[FoRK] Now with magic pixie dust!

Stephen D. Williams sdw at lig.net
Fri May 21 19:35:16 PDT 2004


Gavin Thomas Nicol wrote:

>On Friday 21 May 2004 05:17 pm, Eugen Leitl wrote:
>  
>
>>I don't care about XML. I care about XML+SEXPR+hardware, in above context.
>>I'm also interested in whether XML carries *any* advantages over SEXPR in
>>that context.
>>    
>>
>
>I think the answer, as always, is, "it depends". Most of the binary 
>representations of XML/s-expressions I have seen tend to be optimised for one 
>kind of application over another. At the end of the day, the *syntax*, 
>especially in machine communication scenarios, is largely irrelevant to the 
>"user" (vs. the programmer, and even then, in most cases it is, or should be, 
>irrelevant). 
>
>I've had a few debates over the years about "infoset", "binary xml", or "SOAP 
>vs. RPC", or "binary XML in SOAP", and in 99.9% of the cases, the discussion 
>devolved to discussing a standardised domain-specific API that abstracted the 
>syntax.
>  
>
I think that an XPath based API is pretty general, with certain 
semantics.  You need to be able to get, set (create/replace), append, 
insert.  You need array indexing, array counting, iteration/enumeration, 
subtree operations (get, set, append, insert subtrees).
I've also found that you need something I call "scoped subtrees" in a 
good API.  This allows you to treat an XML-like tree as a collection 
that can be used, without deep copies, in an OO programming environment 
as object-scoped member data.

Compiling Schema a la IDL is bad.  Using schema as a baseline delta is 
good, and optional.  SOAP (as almost universally profiled and 
implemented) and RPC, and CORBA, share the same fundamentally bad 
communications model: synchronous, half-duplex, request response nature.

I feel that validation, both well-formedness and correct against a 
schema, should be optional when it isn't required by the loading 
process.  In an optimized alternative to XML 1.0, validation should not 
cancel all gain.  In other words, optimized format with validation 
should be better than XML 1.0, but if validation isn't required, it 
should be possible for applications to decide to avoid it some of the 
time.  Sometimes data is provably valid already.

sdw

-- 
swilliams at hpti.com http://www.hpti.com Per: sdw at lig.net http://sdw.st
Stephen D. Williams 703-724-0118W 703-995-0407Fax 20147-4622 AIM: sdw



More information about the FoRK mailing list