The ‘Github for Government’ was interesting if not really what I was expecting from the description. Part of the problem is that I think that while there are clearly some parallels between open source and open data there were less than Ben from Github seemed to be suggesting and thus the premise was a bit flawed from the outset.
I have to be honest and say that the more interesting discussion for me would have been around the topic that Jon Foster, from Futuregov, raised about using Github to publicly work on policies and reports. Somehow this became about legislation in the session (partly my fault for raising legislation.gov.uk) but that is clearly more complex.
We touched on Prose.io and that is something I’d like to explore more with its ability to essentially mask some of the complexity of Git(hub) behind a pretty simple writing interface. There seems to be something there that could lead to a new kind of consultation maybe?
The differences between UK/Europe and US city/federal approaches to open data was obviously clear. There seemed to be a feeling that over here there is more top down pressure to do these things whereas in the US it is being led from the grassroots more? Maybe that is true but certainly it seems there is more grey areas to it than that as while we might not have a ‘Code for America’ we do have groups like the OKFN and ODI and a pretty passionate grassroots open data movement. Also I’m pretty sure there have been some pretty significant Presidential decrees about this stuff?
It was also interesting that of a crowd of 20+ only two of us were actually ‘government’ employees (the other being Steven Abbot from Parliament) and so it was an interesting insight to what people believe the attitudes and blockers within the civil service(s) are. Certainly it seemed a surprise to most people that Github is a barred site for the vast majority of us!
The talk of Github uptime currently being better than the US Federal Government was fun but also made a serious point – is it right and sensible for government departments and agencies to rely on a commercial, independent, third-party to provide these services?
The real lightbulb moment for me came right at the end. The discussion about the work Max Ogden is doing on the .dat standard was *really* interesting. An acknowledgment that .git is actually not great for dealing with large datasets and the needs of the open data community has led to a project to build;
“..an open source tool that enables the sharing of large datasets, allowing for a decentralized collaboration flow similar to what git offers for source code.”
Now it seems to me that a tool like that combined with a tool like Github for the community elements could be a game changer. One to watch for sure.
So learning more about that project was one of my actions from this session. The other was to learn how to properly use .git and Github. I really should have taken up the other from Ben for the Github 101 run through.
The Github for Government site has real potential to be a great resource for everyone involved in Government digital & open data so that is something to keep an eye on (and to add more UK/Euro examples..)
Digital by Default 
3 responses to “Github for Government at #mozfest”
I think the use of Git+Github for data has some potential and as Max Ogden points out in his writing about ‘dat’ (which looks very interesting – thanks for that link) for small datasets in simple formats (e.g. csv) – mainly because of the ‘ease’ with which data can be posted online and copied – the problem being that ‘ease’ only exists for those familiar with Git+Github. (see the Cooper-Hewitt work in this area https://github.com/cooperhewitt/collection)
While the Git+Github issue is a hurdle with a decent GUI client (I use GitBox on Mac) it is actually quite easy to copy existing repos and publish new ones – I’d say on a par with installing an FTP client (although no doubt there is more choice for the latter). I’d definitely recommend looking at this route if all you want to do is publish/retrieve repos from Github, and will also do for basic level of version control.
I have to admit that I’m not convinced that using Github to work publicly on policies/reports is such a useful idea – or at least I’d say ‘why use github’? For published legislation it feels a bit more useful as the ability to track change at a very granular level and to see ‘explanatory’ messages about the change (i.e. Git commit messages). However for public development of documents I can’t help but feel a more ‘Docs’ type of environment might do the job?
I think I said the policy stuff would be interesting – not that it would work 🙂 a few places are doing it though (city of San Fran was an example given..)
I’m not that interested in small, simple datasets as this is more of a day job thing for me and our datasets are anything but..
Matt, I went to the Github Gov meetup in London last night. A somewhat unfocused session, but in terms of Ben’s presentation, I had similar thoughts to you. I think it is the philosophy of open source that we need to see transfer to the open data and ope government space. This will then drive the creation of the right type of platforms, based on Github or whatever.
I feel that we are some way yet from creating the eco-system that, for example, I imagined for UK Location back in 2010 – an open, virtuous circle, dialog between data creator/publisher and data re-users. I’ll write more on my blog in the next few days.
And yes, there is no point in creating a platform that can handle simple CSV files, but not larger datasets, e.g. spatial.