More on browsers

Meanwhile, by the by, I installed SlimBrowser this morning – tabbed browsing inside IE (you get to keep your bookmarks and not fiddle around importing them back and forth, you get a neat search toolbar and it blocks popups too). Can’t find any quirks so far. I’m running it on my PC of course – I played with 5 different browsers on the Mac and none worked on all sites all of the time – talk about browser inconsistency. In the end, I settled on Firebird which seemed to do best (but not great).

Browser Stuff

Again prompted by David Fletcher, I’ve been looking this morning at a posting on the Apple website that focuses on web design and mostly about CSS. One of the biggest frustrations in developing online apps for government has been the need to support a vast range of browsers. Many commercial entities can just say “works best in IE 6 or Netscape 6.x” and go from there, living with the inconsistencies between browsers … as Apple says

Although the standards are not yet fully supported by all browsers in all circumstances, creating standards-compatible pages is the best way to ensure good rendering. As always, learning to use new technologies will take some time and will give you some incompatibility headaches. Nonetheless the results will be well worth the investment.

The public sector though faces the need to support many browser versions across many operating systems, despite relatively low usage for everything other than IE. For instance, looking at December traffic on one key government website:

Over 96% of traffic was from Windows PCs, but: 41% was XP, 20% was Win 98, 20% was Win 2000, 7% was Win ME, 6% was Win NT, 2% was Win 95 and .2% was Win CE (that won’t add up to 100%, there were some others in there, including something called Windows 32 and I have no idea what that is).

The total count of different browsers that accessed the site breaks down like this:

Linux: 3 browsers, 0.01% to 0.13% of total
Mac: 3 browsers, 0.02% to 0.71%
Motorola: 1, 0.01% (someone using wap to access the site!)
Windows: 5 browsers, 0.16% to 89.9%

And then there’s Nokia, Siemens, OS2 (does that still exist?), SonyEricsson, Unix and a small amount of unknowns.

And, in version terms, a small number of IE 3 users, quite a few IE 4, 26% IE 5 and 65% IE 6; likewise, Netscape 3,4,5 and 6 are all in use; and 3 versions of the AOL browser. In the “other” category – what I usually call my “n squared” problem (a big, ugly, complicated matrix) are things like: blackberry, knowqueror, lotus notes, icab (?), opera, lynx and even webtv.

The site has to work well for all of those browsers … text has to position correctly, fonts must render exactly, accessibility features must work and so on. That’s a tough challenge – how many commercial sites do you think work well for such a variety? The ROI falls off very, very steeply once you go past IE 5 and 6 / Netscape 5 and 6 … but if it doesn’t work [well] on one browser and that impacts someone’s ability to access government information, I don’t think we’d get very far if we suggested that they upgrade to a more current version. The browser inconsistencies are also one of the reasons that digital certificates didn’t work well.

The Apple article talks a lot about CSS and how hard it can be to get right, but how worthwhile it is. I agree with that – we use it on our platform and by doing it once (at the centre of government), anyone that uses our platform gets to take advantage of all of that work that we’ve already done. So they can forget about the technology and the testing and compatibility issues and get on with publishing high quality, focused content. Much smarter and, to use a current word, efficient.


Prompted by a post on David Fletcher’s blog, I tried out Kartoo just now. Typing “government” gives an interesting result that I have a picture of, but can’t upload to blogger for some reason. I’ll come back later and try again (ah fixed it, new browser blocks popups, but you can recover them and put them on a whitelist). So here’s the screenshot:

It’s an intriguing mini-map of government sites and not at all what I’d expect. Vertex get an entry, apparently because they’re a sponsor – at least they don’t appear at the centre. Both No10’s site and ukonline appear, but so does an Irish government site and MSN. And then two sites that are both retired – (which was last seen in mid-2001 I think) and which I think went around the same time (following the election?). That’s the harsh stuff out of the way.

What I really like is the categories on the left hand side. I can’t always figure out why some sites come up in the picture, but it’s interesting to see the variety none the less. I’m sure I could waste hours here discovering new sites, both from the UK and outside (even though I restricted to the UK, for instance, there’s a site called Clicking on the site can take you to the home page, not necessarily to anywhere that your search phrase occurs, but hovering on the site shows you the link text on the left hand side. You can also increase or reduce the number of sites that come back in the map (on the options button), or change the search engines used to aggregate the results.

What Kartoo really tries to get to though is the relationship between sites – it’s kind of a technorati for the mainstream web. Who is linking to who, and what are the associations that they create. Fascinating, and as David says, a sign that the be all and end all of search is not google (although I’ve been using all the web and teoma a lot recently). A serious bit of work has gone into this, that’s clear.

While you’re looking at Kartoo – type your own name in and see what comes back.

Open source? Or not? Or Not in Newham anyway

Intriguing piece in Kable that says Newham have pulled out of a PoC to evaluate open source on the desktop, because the scale of the project was too big for them to handle (and that appears to be a direct quote from the IT head there). Making a move to a new desktop across the organisation would be a big deal and involve a lot of effort from the team so I can understand the reluctance, but the absence of data on how such a move could be achieved is hurting the public sector generally I think. There are still another 8 PoCs in the OGC list, so there is still time to get data. One of the things I’ve wanted to do for a while is run three desktops in parallel – SunRay (thin client), Open Source and Windows – perhaps on 3 floors in one building or in three teams handling the same functions in three departments and see how it worked out. Keeping the numbers down to maybe 20-30 people would reduce the support overhead and probably allow a relatively quick implementation. I think it would be fascinating to compare not just the cost dynamics but also how people felt about each platform and how they used them differently (for instance, would productivity be up? Would people spend more time surfing on a browser-based platform like a thin client, or less time?)

ID Cards that verify, not identify

Interesting article in this month’s Wired magazine on ID cards and a caveat:

An ID card might do little to prevent another terrorist attack. Bruce Schneier, one of the world’s leading computer security experts, says that any identification system that relies on background checks creates three categories of people: trusted, untrusted, and trusted-but-malicious – people who aren’t on watch lists but turn out to be terrorists anyway.

The Year Ahead

I started working in this space – e-government – in early 2000. Nearly four years on I am, in many ways, disappointed with how far we’ve come. It’s not that we haven’t done a lot collectively. It’s just that the expectations were so high then and it’s hard to look back and know about all of the hard work that has gone at so many different levels of government and see the outcome achieved without wondering where the rest of it all went.

When I started I figured that three years on the idea of e-government would be absurd. Business would revert to the norm and it would just be business. Whenever any new service was launched there would be an orchestrated launch across all channels with a significant reduction by now in the number of services with any kind of paper channel. Patently we’re some distance from achieving that in the UK and probably in many other countries too. The web is still something separate. Paper is still paper and everything done has to be done there too.

Well, I think that finally there’s some change there. New services were launched this year that mean that paper isn’t the only way in and, even, where there is no paper equivalent. How would you deal with real time updates via SMS about where you’re supposed to be when you’re a witness in a court case in a paper world? How would you get total access to the 1901 Census data without wandering down to Kew and poring over endless amounts of microfiche? How would you see emerging trends in your neighbourhood as offered by Neighbourhood Statistics in a world of paper correspondence? The power is being harnesses and the pace at which it’s being done is quickening.

That makes 2004 a pivotal year to me. In July, I talked about a 50 year storm emerging …

“the 50 year storm that looms is the set of events that will take place this year and early next that will warrant the catalyst. Some senior figures are moving on (perm secs in at least three departments), one or two cross-government figures too. Issues like we have seen over tax credits where technology and business issues conspired to cause enormous pain mean that we will have to rethink delivery controls. Spending will tighten as we enter another financial review round. The potential for central infrastructure, like our own Government Gateway, will be fully realised and people will commit resource to exploiting it rather than exploiting ways to get out of it”

Well, for sure the spending is tightening – that’s become pretty much daily news for FT readers.

The figureheads are indeed moving on – witness the recent announcement of a transition from Office of the e-Envoy to Office of e-government and the appointment of someone analogous to a CIO.

The potential for central infrastructure is, at least, being talked about as are ways to further exploit it and other products from a variety of departments. Once the debate moves on from why it should be used to how it should be used, then joining up can really start – because it will be happening from a consistent base. That’s not to make light of all of the other work that will have to be done to join up – there’s an enormous amount around business process, data administration, regulation and co-operation too.

Given today is the eve of another year, I’ve been thinking about what emerging trends we’ll see, in e-government, over the next 12 months and how they might impact what people on the outside looking in see. The storm is, I think, upon us and the next few months will make clear what comes next.

There are 5 topics on my mind (there are a few more within my day to day job, but I’m thinking here about emerging trends only).

1. Citizen at the centre

Since the beginning, e-government has been about putting the citizen at the focal point for service delivery. I don’t for a second think that people go out and deliver things without that maxim in sharp focus, but they have been constrained by a variety of things – natural inertia, lack of system capability, old style business process and so on. This year, things change and the citizen will be squarely in the middle. That will mean:

– Dramatically fewer websites (maybe not fewer in number, but certainly fewer that have to be visited to get the task done)

– More focused content that is written in people speak, not government speak

– More transactions grouped together in logical ways (so Child Benefit and Child Tax Credit will next to each other, ditto Council Tax Benefit and Housing Benefit)

– Central and local services will start to be aggregated – with the local government people probably leading the way (many already see themselves as a kind of one stop shop for government), but probably not until close to year end 2004

– Consistent navigation and controls so that there is no relearning necessary. People know how traffic lights work – and we don’t change that from town to town, so why on earth would we keep changing look and feel?

– Consistent multi-channel delivery will show up, again near the end of the year, where trying to get at something via the ‘phone will feel similar to the web. The folks at the call centre will be using the same interface as you would use if you were online, so consistency will increase … and the website will be updated if your question isn’t fully covered by what is already online.

In many ways, government is some way behind the private sector in doing this (just check, and for instance), but is also grappling with the fact that, historically, every section of government is a separate organisation, indivisibility of the crown notwithstanding. So the second trend is a sort of coming to Jesus.

2. The Rationalisation of Brands

This won’t fully happen in 2004 but during the year we’ll see, I think, a shift to service delivery from a relatively generic thing. It won’t be about the department of this or the department of that, it will be about “you’re the citizen, what do you need”. Departments will still exist – maybe a couple fewer – but there’ll be aligning themselves much more closely, to back up the trend identified in (1) above.

Once you start rationalising the brand, all kinds of interesting things probably happen. Who owns the end to end service? Where do you go when it goes wrong? How do you track something that is being dealt with through multiple back office units, all with different processes? These issues and others like them will force more co-location of resource, more interworking of systems and a greater ability to join up in the future (after all, once the wave hits, why get in the way of the next one?).

This will mean things like syndication going mainstream – to the point that any content will be available anywhere and so you won’t necessarily even know the source brand. Once you start doing this – you can get definitive information about, say, child care from three government entities, Mothercare and Boots – you probably spend less time looking and more time using the information, and you probably get it from brands that you trust on an every day basis rather than having to think about where the definitive source would be. If you’re shopping on (Ocado) and you happen to buy nappies for the first time, it’s probably logical that they pop up some screens about what government services you’re eligible for as a recent parent. Maybe you even save money on your shopping because they can process the claim in real time, using services and data sourced from government but presented in a non-government branded way.

3. A Shift From Silo To Enterprise

Allied to (1) and (2) then is a change in the way that things are designed and constructed – to remove the issues about end to end ownership, support and delivery. As long as things are built inside fortresses, bashing down the walls and linking a couple will be hard (impossible?) and relatively pointless. I’m guessing our would-be CIO will focus first on how we shift from the silo to the enterprise or, at least, from talking about the enterprise to doing it.

Central infrastructure is part of this (but I would say that, wouldn’t I), but it’s not all of it. A base of solid standards for, say, web services security and interoperability would be fundamental – after all, usage of MMS (whilst still low) was pretty non-existent until you could guarantee that sending from Vodafone to O2 would work. A decision on, say, what a trusted government mobile phone number for SMS messages would be is important. Then some components are built a few times, in a sufficiently generic way, and deployed many times (adhering to the standards identified) and plugged (through a set of bespoke adapters) into the multitude of systems in place.

Once we’re through that battle, maybe we can attack the multitude of systems and see what can be rationalised and componentised there – reducing the amount of infrastructure and the complexity of delivery. Ideally, this speeds up introduction of new services and reduces the risk of failure – if it doesn’t do that, we shouldn’t do it.

4. Business Leadership to the Fore

For (3) to happen though, technology needs to be seen as the servant to the business and the business owners have to take ownership of the widest possible agenda. I’m guessing that, today, most governments commissioning a new service build a new IT system to support it. That might have made sense before (it certainly made things simpler through reducing interdependence in projects and delivery), but it doesn’t make any sense now. If (1) – (3) are to happen then, increasingly, business processes will be designed from the citizen into government rather than government outwards.

That will mean a greater degree of cross-business alignment and rationalisation. Perhaps certain departments will own key processes for all of government, perhaps a department will not only own how it works but also actually run it for everyone else (one way for money to get in to government, say).

This is the hardest thing to see how it works. It requires fundamental changes at a base organisational level in any government that undertakes it. That will mean new structures, new incentives, new controls and disciplines. I didn’t say it would be easy, but the emergence of a trend like this will show a true appetite in a government for tackling the very hardest problems. For a long time (since almost the first week I got involved in e-government), I’ve said that the web allows us to put a veneer over the complexity of government, hiding it from our citizens and buying time to allow us to engineer the really complicated changes beneath. That’s still true and we’ve all bought some time – but it’s time to start tackling the hard stuff for real now.

5. Success stories will be common, and will become a non-event

By the end of 2004 a handful of services will be mainstream, i.e. they will have significant usage when compared with, say, buying books online or banking online. Perhaps 40-50% of people will use an online channel for just a few services, finding that it’s quicker, easier and a richer experience than trying to use the ‘phone.

Incidentally, as more private sector businesses outsource their call centres abroad, I wonder whether web usage of things like banking will increase – if you can do self service, why make the call? If that’s right, then government benefits too – the more people who are online and who are comfortable transacting online, the more people will feel comfortable using the services available from government.

What are these services? Some are already there – the congestion charge claims 70% of payments are made online or through SMS. I think they got there by making the offline (in this case telephone) process so ridiculously painful that pretty much everyone found the path of least resistance (the absolute path of least resistance is, of course, not to drive into London and many people it seems chose that one). The online driving test service is already doing well, so the papers say and is a natural start point (kind of a “my first government transaction online”) for a generation that already expects the Internet to do pretty much everything that they need. This will be closely followed, I imagine, by Student Loans

(I notice that this site is one of the first to put the legal terms and conditions for use in as the default page when you first visit – which I guess is compliance with the recent EU law on opt-in, although I hadn’t expected it to be that way. It’s also “strictly copyright 2002” which is a bit of a shame. Is it up to date or not? And then there’s this line “You agree to indemnify and to keep SLC indemnified for and against any costs, claims, demands, expenses and liabilities suffered by SLC arising from or which is directly or indirectly related to your access to and/or use of the website” which would worry me if I was about to use a website.)

Today, the site offers only pilot access to student loan applications (and my postcode isn’t in the pilot area so I couldn’t test how easy the process is).

Once you’re booking your driving test, applying for a university place (which isn’t strictly government I think) and getting a loan then your interactions with government as a student may be few for a while (unless you are applying for benefits, such as reduced price spectacles)

So, fast forwarding a few years on in your life cycle, the next place where usage ought to be significant this year is in the area of benefits and tax credits. The latter already had a banner year in more ways than one and with renewals due in April, online has surely to be the way to go.

On top of those, I think there will likely be another sleeper hit or two. Something that will catch most of us by surprise, like the 1901 Census or the Flood Warning site – perhaps something like Diana’s Inquest, or the publication of the Hutton report or maybe news that London will host the 2012 Olympics (followed by publication of the plans for development to meet that need). Events like these will drive traffic to the web from both existing and new users who can then move on to other government services.

Beyond that, I think the services that really catch on will be invisible government services – the things that government does that aren’t really associated with government. The 1901 Census was one such thing – I wonder how many people really connected that with government? Booking a squash court at your local leisure centre might be another – after all, many such centres are run by local authorities.

But what really interests me are the spontaneous things that we might get people to sign up for as a lead in to other things. Let’s say that when you send in your (paper) tax return, we take your mobile phone number – and text you when the processing is complete and ask you to visit a website to confirm payment details for the refund; or perhaps we arrange to text you when the cheque for your child support money is in your bank; or we make a deal where we’ll email you when we have something to send you, rather than adding to your mail pile at home and you can visit a secure area to check what it is and decide whether you want a hard copy (printed right where you are or sent via snail mail, but your choice). These services won’t be obvious “tick in the box” services – i.e. they don’t exist offline and so when they go online it’s hard to know how to count them (after all, 100% online makes sense at a certain point in time, but at some point, the baseline has moved and you might be putting 50% online of what you had then and another 25% of services that didn’t exist before).

2004 is going to be lots of fun, one way or another. There’s a huge amount to do and, from tomorrow, only 24 months before the deadline in the UK to have 100% online and high usage of the key services. That means 12 months from now we better be at least ½ of the way between where we are now and 100%/high usage!