[
Lists Home |
Date Index |
Thread Index
]
- To: "'John Cowan'" <jcowan@reutershealth.com>,"'Rick Marshall'" <rjm@zenucom.com>
- Subject: RE: [xml-dev] Query decomposition [was: What niche is XQuery targeting?]
- From: "Michael Kay" <mike@saxonica.com>
- Date: Wed, 15 Dec 2004 13:01:51 -0000
- Cc: "'XML Developers List'" <xml-dev@lists.xml.org>
- In-reply-to: <20041215122702.GA8199@skunk.reutershealth.com>
- Thread-index: AcTiodS9YDf9sJDKTqKPbzpABbIQ4gAAmO4A
> The bank I used to work for estimated that about 80% of all accounts
> had errors in the "indicative" (non-monetary) data: address,
> phone number,
> SSN, etc. Part of the problem, of course, was that the bank's
> systems had no concept of "customer", so if a customer had multiple
> accounts, this information was stored multiple times.
Back to the intermittent permathread on data integrity.
Errors in data arise most often because the data was right once, and isn't
right any more. Integrity checks performed when data is updated can never
catch errors that arise because it isn't updated when it should be.
When British Telecom published their phone directories on the web, I
discovered that I was still listed at an address where I lived for a few
months about 25 years ago. When I rang to point out the error, they informed
me that the phone there was still registered in my name, and that I had an
excellent payment record. It appears a succession of new tenants had left
the phone in my name to avoid paying a reconnection charge. So BT haven't
the faintest idea who their customers are, even if they have been paying
their bills regularly for 25 years.
Michael Kay
http://www.saxonica.com/
|