Tell Them What You Think

TTWYT’s development blog and related musings

I meant to write about this earlier, and the blogosphere has punished me for my tardiness — Simon beat me to it.

In any event: DIUS have launched an online consultation exercise on the Science and Innovation paper that they published in March. It looks great, and much like MyLifeMyId, it’ll be interesting to see what kind of results it produces. It is inspired by the Open Rights Group’s platform for collaborating on consultation responses, which features similar paragraph-level commenting. Apparently, this functionality is available as a plugin.

I’m particularly interested in the outcome of this experiment because I’ve wondered, since seeing ORG’s version, how well paragraph level commenting works as a discussion medium. It’s extremely granular, which is fine if certain paragraphs attract a lot of attention and thus spawn discussion, but doesn’t work so well if the feedback is more diffuse.

It is rather similar to an example of bad practice in online fora: a new site seeking to develop a community will often deploy forum software. A classic mistake is to make umpteen fora, one for each topic that could possibly want to be discussed, or, more commonly, for each topic the administrators would like to be discussed. A new user, upon encountering pages of mostly empty fora, rapidly gains the impression that there is no activity, and no point in staying around. A rather better strategy in this situation is to make very few fora, to funnel new users into a smaller number of places and give the impression of greater activity. New fora can be added later, as demand dictates.

I don’t know whether this problem will apply to these kinds of exercises. It will be interesting to see.

Some splendid people in the US have released a paper recommending that Government should focus on releasing its data in reusable ways, rather than designing better user experience — their thesis being that third parties will do this for them if the data is available. Funnily enough, I find myself agreeing.

The paper is a little US-centric but the principles are sound and entirely applicable to the UK. Ed Felton has posted an excerpt if you don’t fancy reading the whole thing.

The paper hits all the right buttons: let innovative private bodies come up with the best ways to display government data and compete for each other’s audiences. Make structured data available first, and then produce a government site to display it, if required — the data should be a priority, not an afterthought. Make sure that such data feeds exist in known, permanent locations. Make government sites operate on the same data they provide to others — TellThemWhatYouThink is powered by its own API. We eat our own dogfood. I wonder how many government departments can say the same^*?

I made a lot of these points when I spoke at Tower08 earlier in the year. It’s definitely true that there are people in Government who agree, but there’re still plenty that don’t — hence the unfortunate need for campaigns like Free Our Bills.

Hopefully, reports like these will continue to be written, good examples of the reuse of public data will continue to be found, and Government will eventually see the light.

¹ Ok, that’s not quite true. I know the answer, as does all of sentient life.

I had a totally awesome time at the Barcamp over the weekend.

Matthew Somerville did two great talks — one on the new video feeds on TheyWorkForYou.com, and another on Bach Chorales. Stuart Langridge did a splendid talk on HTTP codes,  which led to quite a bit of pleasant ferreting around in TellThemWhatYouThink’s logs and code to see if things could be improved (they can). The various Yahoo people demonstrated how to do their code reviews, and allowed us to “ask them anything”, which was excellent fun. I did a talk on TellThemWhatYouThink, predictably enough, and some great ideas came out of that — thanks to all who came. I had long conversations with Rob McKinnon, of TheyWorkForYou.co.nz, who had some great ideas about how to involve people more with the site and each other.

In short, a splendid time was had all round — much thanking and hat-tipping due to Ross Brugies, who sorted the whole thing out — food, beer, swag, ideas and new people. Who could want more? I can’t wait for the next one.

I had the pleasure of speaking to the assembled great and good at the Tower08 Transformational Government conference on Monday this week. I hope that video will be available at some point, and I’ll link to it if it is.

I talked, reasonably predictably, about the resusability of public data, and about why it’s important to embrace the idea that data should be made available in ways that allow people to use it, reuse it, combine it in new and clever ways and produce new, useful tools.

I also pointed out that there is an incredible amount of value to be generated from this data if it can be published in ways that allow more collaboration, and that it’ll be much cheaper in the long run if Government doesn’t try to solve all the problems. I drew a comparison between DirectGov’s fairly awful search facilities and the results produced by DirectionlessGov, which drew both heckles and laughs — an odd response. I am rather surprised to find that there actually are people out there who think that DirectGov’s search is better than Google’s. It’s a strange world we live in!

Being fairly new to the scene, I was most struck by the huge differences in people’s interpretations of what transformational government should be about. In fairness, this shouldn’t have been that surprising: everyone is interpreting it according to their vested interests, which is predictable enough.

At one end, there are people saying that everyone should own their own data, that public data is public property and should be disseminated in ways that make it as useful as possible, that massive data sharing and joined-up delivery of public services through one site is a dangerous folly.

At the other, you have people saying that we need to make identity card systems to share everyone’s data throughout government, that we should make public services usable online by having ultra-secure identification methods, that we need one place to find everything anyone might want from government, and that web 2.0, sharing and mass collaboration are merely the whimsical trends du jour.

I think it’s probably easy to tell where I stand! I’m happy to say that there is a cadre of people in government who also tend towards the former view, and that it is larger than one might think. These ideas are gaining some traction, at least, and that is quite something.

The recent redesign of the DoH website is, as has been noted by others, a great improvement. I was happy to discover that they do now have RSS feeds for consultations, and, more impressively, for consultation responses. This is something people have wanted for years, so it’s great to see that it has finally happened.

Unfortunately, they don’t work. The current live consultations feed, erroneously, contains no items.

I shall let them know…

A few months ago, I responded to a couple of government consultations and, in the process, discovered there was no way to search all live consultations, or to be alerted when a new one was published.

This struck me as more than a little mad: consultations are one of the government’s main methods for engaging in substantive public debate. I am not aware of any other process by which Joe Public can express their views on a particular topic, in detail, and can be reasonably certain that they will be read, and justifiably hopeful that they will be taken into account. That said, the entire process is worse than useless if no one finds out about them in time to contribute.

I thought that there must be a solution to this problem. My first port of call was a quick survey of government websites to determine which ones provided their data as RSS feeds. To my utter astonishment, only the Welsh Assembly Government did. With this came the realisation that the problem was a bigger one than a mere aggregator could solve (call me naive, but I had hope!), so I set about writing a screen-scraper for the DCMS as an experiment. I subsequently discovered that I had unwittingly picked one of the worst departments to scrape: their code is particularly unstructured. After implementing scrapers for ten or so departments, I had a look at the horrendous spaggetified ascii pudding that my code had become, and started again.

Over a couple of weeks, my little collection of scrapers grew until I had twenty or so departments. At that point, the data was only available as RSS feeds, and I had no plans to do anything more with it. After showing it to a few people, however, its potential became clear, so I started on a proper website to do all the things that people would want. In particular, Tom Steinberg’s reaction was effusive, and his enthusiasm infectious. I was also much encouraged by the numerous splendid people I met at the inaugural UKGovWeb Barcamp, who all said very useful things. Kudos in particular to MySociety for offering their hosting services.

In the weeks following Barcamp, I continued to beaver away at the scrapers and the site, until it went live on 11th Febuary, 2008. I have been very pleasantly surprised, to say the least, by the level of interest TellThemWhatYouThink has received. I have already had a few meetings with civil servants to discuss changes to their sites which would improve the quality of the data TTWYT’s scrapers can gather, and several more are planned. Lots of people have emailed me great feedback on improving the site, and I and others have lots of ideas about where it could go in the future. To all involved: many thanks.

Watch this space!