Labour’s “ICT” policy includes a statement on “Open software” in government, part of which is attempting to get two thirds of government agencies to use some sort of open-source software by 2015. This is basically what you expect from politicians when talking about “ICT” (or nearly anything, unfortunately) – they jump on whatever bandwagon/buzzwords are popular without any understanding of what should be done.
Firstly, I would be shocked if more than two thirds of government agencies were not already using some form of open-source software already. For a start, it’s nearly impossible to use the Internet without accessing something running Linux or Apache.
More importantly, it makes no sense at all to aim for governments to be using more open-source software. They should be using the software most suited to the job at hand, whether closed source or open source. What benefit is there in requiring open source? If they think it’s cheaper, then they should look more closely, because it’s often not. The same applies to being more secure. There are absolutely situations where the best choice is an open-source one – but there are absolutely situations where closed source is better. I don’t want government employees forced to use less than the best tools because of some ideological burden placed on them by someone wanting a cushy job for the next three years.
They have a few other requirements:
Software developed in-house will be made publicly available. A nice idea, but (a) I’m fairly sure that most of the in-house software is of no use to anyone else, (b) I suspect most of the software you’d assume was in-house was actually developed by non-government contractors, and, most significantly, (c) not all code is ready to be shared. If a government sysadmin writes a quick script to do a job, do we really want to add the pressure that it will be publicly available (and given that it’s the government, it’s reasonable to assume that someone will be looking at everything). As long as the software does the job, that’s good enough (in the “in-house” context). What would be worthwhile is ensuring that government agencies consider whether software should be released to the public – I’m sure that there is some that would be of general interest and where the quality is suitably high.
Agencies considering technology purchases over $2 million would first evaluate whether publicly available technology would substantially meet their requirements.
Ugh. A technology purchase over $100 should involve consideration of whether there was an existing public tool that could do the job – this should always be the case, not just for obscenely costly jobs.
Labour would also create a government “app store” to provide “a short circuit for fledgling NZ software developers to get to market” which would allow local developers to submit software for purchase by government agencies. Labour promises to ensure “informed neutrality” in software purchasing with due consideration of open-source software.
I really don’t understand what this is trying to achieve. There’s no indication that local developers would be preferred (only, with no reasoning, open-source software), so how does this help local developers? This sounds like a way to spend a lot of money for very little gain.
Governments should only be using software that produces and accepts files that are in formats that are open standards (this includes Microsoft Office documents). We do have a “New Zealand official interoperability framework definition” already, but it could be significantly expanded on (see what other governments require, for example). Everything the government produces should be in some sort of open format and everywhere the government accepts data at least one open format should be accepted. Open standards apply to other areas too, but open formats is the key in “ICT”.
We have a limited commitment to open data already (see data.govt.nz). This should be significantly expanded on – every data set that the government creates (and there are a huge number of these) should be publicly available, so that (a) we have transparency, and (b) other researchers can benefit from the data. The only limitation is privacy – unfortunately even when a data set appears to be anonymous it is often possible to identify individuals - it would be worth spending money on figuring out how to get past this so that we can share just about everything without breaching individual people’s privacy.