?

Log in

No account? Create an account

Previous Entry | Next Entry

Why the ID database will never work...

...you just have to look at the problems people have with CRM systems...

Standard data cleaning metrics show that data in any contacts database degenerates at a rate of around 30% a year, just through people moving, changing jobs, getting new phones - all the trivial things we never really think of. It's a massive problem, and the companies set up just to help people clean corporate contact databases struggle to keep up with things.

Now imaging trying to keep the data of everyone in the UK clean. It's going to be virtually impossible. And the national database will be storing much more than just where everyone lives.

So, you have to ask yourself, are you able trust a national ID database?

Comments

( 12 comments — Leave a comment )
megadog
Jun. 30th, 2006 08:33 pm (UTC)
The perils of database cleaning.
A friend of mine some years back undertook a project for a major UK electricity utility - basically doing a database-clean on their customer list to remove dupes.

Result: something in excess of 600,000 hits [principally where they had Jane Doe, Jane E. Doe, John Doe and John Q. Doe all living at 16 Acacia Avenue, Witley Scrotum, Middle England, WR31 2BQ].

So far, so good. Except that the company's business-managers threw their hands up in horror with the realisation that - if the truth got out - it would appear that this particular utilityco had just suffered a massive slump in its number of customers [with consequent massive slump in share-price].

You can bet your sweet bippy that any government agency [or outsourced outfit like Crapita] running an ID-card-type database will *love* to retain dupes in their system on the basis that they can then claim "we have 50 million of the UK's 57 million population registered".

ccomley
Jun. 30th, 2006 08:59 pm (UTC)
specious - it doesn't matter if a national ID database doens't have your current mobile phone number or yahoo messenger nickname.
sbisson
Jun. 30th, 2006 09:09 pm (UTC)
I'm not sure - it's a matter of complexity. If a relatively closed problem domain can degrade at 30% per year, what will be the degradation on an open system with no clearly defined data model?
purplecthulhu
Jun. 30th, 2006 09:06 pm (UTC)
This problem has been legislated out of existence - the errors in the database are the fault of the record holder. If there is an error in your entry and you do not tell The Powers that Be that will be a 2000 pound fine or 2 years in prison.

This is in the law already. They don't say whether you need to know about the error for it to be your fault...
sbisson
Jun. 30th, 2006 09:10 pm (UTC)
Well, that's 30% of the population in jail every year...

I wonder what that'll do to the prison building programme...
andrewducker
Jun. 30th, 2006 09:39 pm (UTC)
It's an interesting one. I've long thought that the only way to keep that kind of data up to date was to give everyone access to update their data. Only you can imagine the fun that would cause...
surliminal
Jun. 30th, 2006 10:47 pm (UTC)
Make the ID database a Wiki!! Hell, stats show it'll be more accurate than central cleaning??
surliminal
Jun. 30th, 2006 10:48 pm (UTC)
they have to find you to prosecute you though. And your contact details will be wrong..
purplecthulhu
Jul. 1st, 2006 07:20 am (UTC)
Which would be a great justification for random spot checks on the street...
purplecthulhu
Jul. 1st, 2006 07:18 am (UTC)
Exactly - shows how well thought out the bloody thing is!
stillcarl
Jun. 30th, 2006 11:11 pm (UTC)
So, you have to ask yourself, are you able trust a national ID database?

But a national ID database is about fear, not trust. A fearful population is a compliant population.
(Deleted comment)
andrewducker
Jul. 1st, 2006 10:05 am (UTC)
The first of those three methods being the one most likely to work. The way to write effective big systems being, of course, to write effective small ones and then make them bigger.
( 12 comments — Leave a comment )