I’ve written before about the myth of software support. Software mostly doesn’t work, vendors mostly don’t fix stuff, and the world keeps turning anyway. Everything’s broken and nobody’s upset.

So why do companies stop using perfectly good stuff (hardware and software) simply because they can’t buy support from the original vendor? This position seems nonsensical to me, but I’ve come across two examples in the past week:

IRIS

The Iris Recognition Immigration System (IRIS) was a perfect example of trading privacy for convenience. I let the UK government take pictures of my eyes, and in return they let me get out of passport control quicker by going through an automated gate rather than standing in line. You didn’t have to be a UK Citizen to use it, so it was particularly beneficial to frequent travellers from outside of the EU who might otherwise face a long line at passport control.

One of my friends complained the other day about IRIS not being available, and quickly got an answer from another friend about the reason:

The machines are still there. I’d bet that the machines still work. But somebody is too scared to turn them on in case they break because Microsoft won’t any longer be on the hook to ‘support’ them. I can make two obvious points at this stage:

1. The appropriate time to stop using the machines is when they actually stop working.
2. If they had stopped working when the ‘support’ was still in place then it would still have taken days (at least) to get a fix.

What is it about software support that causes such insane risk/benefit trade offs?

Part of the problem is of course ‘security’. Microsoft ending support for XP means no more patches for vulnerabilties. This is a problem if you’re an Internet connected PC surfing the web. This shouldn’t be an issue for an applicance with some embedded software.

I’m left wondering how much the UK Borders Agency paid MS for support for IRIS, and how much utility they ever got out of that contract? I expect the answer is something along the lines of too much and not enough.

Before moving on it’s worth noting that IRIS has been partially replaced by new automated gates for those with EU issued RFID enabled passports. This is no good to Kirk, because he has a US passport, and it’s turning out to be no good to me because the regular travelling public has a much higher moron that can’t work things ratio than those that opted into IRIS.

Update (26 Jan 2013) – Kirk tweeted yesterday evening that the IRIS machines are back in action (at least at Heathrow T5). Perhaps somebody saw sense.

Pano Logic

The next example I’ve come across is Pano Logic, who made a thin client device (which they branded ‘Zero Client’). It seems that at least one of their customers has decided to trash over a hundred of the devices as their IT support guy is now looking for creative ways to hack the hardware.

Presumably the day before support ended the devices were working fine, and the day after they were still working fine. The lack of support shouldn’t be an issue until something actually goes wrong. The main causes of going wrong are installation issues (which had already been passed) and age issues (when component failure comes along) – hence the bathtub curve. The age end of the bathtub isn’t all that often an issue in IT because we tend to replace stuff long before it falls apart, though the Pano Logic device might have been an exception – I’d expect that obsolesence would catch up with them due to changes in monitor technology or something like that.

Conclusion

The ability of IT vendors to support things is poor at the best of times. Withdrawal of support shouldn’t be an immidiate trigger to get rid of something – it’s just one of the indicators that it’s time to draw up plan B for when failure actually does occur.