RE: [xml-dev] Multidimensional XML

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

I don't think it reasonable to approach XML For Analysis without some

background in OLAP, and OLAP isn't easy.   I don't know if that makes

XForA harder than other XML approaches to OLAP because I haven't

seen any.  The papers Fuller referenced included a slideshow

from Yannis Stavrakas

 

http://www.dbnet.ece.ntua.gr/seminar/01-10-29_mssd_ys.ppt

 

that was illuminating in that it extended XML DTDs to include keywords for

things like <!MULTIDIMENSIONAL.   That's a shocker to the XML 1.0

crowd.  The examples of how the author proposes to include dimensionality

into schemata/DTD are intrigueing.

 

I don't think of UBL as groundbreaking.  The concepts there have been

around for ages and ages.   I think of it as well-done craftwork with

a toleration for the tedium of negotiation in committee, admittedly,

rare skills.  In other words, crunching it down to spare structures,

but not having to innovate new concepts.  I could be wrong about that.

I don't pay a lot of attention to UBL because that part of my career

was over ten years ago, although, we would be very happy to see

crime syndicates using UBL. ;-)

 

Usually one sees OLAP where one has to data mine and analyze

large sets of context dependent information.   It is something I

wanted to bring up to the HumanML folks in the future as an

approach to some of the problems of semiotic sign systems.

 

Anyway, SVG looks daunting to some graphics artists, so I guess

preparation is everything.

 

len

-----Original Message-----
From: bryan [mailto:bry@itnisk.com]
Sent: Friday, August 09, 2002 8:46 AM
To: Bullard, Claude L (Len)
Cc: xml-dev@lists.xml.org
Subject: RE: [xml-dev] Multidimensional XML

           

 

>As in "too complex for work" or "too complex for work with multidimensional data"?

>The first is "too loose to decide" and the second requires OLAP expertise.

            Somewhat agreed,  although I think there is a third condition which is the one  applying to me of

            “too complex to use as springboard to the technology” , that is I felt from dealing with it that I had to back out and start from OLAP, this is like starting from tex (or some other layout technology) to deal with xsl-fo, may be a good idea but shouldn’t be necessary in my opinion. I realize this is arguable, I can sit here and argue it to myself, nonetheless I hate it when a domain-specific application of xml is so insular that I cannot come to a clear understanding of the domain from reading about the spec.

 

 > OLAP is one of those neat technologies that doesn't get much attention in the XML

>world.  I've always considered XML for Analysis a ground breaking and much

>ignored innovation.  Locally, I find when I inquire about it that the SQL practicioners

>want to ignore OLAP as "too hard and we don't really need it".

            I do consider OLAP from what I’ve learned of it in my spare time, a neat technology, I also can’t help thinking that for my SQL practice it is “too hard and I can do without it” if such technical schizophrenia is acceptable,  I’m pretty sure my co-workers would be divided into the “too hard and we don’t really need it” and the “shut up Bryan” camps. As for the groundbreakingness of XML for Analysis I have some feeling for the specification, complicated by its near unreadability(for one of my average abilities, there are no doubt people out there for whom it is a transparent spec) and my need to go cross-referencing all the time and thinking “hell I will never use this”. For me when I think  of groundbreakingness in XML I tend to think of stuff like UBL , things that I feel I will one day come to use and applaud. Does this seem wrongheaded?

 

 >So one might

>approach this first by enumerating the cases for which OLAP is the simpler

>approach and then the subset of these for which XML For Analysis is the

>simplest approach.

I think that would be cool, and a useful way of evaluating many of the standards that come out. In some way this is what some threads on this list attempts to do for various specifications, it would be nice if these comparisons were on some site together

- could call it www.xmlcomparisons.org

Maybe some Microsoft people on this list have specific insights on the matter and would like to do concise evaluations?