> to XMLspy and kin. Why in the world would a sane developer  create their 
> own new niche encoding when the rest of their system is  going to be built 
> using modern software tools and will already have  to support XML anyway?
>
> I'm afraid I'm not going to be spending time on Lumas analysis, I'm  sorry 
> to say. The IETF is a tiny part of the software universe and  needs to 
> learn to play nice with the 99.9% of the developers that  don't show up at 
> IETF meetings.

But people still seem to be happy to invent new binary encodings such as 
BFCP, AVPs, and new ABNF grammars such as MRCP that have all the problems 
that you mention above.  Why is that?

People are even pushing JSON, when they have access to all the XML tools 
that you mention (http://en.wikipedia.org/wiki/JSON).  This is used 
extensively by the likes of Google and Yahoo even though they have all the 
resources to run XML services.

Like you, I would like the encoding debate to go away as well.  Your 
approach is to pretend that there's no problem.  My approach is to design a 
system that has the high-level design benefits of XSD, the debuggability of 
a text protocol, and the compactness of a binary protocol.  With Lumas there 
would be nothing to debate, and we could get on with more interesting 
things.

It seems I also have more faith in the development capabilities of 
developers.  If Microsoft (and Borland) can knock out C# then they're not 
going to be bothered by Lumas.  Heck they can get the C and Perl versions 
from my web site now.

Don't get me wrong.  XML is great.  But it is not the be all and end all. 
In programming I use C++ for some things, Perl for other things, PHP/SQL for 
yet more things, C# for other other things.  It depends on the problem I'm 
tackling.  Why shouldn't I have the same choices for message definition?

Pete.
--
=============================================
Pete Cordell
Tech-Know-Ware Ltd
                        for XML to C++ data binding visit
                        http://www.tech-know-ware.com/lmx
                        (or http://www.xml2cpp.com)
=============================================
. 



_______________________________________________
XCON mailing list
XCON at ietf.org
https://www1.ietf.org/mailman/listinfo/xcon