Tuesday, 7 May 2013

$10k to upgrade?

From: http://www.flickr.com/photos/72636606@N08/7678786610

Slashdot recently ran an excellent story about how Some Windows XP Users Can't Afford To Upgrade and it made me really rather upset. You see Windows XP is the bane of my development life as I have to sort out support for it on each and every project I seem to work on. When I'm making things for myself I simple sling BootStrap and jQuery in there and everything just works. I even use jQuery 2.0 with it's lack of support for older browsers! It's not just the extra bother that I have to go through that upsets me though - I'm also really sad for the poor users!

They end up being stuck with old hardware because they can't afford the shiny new software that they'd need in order to upgrade their hardware... poor sods!

The example quoted is of Optometrists practice, I think, and the specialised software she'd need in order to run her practice. The upgraded software was something like $10K which is a nightmare - it also left me wondering what sort of software that she might possibly require that would cost that much?

I'm doubting that she'll be controlling sophisticated machinery or anything like that - just recording the results of her examinations and scheduling appointments... not exactly rocket science. Sounds almost exactly like something we could crack within a couple of days at work or a couple of months of my Sunday's (Sunday morning are brilliant as 'er indoors goes to Church so I go to work and use the blindingly fast connection and machine there to do fun things - often alongside catching up with work things).

But this is interesting in all sorts of other ways... approach one Optometrist and build her her software for pennies, then go on and sell it to other Optometrists for more than pennies but significantly less than $10K. What a business model!

Everyone's happy - we build a solution that will be used in anger and has been optimised for the actual users of the system, they get a cheap alternative to their traditional route that works just the way they want it to work and everyone else gets a cheaper solution.

The only possible fly in the ointment is the old data - transferring the data from the old system is often a major pain. It might be stored in all sorts of odd formats, which may or may not be proprietary, but generally speaking it's possible to extract it as a csv file. But massaging that old data into our nice new system is often a pain. Thinking has come on in terms of databases in particular - is the data normalised? And to what degree is it normalised?

In the past I've spent ages writing parsers in PHP to take two or more distinct tables and de-normalise them. I'm pretty sure that there are faster alternatives though and I've heard good things about jitterbit, but this is basically the approach that's required to get old data into news systems unless you're very fortunate and the old data doesn't need cleaning. I'd not hold my breath that that's the case though!