[Omega-devel] StatDataML

Hi,

I just had a very quick look at the StatDataML proposal --- nice
work!   At the risk of showing my ignorance, I want to mention 
my first impressions.

My first impression is that defining datasets in terms of
arrays and list is a bit too high a level.  What about 
simpler vectors, scalars? (I know that R/S don't have scalars,
but other systems/applications do.)  Can we think of a core
set of "basic" data types (factors, strings, integers, etc.)
from which to build on other, possibly recursive types (perhaps
similar to corba's IDL basic data types or S's datadump?).  
Would it make sense to imagine, say xlispstat/python/java applications 
reading  and interpreting an StatDataML document without serious difficulties? 

My gut feeling (which is often wrong) is that the DTD should make
the data self-describing:  e.g., the factor "machineId" has 
levels (or defining set) "Stepper1", "Stepper2", ... "Stepper20", 
eventhough the particular dataset at hand has only a
subset of those.  Similarly, perhaps allowing  units and classes
to be included in the dataset (in the case of currency, it is certainly 
a number, perhaps single precision, perhaps not, with specific units 
dollars, euros, pesos, etc.)

More long-term, how about application-defined data?  Application may have
it's own set of data objects that fully exploits contextual 
information that could be extremely useful to capture and 
communicate.  Also, do the data have to be in ASCII format?  What about 
(possibly mime-encoded) images? sound?

As I mentioned, these are questions coming from my lack of experience
with XML, but may be worth raising now better than later :-)

David A. James
Statistics Research, Room 2C-253            Phone:  (908) 582-3082       
Bell Labs, Lucent Technologies              Fax:    (908) 582-3340
Murray Hill, NJ 09794-0636
-----------------------------------------------------------------------------