06 October 2009

Open Source and the Fear of Failure

Yesterday I took part in an interesting event organised by BT called "Accelerating Enterprise adoption of Open Source Software" (disclaimer: filthy lucre was involved.) One topic that elicited much comment was why the public sector has singularly failed to deploy open source. As well as political issues (Tony Blair was and presumably still is manifestly in awe of (Sir) Bill Gates), there's another important issue to do with a fear of failure.

Nobody in government wants to take a chance on something new, so they stick with the old suppliers and the old solutions. When those (almost inevitably) fail, this causes people to be even more cautious, and so the vicious circle continues.

That's clearly bad news for open source, but here's a particularly good articulation of why the fear of failure is bad for governments more generally:

When I’ve spoken with government people, they confess a phobia of failure. Yet without the opportunity to fail, government – like industry and media – cannot experiment and thus innovate. We must give government the license to fail. That is difficult, especially because it is the citizenry that must grant that permission. I think government must begin to recast its relationship by opening up pilot procts to input and discussion, to smart ideas and improvements. I’m not suggesting for a second that every decision be turned into a vote, that law become a wiki. Government still exercises its responsibility. But it needs to use the new mechanisms of the web to hear those ideas. I would look for examples to Dell’s Ideastorm, Starbucks’ My Starbucks Idea, and Best Buy’s Idea Exchange.

Follow me @glynmoody on Twitter or identi.ca.


Matt Aslett said...

I think Mark Shuttleworth encapsulated the avoidance of risk perfectly when he said "The only thing worse than being locked-in to a vendor is being locked-in to a different vendor than everyone else".

glyn moody said...

Indeed: thanks for reminding me of that.

Fox said...

We discussed this at length at the post-event event - it is all about responsibility (as JP put it) and being able to attribute blame.

In an OSS implementation, you have to take responsibility for your own work, and if something doesn't work then it is up to you to fix it. In government, where IT projects inevitably don't work, the civil servant being able to cover their behind is a prime procurement objective.

Ironically, OSS enables you to fail faster, and fail more cheaply, than proprietary software.

More on my blog: http://foxonsoftware.blogspot.com/2009/10/bt-foss-and-national-programme.html

glyn moody said...

@fox: good points - thanks for the link

Anonymous said...

Posting as Anonymous for obvious reasons...

I'm a Sysadmin for a UK government department so I see how this all pans out, from upper-management strategy meetings down to day to day maintenance of the systems.

Software purchases are invariably led by the current vendors of choice - the shop-floor geeks are /never/ consulted to find out if a particular system/solution actually works or can be managed effectively. The only time I've ever seen a technically competent person making decisions is if a contractor's calling the shots for a particular project. Depending on your contractor, this can be a lottery (although my team has been pretty fortunate to date in this regard). Usually by the time anyone technical has been invited to contribute, the software's already been bought, so we just mop up and try to get whatever we've been given working.

Any software that touches any live data /absolutely/ /must/ have a support contract so that management asses are covered (understandable from their point of view, I suppose), and that rules out a large amount of very suitable and very high quality Open Source software - even though there are admins in-house who can support Open Source software better than the commercial vendors can support their offerings.

The basic problem is that those with the power to specify software purchases are unfortunately utterly unqualified to make such decisions.
Vendor pitches are taken at face value. Requirements are 'tweaked' to ensure that the favoured vendor's product comes out on top (usually because management are sold on a particular feature that makes their life easier). Technical issues are hand-waved.

That said, it's not all bad news. Most of the (Linux/Unix) admins I've met in government are actively sneaking Open Source stuff in through the back door. Linux machines pop up doing network and system monitoring, MySQL/Apache/PHP Wikis get deployed under the radar, that sort of thing (we keep most of this stuff secret until the risk of removing it is higher than the risk of keeping it ;) ). It's 'small beer', if you'll pardon the management-speak, but it's slowly adding up - it lets us show how good Free Software can be, without doing anything that would be considered a Business Risk. And good, Free software (even in small measures) makes admins' lives a lot happier.

Apologies for the rant - I've spent the last two days on conference calls with a (proprietary!) software vendor trying to tell them why their software isn't working. Again.

glyn moody said...

Please don't apologise - that's fascinating stuff, even if it sadly confirms my fears. Thanks for sharing it with us.

Fox said...

Thanks Anon, although sadly my 20 years in the software industry (mostly in the private sector) have taught me that the situation you describe is not limited to public sector organisations - I've seen bad management in Publishing, Banking, Telecoms, Healthcare and Media.

The fact therefore that it is also to be found in the public sector is not too surprising!

On the plus side, you can make things better by a) taking your FOSS skills elsewhere, or b) proceeding into management yourself, and working to improve the organisation from within :)

Anonymous said...

i can never understand why opensource isnt more popular in government (well apart from wanting to cover their own arses)
considering the EU is currently investigating microsoft (again) you would have thought they would want to keep microsoft at arms length for a start.
plus theres probably the cost savings
read this article on the LSE http://www.ibspublishing.com/index.cfm?section=news&action=view&id=13440
then of course theres all the electioneering sound bites they could use to sell it, 'freedom','open' etc etc

Anonymous said...

I'd say it's not even a fear of failure. Rather, it's a fear of being *blamed* for the failure. But given that (at least in the USA) getting fired from the government nearly takes an Act of Congress, the question is:

WHY are they afraid of being blamed for failure?

The answer, I've learned, is this. Governments don't give you bonuses or merit-based raises. And government salaries aren't typically as high as those in the private sector. Promotions don't offer you *that* much money. Therefore, the currency in government becomes power and perceived importance.

How do you get power and perceived importance? By getting bosses to like you personally. I've seen several jerks get promoted in the private sector, or get large bonuses, because they could do the job well. Not so in government.

It's the same in education. Parent-Teacher Associations have significant influence. One school tried piloting OpenOffice.org in a computer lab because it ran out of money. A kid told her parents, "hey, we're using OpenOffice! Teacher even gave me a copy to take home! Neat, eh?" The parents went hog-wild and made a huge stink ("WHY AREN'T YOU TEACHING MY LITTLE JANEY MICROSOFT OFFICE??"). That school's PTA backed these stupid parents. To shut them all up, the school cut a teacher position and bought a bunch of MS Office licenses.

Now you tell me who benefited there. The kid? Nope. The school? Wrong again. Only Microsoft benefited.

Jose_X said...

Avoid blame:

Eg1, take out a Red Hat contract. See:

Eg2, leverage the voting so that the blame can be passed on to the affected citizens. This might mean opening up a few more things and perhaps listening to further suggestions as you go underway.

Laws by wiki is a great idea, what are you talking about! [OK, the wiki would be a staging/incubating area and subject to formal procedures afterwards.]

Jose_X said...

Ooops, forgot to include the Red Hat link. It's a video interview hosted on cnbc. The link can be found here: http://boycottnovell.com/2009/10/08/ffii-red-hat-vs-swpat/ (as well as a link to a great FFII submission to the Bilski patent case).

.. alright. The actual Red Hat link is http://www.cnbc.com/id/15840232?video=1286390389&play=1

ajt said...

How different things are in France...


glyn moody said...

@ajt: great link, thanks.

Yes, the contrast is extraordinary.

Madhukara Phatak said...

Better customer support policy is required in oss world

glyn moody said...

Well, it depends on the area. Support for mainstream LAMP project is pretty good; some of the smaller are slightly less well-established. But I think things have moved on a lot from a few years ago.