[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [xml-dev] Caught napping!
> -----Original Message-----
> From: Linda Grimaldi [mailto:lgrimaldi@neocore.com]
> Sent: Thursday, November 08, 2001 6:06 PM
> To: bradford@dbxmlgroup.com; Champion, Mike
> Cc: xml-dev@lists.xml.org
> Subject: RE: [xml-dev] Caught napping!
>
>
> One question that comes
> to my mind is that if the relational model is the only model worth
> considering, why is 70% of all corporate data still in flat files or
> VSAM/other hierarchical formats, despite the data management issues
> these formats present?
This very subject came up in his previous Against the Grain column (sorry, I
know I quoted this before):
"What is even more fascinating ... is that the most infamous hierarchic
DBMS, IBM's IMS, was developed exactly in the same way as XML ... Its actual
implementations were discarded in practice."
I innocently pointed out that according to IBM,
"More than ninety-percent of the Fortune 1000 companies use IMS. IMS serves
200 million end users, managing over 15 billion Gigabytes of production data
and processing over 50 billion transactions every day" and questioned
whether the term "discarded in practice" could be applied to IMS.
Mr Pascal's reply was, as usual, carefully reasoned and polite:
http://searchdatabase.techtarget.com/forums/0,289802,sid13_fid1,00.html
"I wish there was a way to say it differently, but there isn't: you do not
possess the intellectual ability to understand my arguments and repeat the
same nonsense which I have already debunked explicitly in my response."
Seriously, I think the "70%" statistic illustrates some truths we all have
to deal with:
- Very few organizations throw away working systems and live to tell the
tale, no matter how ugly the underlying code. Joel Sapolsky has a great
rant on this subject http://joel.editthispage.com/stories/storyReader$47
- Effectively normalizing a nasty structured document format like an
aircraft maintenance manual is several orders of magnitude harder than
normalizing a personnel database (the canonical example in most every DBMS
book).
- 500-way joins just don't work that efficiently on today's RDBMS systems.
- Your CFO doesn't care about the mathematical foundations of your database
design, only whether it gets the job done.