Lists Home |
Date Index |
I strongly and emphatically disagree with your assertion that " There is no
such thing as a "rigorous scientific proof"..."
I cite physics as a prime example. Various proofs, including proofs of the
theory of relativity and the speed of light, for instance.
In this context (XML, RDBMS) I will state the specifics as:
- a test bench or test bed of consisting of computer(s) running Windows or
- an XML parser (approved by or compliant with W3C specs).
- a dbms such as Oracle
- an application specification and requirements
- a definition of terms including best practice, best system, cost factors,
essential items, desirable items, and specific yardsticks or measurement
areas, prediction in advance of measurement results for all areas,
measurement areas to include technical (performance, reliability,
usability, complexity) and non technical areas such as costing, ease of
use, ease of interface to other non technical systems by non technical
people (ad hoc integration).
- applications that satisfy that specification and requirements written
1) solely in XML (HM)
2) solely in dbms (RM)
3) as a combination of XML and dbms technology (HM mapped to RM and back).
Tests or proofs would consist of:
- initial cost factors for 1, 2, and 3 above
- support and maintenance cost factors for 1, 2, and 3 above.
- performance test results per the specifications and requirements
Applications I suggest would be good to test are:
a) company phone book for company with six offices in four locations across
boundaries such as states or zip codes or phone area codes.
b) contact system similar to various commercial systems like ACT
c) multimedia system, such as movies and soundtracks and musical scores to
plays and operas, and text of those movies, screenplays, operas, and
reviews of these in a searchable format (where was this line used... )
d) financial data such as a corporate book of accounts
e) document data such as newspaper text
f) mixed media historical data such as all 20/20, news casts, newspapers,
movie news reels, Nightline and BBC world news shows.
When people say no system is like their system, they just have not seen
enough systems. I respectfully submit that endeavours towards these proofs
will be highly valuable to the advancement of the science in XML as well as
in DBMS work yet to come. Theory is fine. Practice is harder. Generating
standard test cases, for systems such as the above, will be well worth the
effort as it brings us together to solve problems in a coherent and
systematic way, with open source offered to any and all who wish to
participate or contribute.
Without doing a standard set of proofs, I just do not see how the
discussion can advance effectively. Of course people get frustrated, of
course people get emotional, of course people are talking at cross purposes
most of the time !!!
But it does not have to be that way. Let's get a set of proofs, such as the
above, underway world-wide, right now.
If we are all on the same page, we can speak much more effectively and
address the real issues much more directly than is usually the case today.
Just an idea.
At 11:51 PM 8/25/2003 -0500, Bob Foster wrote:
>From: "John Cowan" <email@example.com>
> > pop3 scripsit:
> > > Nor can I support folks in a CMM or ISO9000 shop
> > > utilizing XML to any significant degree until they can show that XML is
> > > proven best practice, by rigorous scientific proofs.
> > What programming language can you justify by "rigorous scientific
> > [meaning mathematical?] proofs"? Scheme and Prolog, maybe.
>Well, exactly. There is no such thing as a "rigorous scientific proof".
>Asking for one just reveals a misunderstanding of science. You can have a
>rigorous mathematical proof, or a scientific theory that agrees with
>empirical observation and is consistent with other theories, but you cannot
>mix the two and somehow arrive at a proof of a theory.
>The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
>initiative of OASIS <http://www.oasis-open.org>
>The list archives are at http://lists.xml.org/archives/xml-dev/
>To subscribe or unsubscribe from this list use the subscription