Announcements — Leave a comment
23
Mar 13

Government Service Design Manual

In an effort to talk a little more about what we’re up to at work it seems only right to mention the new Government Service Design Manual (and accompanying Digital by Default Service Standard). Their release is the next step in the Government Digital Strategy, providing a guide to what “digital services so good that people prefer to use them” look like and a framework to assess whether new services are ready to launch.

DbD kitemark

We’ve released the Manual as a public beta because we want to get as much input as we can from our colleagues within government and around the wider community that builds services for the modern world. Like many others at GDS, I’ve spent a lot of time over the past few months writing up what I’ve learned about building for the web and making things within government, and it’s great to see that content getting better thanks to all the feedback we’re getting.

So please let us know what you think. We accept Github Pull Requests or feedback through a few other routes. There are also more words about the whole thing in Richard’s piece about where this came from, Andrew’s piece on beta testing the standard, and Gareth’s piece on the ‘devops’ flavoured content.

Uncategorized — Comments Off
20
Jan 13

Micro Services and Storytelling

I’ve found myself watching a lot of InfoQ videos lately. They’ve been a good way for picking up some new ideas, but especially helpful for identifying other people kicking around similar concepts to those we’re using at GDS and beginning to establish more consistent language for it.

This morning’s choice was James Lewis’ Micro Services: Java, the Unix Way which is an overview of how his (thoughtworks) team built a financial services system out of a set of composed services, joined together with HTTP and AtomPub.

His description of his role as “team storyteller” resonated as a way to describe with how I spend a lot of my time these days: sharing the context for past decisions, discussing how we can improve on them, and trying to bring it all together around a coherent narrative.

The call for applications that are “Small with a single responsibility” also stood out as it so clearly also describes the way we’ve tried to build GOV.UK.

  • Each application only does one thing
  • Small enough to fit in your head
  • Small enough that you can throw them away (Rewrite over Maintain)

One of the things James mentions that that allows is putting each of those applications in a separate version control repository. That’s the approach we’ve (mostly) taken and it has been helpful for keeping deployment simple, but it has led to an incredible proliferation of repositories. In recent weeks I’ve begun to try grouping and describing those so that there’s something other than a massive list on a github page to help people navigate the code. It’s likely to be a never-ending task, but there’s a big storytelling responsibility to discharge there.

Announcements — Comments Off
12
Jan 13

Selling a laptop, mac minis and monitor

After making three house moves in the past year or so and clearing out a few remaining items from my pre-GDS office, I have all my gadgets and devices in one place for the first time in a while. It’s now very clear that I need to get rid of a few things.

First up is a 2009 model 15″ MacBook Pro. I bought it in June of that year and served me well till the 11″ air won me over. It’s got 4GB RAM, a 500GB hard drive and a 2.8GHz Core 2 Duo processor. It’s running OS X 10.6.8 but could be upgraded. It’s in great shape, but in the midst of all the moving I’ve lost track of the box and install discs. I’m looking for around £700-750 for it.

There’s an original series Mac Mini. That one’s got 1GB RAM, a 40GB hard drive and a 1.25GHz PowerPC G4 processor. And a modem. It’s not been used for a couple of years but is in good working order. The box is long gone and I’ve not found the install discs. I’m looking for £100 or so for it. It’s now listed on ebay.

The other Mac Mini’s more recent. It’s a 2GHz Core 2 Duo, with 2.5GB RAM and a 120GB hard drive. I’ve lost the install discs and box for this one too, but it’s running OS X 10.6.8. I’m looking for £200 for this one. (now taken)

Finally for now, there’s a Dell 2005FPW 20″ monitor. It’s served me well but is redundant now there’s a shiny apple display in the house. It runs 1680 x 1050 at 60Hz and takes VGA, DVI or S-Video input. Offers in the region of £65 please. (taken)

(We’re also selling a door, but that seems less relevant for this blog.)

I’d prefer to sell these to someone who can collect them, but I’d consider posting the computers within the UK. All prices are flexible but based on some ebay research these seem reasonable. Any enquiries? Email me at jys@ketlai.co.uk

Uncategorized — Leave a comment
27
Oct 12

That GOV.UK thing

A little over a week ago we launched GOV.UK. Except we didn’t, really. There was frantic work well into the evening of October 16th as people sought to ensure that we’d met all our legal obligations and that the right polish was in place. But we didn’t really launch GOV.UK…

What we launched was a cluster of nginx servers that redirect traffic for direct.gov.uk and businesslink.gov.uk (and a host of subdomains) to the correct places on the GOV.UK site that had been in place for months. Our launch was primarily about switching something off, not switching something on. And that was ideal. We made the smallest change we could possibly make, we tested it, and we told people.

It’s a tighter, faster and more expansive gov.uk than the one we released back in February. It’s been put through its paces over and over again, and we’ve cleaned it up (content, design, tech) as we’ve gone along. (We’re Actually Always Refactoring as a very good recent blog post has it).

The fact that we switched off two very large websites and replaced them with one simpler, clearer, faster alternative is a big deal. The fact that this site is now the canonical source for a whole host of information is a serious business. But in practice it was another small, iterative step of the sort we intend to keep making over and over.

Uncategorized — Leave a comment
26
Oct 12

No one was afraid

Now that GOV.UK is out the door I’ve been trying to catch up with my Instapaper backlog. One of the more recent pieces (yes, I’m way behind was Frank Cottrell Boyce’s wonderful description of working on the Olympic Opening Ceremony with Danny Boyle.

I particularly loved his description of the way the core team came together:

They worked so closely they were practically a hive mind. My job was to join up the ideas in a way that the non-hive dweller could understand.

Danny created a room where no one was afraid to speak, no one had to stick to their own specialism, no one was afraid of sounding stupid or talking out of turn. He restored us to the people we were before we made career choices – to when we were just wondering.

That sounds so much better than “cross-functional team.”

Notes — Leave a comment
10
Jul 12

Little Printer hack day

Chris' photo of MetaLoca

An early draft of MetaLoca by Chris

I don’t make it to many hack days. Domestic life with two kids and the tendency of my day job to spill into the weekends make it tricky. But I couldn’t pass over the chance to play with BERG’s Little Printer yesterday the weekend before last. Conveniently it’s also Kari‘s workplace, so she and the kids came along and were ably entertained by the BERG team (special thanks to Helen for keeping an excited three year old busy).

Mostly it was a day of enjoying the camaraderie and creativity of an assortment of friends. BERG do a very good job of bringing together a delightful bunch of people and the day had a nice pace. There was focussed activity but no crazy up-all-night race to the finish and plenty of time to share experiences and observations. Dan, Tom, Lanyrd, and BERG themselves have all written about the day too.

Chris, James and I had been chatting a bit before hand about exhibitions. For several years Chris has been running LocaLondon and I’ve been playing with some related ideas for a while. So building an app that would provide a weekly summary of what’s opening and closing in the next week seemed a good place to start. Chris had started a sinatra app and I focussed on tying that in to the BergCloud API while he and James began making test prints as PNGs. We just about got it all pulled together in time for the show and tell. We learned that the publication we wanted to make would use too much paper, that when you want to print recognisable, usable maps of bits of London it’s hard to beat the A-Z; and that Heroku’s perfect for these events, but getting seed data imported can be a bit of a pain.

The way we were working meant I had a bit of time to play with some other ideas. James had previously been working with another printer-of-small-footprint to make a little “what’s in season” publication that pulled in from the BBC Food site and Maureen‘s cookbook to provide suggested recipes. I took that code and amended it to work with BergCloud. It took all of about five minutes and worked pretty well.

I was also wondering about the use of something like Little Printer in more public spaces like large arts complexes. Places like the Barbican often have screens showing what’s on that day but they’ve never felt quite right. It’d be nice to be able to grab a strip of paper flagging the things going on that day, and especially those which aren’t sold out. To probe that a bit I scraped the Southbank Centre‘s website to grab all one-day events taking place that day and had them printed out using some of the styling for LocaLondon. It wasn’t quite right: lots of festival events made it hard to judge how everything fitted together, and while I could flag free events I had no way of showing what there were tickets left for. But it worked reasonably well given it was squeezed in around our main project.

After a day of working with it, I’m increasingly looking forward to owning a Little Printer. As a device to sit in the corner of a room spitting out reminders of things I might otherwise miss I’m sure it’ll be delightful. But I’m also increasingly interested in how it could be used in more public spaces (or at least shared ones, they’re increasingly privatised) to offer a little welcome and summary to those passing through. Wherever it goes, it’s great to be reminded that devices like this can have nice simple, basically webby APIs. And to spend time playing with friends.

Notes — Comments Off
08
Mar 12

Misleading infographics

During the opening keynote of QCon London yesterday Martin Fowler and Rebecca Parsons explored “The Data Panorama” and naturally that involved talking about visualisations as tools to explore large volumes of data. It was mostly very sensible but it reopened my unease about the way we all too often gloss over the fact that the very properties that can make visualisations so effective to communicate large lumps of data also make them a very effective means of misleading us.

So this tweet from Dominic Campbell and the article it links to felt timely, and worth flagging.

Commentary — Comments Off
05
Mar 12

Kestrel in Gauges

I really appreciated John Nunemaker’s recent post about the way they’re using Kestrel (a distributed message queue) in Gauges.

There’s nothing revolutionary in the way that Kestrel’s being used here, from the post it seems to be a fairly standard use case for it, but it’s a lovely example of detailing the day-to-day work of growing a web app that blogs are so good for. And it’s particularly good to read about the stages they went through as they tested each component in turn, in production, before switching over to them completely.

Notes — Comments Off
05
Mar 12

Inside Government

The second of the three GOV.UK beta releases was unveiled last week. “INSIDE GOVERNMENT” is the promised “corporate publishing platform” designed to bring together the core web publishing activity of all government departments in one place. Neil’s written very eloquently about it on the Government Digital Service blog, outlining some of their core challenges, and James Mead has added the developers’ perspective on the Free Range blog about their involvement in it.

It’s a strange release for me—it was the first of GDS’ launches that I didn’t press the button for, or even attend as I’m currently out on paternity leave—but I’m really delighted to see it out there for the world to feed back on. Since I first heard about the vision for the single domain I’ve been excited about the possibility that government information could be published in a way that allows it to be sliced along axes other than ‘department’ and “INSIDE GOVERNMENT” begins to give some life to that.

It’s also a really interesting counter to the citizen end of our emerging publishing platform. Where the citizen content is conceptually very flat with very few formal connections, the content in the corporate system is very much a series of nodes interconnected in numerous ways. Free Range’s work on the app, guided by Neil and co’s careful analysis, is especially fascinating for the way it lays out an encoded domain model for such a notoriously complex world.

This commit is a good example of the level of thought going into the process:

“First Secretary of State” isn’t really a role, so we can’t use its name directly to order.

Ideally we’d model which minister was also the “First Secretary”, but that introduces more problems. If we’d made “First Secretary of State” a ministerial role, William Hague would appear twice in the list of cabinet members. If we’d made it a non-ministerial role, it wouldn’t have been sortable in the set that MinisterialRole.cabinet returned.

This is a quick, pragmatic solution, but more thought probably needs to go into whether or not the lists of ministers are lists of roles, or should instead be lists of people (along with the roles they currently have).

My only involvement in the code to date is a fledgling branch that started to map out what our API for this side of gov.uk might look like. As it worked out there wasn’t time to do that justice and ship the website so we left the API to return to later (Paul Battley’s working on part of that now, and taking a slightly different tack). My time on it was enough to really appreciate the amount of insight captured in the code.

It’s also been a delight to have Free Range in leading the development of the ‘whitehall’ app that drives “INSIDE GOVERNMENT”. GDS is committed to building a world-class team of in-house developers, designers, and the like, but we’re not going to achieve the transformation we’re aiming for on our own. So we need talented outside teams, small companies, freelancers, and the like to help us. This release is our first evidence of just how well that can work.

Announcements — Comments Off
29
Feb 12

Backing up flickr

It’s taken me a little while to get to Aaron Straup Cope’s write up of his Personal Digital Archiving conference talk, but I’m rather glad I have.

The talk is an exploration of what we might do if flickr disappeared tomorrow; it’s a topic many of us have been pondering at least since the news broke of yahoo’s decision to “sunset” delicious. Two elements of the talk really grabbed my attention.

The first was the detailed exploration of how difficult it is to back up or transfer the “social” aspect of a social website. Copying my files and some related metadata is relatively easy. Preserving experiences and relationships is a lot harder:

“Privacy is genuinely important no matter what people are passing off as industry best practices. It is doubly important for anything that archives Flickr because a respect for privacy remains core to what the site is about and the ways that people use it.”

“This is the actual hard part of the personal archiving problem: How to deal with authentication and authorization controls defined by a third-party site that may or not exist anymore.”

“This problem is also why parallel-flickr is not the mythical archiving of all of Flickr. Because you can’t back up Flickr. Or rather: The only way to back up Flickr with any kind of credibility or ethics is to swallow the thing, whole.”

But in the middle of the piece there’s also a little something about a past event to backup flickr which involved building out complex metadata and wrapping them in lots of “standards”: “all the best practices around XML, the Semantic Web and static, linkable resources”.

It is incomprehensible gibberish.

Worse, it’s hard to do anything with. Not only are the data models overly-complex but all the stricter-than-strict, standards-compliant tools that grew up around them are hard to use.

This is a really important point, especially if we’re going to talk about personal archiving.

If I can’t stand to look at this stuff seven years later then what hope is there that someone who does not live and breathe the technical details will?

For those of us living in a world where there’s a lot of excitement around ideas of “linked data” and associated standards, experiences like this one are very important to hear.

← Older Entries