Power breakfast
This morning I was at a breakfast meeting at the Scottish Parliament, hosted by Patrick Harvie MSP, about how to promote the adoption of free software (also known as open source software) by the Scottish public sector. Patrick's been pushing this for years, but keeps meeting a wall of incomprehension and indifference. The meeting was billed as a briefing by the Free Software Foundation's general counsel, Eben Moglen, but, while he did most of the talking, it was more of a round-table discussion. I'm not entirely sure what I was doing there; I wasn't able to contribute much, but I had a very interesting time and learned a lot. Here are my notes from the meeting - from memory, so not in any particular order.
Scotland's a great place to work on public sector adoption of free software: it's small, has a burgeoning tech industry, and a government procurement infrastructure small enough that it's feasible to set up a special meeting for all the key players rather than trying to hijack an existing event (as has to be done in the States or Canada). Apparently, Scotland has a simpler procurement setup than many Australian state governments. It's probably the most "vulnerable" English-speaking jurisdiction, and is next to the large UK market. The UK government was previously not thought to be worth working on since Gordon Brown was so completely an MS fan; it's not clear that the new lot will be any better, but if they are then Scotland offers a viable way in.
We need a pain point to sell free software to the public sector, and the obvious one is the 25%-ish budget cuts that are coming in November (with a preview in September). We need to stress the "free as in beer" part of "free" for a change, because "free beer is all that most governments can afford to drink at the moment". Eben mentioned VoIP (using Asterisk) as an obvious cost-saver: "there's no reason for any government to pay for another telephone call". Alastair (from Valley Technology) and/or Paul (from the Scottish Society for Computers and Law) mentioned that open source reduces business risk (your supplier has gone out of business? Never mind, you have the source code), and hence insurance premiums - insurers (the smart ones at least) are starting to really get on board with this aspect of open source.
There are key people at most levels of the procurement chain who Get It, in particular the Scottish Government CIO (who investigated free software as a result of a question Patrick asked in Parliament in 2007), edit and librarians (including the Scottish Parliamentary Library's librarians) but they're largely working in isolation. We need to get them all into a room together to work out a unified strategy. Lock-in to long-term contracts isn't much of a problem here, as it is elsewhere: there are long-term contracts, but there's almost always something up for renewal soon. The CIO has gathered detailed figures on this stuff.
EU non-preferment rules are possibly a problem: above a certain size, contracts have to go out to UK tender, and above another certain size, they have to go out to EU tender. I'm not entirely sure how that's relevant, to be honest. We could possibly make a case that free software (or rather, specific bits of free software) is an indigenous Scottish thing - it allows much more scope for local support companies, for instance.
Currently Scotland has lots of free software-aware geeks, but not enough (?) large-enough companies explicitly offering support or supply of free software systems. We have a demand/supply chicken/egg problem. Eben said that the best way to ensure long-term supply of free-software support services is to think five years out, and get free software into high schools. Until that cohort graduates, you may have to pay up to 30% more for Linux support over MS support, but after that it will get cheaper than MS support. However: the Scottish government currently has a £6m supply-and-support contract for their desktops with a company in (?) Barrhead which does all the support themselves because Microsoft don't offer the necessary turnaround time - moving to Linux could be a big cost-saver for them, even in the short term.
They tried introducing free software in schools in Iceland, where the Minister of Education is a left/Green who Gets It, and found that they can be stymied at the school-sysadmin level: where there was a free software-aware person in the school, they succeeded, where there was someone with a wall of Microsoft certifications they failed. Possibly some sort of negotiated discount on Linux certifications would help to break the "sysadmin's career is locked into Microsoft" deadlock. Private sector adoption of free software largely worked by samizdat: low-level geeks installed free programs because they were free-as-in-beer and thus required no purchase order, and by the time the CxOs heard anything, the company depended on it. In the public sector, this doesn't work: change has to come top-down, but requires bottom-up buy-in.
Scotland's a great place to work on public sector adoption of free software: it's small, has a burgeoning tech industry, and a government procurement infrastructure small enough that it's feasible to set up a special meeting for all the key players rather than trying to hijack an existing event (as has to be done in the States or Canada). Apparently, Scotland has a simpler procurement setup than many Australian state governments. It's probably the most "vulnerable" English-speaking jurisdiction, and is next to the large UK market. The UK government was previously not thought to be worth working on since Gordon Brown was so completely an MS fan; it's not clear that the new lot will be any better, but if they are then Scotland offers a viable way in.
We need a pain point to sell free software to the public sector, and the obvious one is the 25%-ish budget cuts that are coming in November (with a preview in September). We need to stress the "free as in beer" part of "free" for a change, because "free beer is all that most governments can afford to drink at the moment". Eben mentioned VoIP (using Asterisk) as an obvious cost-saver: "there's no reason for any government to pay for another telephone call". Alastair (from Valley Technology) and/or Paul (from the Scottish Society for Computers and Law) mentioned that open source reduces business risk (your supplier has gone out of business? Never mind, you have the source code), and hence insurance premiums - insurers (the smart ones at least) are starting to really get on board with this aspect of open source.
There are key people at most levels of the procurement chain who Get It, in particular the Scottish Government CIO (who investigated free software as a result of a question Patrick asked in Parliament in 2007), edit and librarians (including the Scottish Parliamentary Library's librarians) but they're largely working in isolation. We need to get them all into a room together to work out a unified strategy. Lock-in to long-term contracts isn't much of a problem here, as it is elsewhere: there are long-term contracts, but there's almost always something up for renewal soon. The CIO has gathered detailed figures on this stuff.
EU non-preferment rules are possibly a problem: above a certain size, contracts have to go out to UK tender, and above another certain size, they have to go out to EU tender. I'm not entirely sure how that's relevant, to be honest. We could possibly make a case that free software (or rather, specific bits of free software) is an indigenous Scottish thing - it allows much more scope for local support companies, for instance.
Currently Scotland has lots of free software-aware geeks, but not enough (?) large-enough companies explicitly offering support or supply of free software systems. We have a demand/supply chicken/egg problem. Eben said that the best way to ensure long-term supply of free-software support services is to think five years out, and get free software into high schools. Until that cohort graduates, you may have to pay up to 30% more for Linux support over MS support, but after that it will get cheaper than MS support. However: the Scottish government currently has a £6m supply-and-support contract for their desktops with a company in (?) Barrhead which does all the support themselves because Microsoft don't offer the necessary turnaround time - moving to Linux could be a big cost-saver for them, even in the short term.
They tried introducing free software in schools in Iceland, where the Minister of Education is a left/Green who Gets It, and found that they can be stymied at the school-sysadmin level: where there was a free software-aware person in the school, they succeeded, where there was someone with a wall of Microsoft certifications they failed. Possibly some sort of negotiated discount on Linux certifications would help to break the "sysadmin's career is locked into Microsoft" deadlock. Private sector adoption of free software largely worked by samizdat: low-level geeks installed free programs because they were free-as-in-beer and thus required no purchase order, and by the time the CxOs heard anything, the company depended on it. In the public sector, this doesn't work: change has to come top-down, but requires bottom-up buy-in.