Sunday, August 30, 2009

The Gateway 9 Years On

Looking for something entirely unrelated, I came across these two slides from July 2000. When I used these, they were printed on transparencies and displayed using an Overheard Projector. From there to pico-projectors!



Looking at that second slide, not one of the departments named still exists:

IR and C&E became HMRC, DSS became DWP, DTI iterated a few times and is now DBIS, MAFF is now defra (and DECC).

Friday, August 21, 2009

Counting .Gov Pages

I've been fascinated by the government website page count for the last couple of weeks. The numbers just seem so extraordinary.

On August 5th 2009 I wrote that there were 112,000,000.

Today, there are 114,000,000. That would appear to mean 2,000,000 were freshly created in the last 15 days or so. I'm suspicious because the numbers are so round. It's not 112,123,456 and 114,654,321 ... but maybe Google rounds up at numbers that large rather than count exactly.

In a later post I showed the timeline that, at first glance, showed pages dated from the 1450s - but Google's timeline function seems to look for instances of a date in a document. Dan and I have debated why that might be and whilst he leans towards the view that it has some use, I'm a bit more sceptical.

But what you can do is look at pages modified in the last 24 hours, last week and last year (which I think it does by looking for dates within those ranges in the document, rather than looking at the upload date). For government this is:

Last 24 hours: 35,300

Last Week: 1,540,000

Last month: 2,080,000

Last year: 5,390,000

Again, those numbers don't seem to work ... 7 * 35,000 isn't 1.5m ... 52 * 1.5m isn't 5m. Checking every month through 2009, there's a reasonably consistent number of between 1.9m and 2.1m pages.

It's likely that edit frequency varies but it's been a busy week for UK government if 1.5m pages have been updated - if it cost just 1p to update a page, that would be £15,000 of effort ... if it costs £5, that would be £7.5m! But at the annual end, fi it's 5.3m pages at 1p then we're at £530k ... that's about 10 people at fully loaded costs. If it were £5 ... then we have armies of people updating pages.

In July 2001, google says only 20,900 pages were updated. In July 2009 it was 2,120,000. That's over 100 times as many pages.

I feel a table of analysis coming on to see if there's any sense in this. In the meantime, any clues what is going on? Is Google doing something that I haven't accounted for?


In May 2002 I spoke at a conference about the prospect of government making more use of text messaging - I was specifically thinking about exam results at the time (although there was a separate thread about "your benefit cheque has been deposited"). The idea was written up by Computing and then picked up by some daily newspapers (a slow news day I imagine). The Independent, for instance, said:

Children could receive their exam results in the form of text messages sent to their mobile phones next year, as part of government plans to expand "online" services. Similar systems are being considered for benefits claimants, to tell them when their money is being paid, and to inform people of the progress of their passport applications.

The plans were revealed by Alan Mather, of the Office of the e-Envoy, who said that the necessary security could be in place within a year. He also added that banks could join in, using software developed and tested by the Government.

"There are organisations that want to issue notification of important things through SMS text messages, such as exam results or benefits information," he said. "This can be done in a 12-month timeframe. And if we get this up and running there's no reason why the banks couldn't do the same – we could extend the model to any commercial provider."

However, a stumbling block could be that the proposed services would not work on "pay-as-you-go" phones, because they can be bought by anyone without further checks. At the moment some government services can be accessed via PCs with a password sent through the post.

Looking at it now, 7 years on, I have no idea what I was talking about with some of it. I remember being concerned that given 70% of phones (at the time, probably the same now I suspect) were pay as you go and so had no address information, we wouldn't be able to use it as an authentication token or as something to send semi-secure information to. I can't remember what I was talking about with the banking reference although at the time we were wondering if we could get banks to do authentication checks for us and then pass the confirmed check to us so maybe it was that.

There were some odd reactions from some people, such as:

Exam company, Edexcel thinks it would be difficult to co-ordinate all the phones numbers for 4 million students who take exams each year.


John Lettice noted: We do however doubt the examining bodies' and/or schools' willingness and capability to collate and distribute results in secure SMS form. Or indeed as email, or posted on a secure web site. Reality check: just yesterday The Register supplied sprog one's school with a single first class stamp, GCSE result delivery for the use of. Under the circumstances we do not expect them to be asking us to stump up for our share of an SMS server in the foreseeable future.

When someone at work mentioned yesterday that her daughter had checked her results online and had to login to a website to do it, I wondered how far people had got with exam results by text. I found only a couple of references - and, as far as the UK goes, it's Scotland leading the way.

From the BBC:

Nearly 160,000 school pupils across Scotland have received their examination results. Almost 30,000 students received their grades by e-mail or text message.

From Australia:

Interactive communications services supplier Legion Interactive has been chosen by the New South Wales Board of Studies to deliver exam results to students via text message. Up to 63,000 students will now receive the results of their HSC exams [by text]

From Tanzania:

Tanzania`s university students will from May this year access their examination results through mobile phone text messages (SMS), according to a local Information Technology (IT) expert. The executive director of Easy Life Group (T) Limited, Benjamin Sitta, said yesterday in Dar es Salaam that the use of mobile phone has widely expanded and starting this year, students will access their exam results through their mobile phones

What, I wonder, is stopping more exam authorities doing it? Back in 2002 I was told that youngsters receiving exam results might need others around them to counsel them ... so how does sending them in the post address that?

HMRC has been texting me for perhaps 5 years to tell me that my Self Assessment form has been received, the refund is on the way or that I need to pay them more money.

John Lettice was certainly more right than I was ... I just can't see why it isn't more common.

Thursday, August 20, 2009

Charles Cox - An Update

A kind friend forwarded this news today, which I'd missed ... from the LBC website:

Manslaughter In Soho

Police have charged a man with manslaughter after an incident in Soho.

Charles Cox was attacked in Wardour Street in November 2007 - he died at the beginning of this month.

35 year old Jeremy Mark Aylmer is due to appear before magistrates later.

According to the FSA website, he's a trader presumably in the oil business:

Wednesday, August 19, 2009

It's older than you think

Back when Gutenberg was inventing his first printing press and the Renaissance was just beginning, google tells me government was already beginning to post pages on the web ... with volume steadily increasing all the way into the Industrial Revolution before peaking recently during the Information Revolution.


Wednesday, August 05, 2009

While We're Talking Rationalisation

Here's a site that seems to be no longer relevant


In case you were wondering ... this is to be found at

Delete this page and right away we'll be at 112,999,999 pages ... it's going to be slow!

SportTracks, Rubitrack, Ascent - All For Mac?

If you're a regular runner you've probably already bought a GPS tracking device - perhaps a Forerunner (maybe a 305, 405 or the new 310XT) or some other device. You've probably also concluded that whatever software came with it is no good and looked for a replacement.

If you're using a PC then the only software that I came across that was worth having was SportTracks by ZoneFiveSoftware. It's technically free - but they ask, and I'd urge you to make, a donation. SportTracks is really very, very good. But it's PC only and whilst it does work inside software like Fusion or Parallels, it can be difficult to get working and when I was using it I ended up keeping a PC on my desk, next to my Mac.

There are rumours about a version of SportTracks for the Mac, but they seem to still be just that, rumours.

For Mac users, I'd always thought that the choice was more limited but there are at least two great options: Rubitrack and Ascent. Both of these are nearly perfect options to use instead of SportTracks not just whilst we're waiting for SportTracks to come out on the Mac, but for good.

I've written before about using Ascent so won't repeat any of that here. Rubitrack is very similar and it has, just, the edge for me. Trying to boil it down to the biggest differences, Ascent is slightly more technical in its capability - it has greater graphing and reviewing options - but Rubitrack, for me at least, has better presentation and sifting and sorting. Both of them auto-label your runs based on prior runs, but Rubitrack makes equivalent runs easier to find - allowing you to sort by location, by timetable and by distance with selections from the left hand menu bar. I also like the way Rubitrack can present the runs as icons - like the picture below.


Neither Ascent nor Rubitrack include one feature that I used a lot with SportTracks - the ability to tag runs with a piece of equipment and then keep a count of how far you'd run with that gear - for instance, you could note when you started using a particular pair of shoes and it would keep track of the mileage you'd got to, allowing you to figure out when to replace them (or at least to get ready to - given it's all about how far and hard you run I suppose).

Two very good choices then, with Rubitrack having the edge for me. The good news is that both are available as free trials allowing you to put a limited number of runs in without making a big commitment. Try them out and then pick the one that suits. And when SportTracks comes along for Mac, if it ever does, you may stay just where you are ...

Spezify on e-Government

This both impressed me and disappointed me ... it's from a kind of search engine called Spezify. I wanted to be able to click on some of the results and find out where they were from.


For instance, there was this tweet "The UK e-Government agenda has been delivered on tactical infrastructure that will leave an unsustainable legacy for public sector services" ... which I could only find by going to twitter and typing in some of the key words. It turns out to be from David Gale (who, based on his other tweets, you wouldn't expect to be ruminating on government IT but what he says resonates with what I used to warn public sector conferences about 5 years ago).

But there's something very clever about what Spezify does ... I can't think of why I'd want to use it often, but it's impressive nonetheless.

Just How Many Webpages Does Government Have?

After yesterday's post, Dan suggested that we count the number of instances of "gov" in domain names within the "" universe. That seems pretty smart - there will doubtless be a few sites with "gov" in the main part of the URL but he doubts there are too many of those.

Doing that gives the following astonishing result:

200908051210.jpg 200908051210.jpg

Meaning, it would seem, that there are around 112,000,000 (112 million!) web pages within the domain.

If you do the same for (or just .gov since the US is, unlike in every other area, silent on the web), you get 236,000,000

If you switch to French and use ".gouv" you get only 29,700,000 pages

Amazon, by the by, appears to have only 13,500,00 in the UK

Anyone care to contest the figures or have a better way of estimating page count?

Tuesday, August 04, 2009

One Government Website

I've been catching up on posts from Emma Mulqueeny (who, amongst many responsibilities, is at the sharp end of moving content to the core handful of websites that will be government's public face) and Jeremy Gould (who noted in a post about transforming government that it wasn't about closing down websites, which, of course, it isn't - but one sign that you've grasped the fundamental point about joining up government, thinking like a citizen and making life easier for them is that you, inevitably, reduce the number of places someone has to go to get whatever they need done).

[This was the team goal back in 2001]


I've been writing [ranting?] about the need for website rationalisation for a long time, nearly as long as I've been blogging here (about 7 1/2 years I think). I remembered a post I'd written - in July 2004 it turns out - called "There Can Be Only One." It pretty much says everything that I would say now - but I've made a couple of updates, pulled out inside [ ] and added some new charts and slides.

[This slide made its first appearance as far as I can tell in November 2002]


So, this is the post with the edits:

Ian Dunmore [picked up] on a post here from earlier in June that noted that the battle was over for whether there would be only one government site or not, but that people hadn't figured that out yet. He did am ad hoc survey of the regular visitors to his site who filled it up with great comments, coming out roughly 50/50 I think in either total support or strict opposition to the idea.

First up though, the thing that gets me the most is we seem to have an acceptance of either 1 site or 3000+ sites. I've always thought that aspirationally one was the right answer, but I'd settle for 100 or even 500 on the basis it would reduce the problem of information fragmentation. One post notes that "The portal partners can't even agree what should be in the A to Z which, I guess, just shows the absurdity of letting government design websites". A while ago there was an A to Z on a central government site. Where do you think the "Treasury" were filed? T? Ha! It was under "H", for "HM Treasury". People don't think in alphabets and, if they do, they don't think in government alphabets. Besides, with hundreds of services, each set of topics under one letter will cover 3 pages.

An interesting idea was this one "One site implies one entity, one controlling force, no local democracy. How about 4 sites - My Country, My Region, My County, My Local Council !". Whilst I disagree that one site implies one entity (when you read a newspaper, apart from the Daily Mail, do you expect to get only one point of view from it?) as authorship can and would be spread across the entire constituency, I do like the idea of this kind of disaggregation. There's probably a "My Community" site as well - people in or near my area with my interests.

Or how about this one "The practicalities of a central organisation doing this for the country make this idea a joke." One site doesn't mean one controlling entity - it might mean one "voice" in terms of style of writing though. We have 5,000,000 pages of content in government across 3,300 sites. How many of those pages are written in any kind of consistent, understandable, accessible style?

[Out of interest I searched for the number of times "accessibility" is mentioned on websites and was more than a bit surprised to see a count of nearly 1.5 million come back. I wondered how many sites have figured out A or even AA accessibility now - not long ago it was single digits percentage for even just a single A. Sitemorse's survey says things have changed ... bravo!]


One person, obviously well connected, said "like Andrew Pinder, I grew to realise that Departments will just not allow themselves to be joined up". Tell that to the Inland Revenue and HM Customs (Filed under "R" for revenue and "H" for HM in the A to Z). The Government Gateway joins up a dozen departments today, the Knowledge Network over 40. The days of Fortress Government or Super Silos are declining.


"Most people look for something via Google or some search engine or other" - go type in "disability living allowance" in google and restrict it to and count the occurrences (16,900 today, up from 9,900 a year ago). Tell you what, type in "I'm a new parent, what can government do for me?" and see if it works. Search engines are great when you know what you want, but they don't find what you don't know nor do they intuit what you might want.

[I used to track this kind of statistic out of mild interest but long ago stopped. I just ran "disability living allowance" through google and the answer astounded me: 156,000. Wow! 93,000 of those are in the last year. The good news is a large chunk of page one's results were on, but not a large chunk of the ones from the last year. Google also has a new thing called a "wonder wheel", which looks like this for DLA]


But, actually, "The more there are the more competition there is, the better sites become". I'd missed that - I hadn't realised that government entities were supposed to compete against each other. I thought we were in the business of serving the public and making it easy for them to find things. Besides, the more money we spend competing, the better, right? We must spend north of £1/2 billion a year on websites right now - another couple of hundred million widely spread would get us what exactly?

[I don't have details on the current web spend and I'd like to think it's a lot less now, but I wonder with all of the effort going into twitter, social networks and so on whether we have slimmed the technology cost but massively increased the human spend. That might be no bad thing perhaps]

I'm delighted that so many took the time to respond and I have, in turn, responded largely in the spirit of the posts that were made. My contention is: - 3,000 sites is too many; the right answer is closer to 1 than 3,000 - 5,000,000 pages is too many; too many are out of date; too many are never looked at; the cost of maintaining a page that's never used is infinite as a ratio against usage.

[I was wondering how to count how many web pages government has now. Dan came up with a way once - I think he searched for a specific word that we thought might be in the footer of every page. I just tried with ' home - "home office" ' and got 21,600,000. If you have a better idea, I'd be interested]

[Google also has a timeline graph - this is all new to me, I haven't noticed it before so call me an idiot if it's been there for ages. This is the timeline for DLA again. The first reference is to a document published in 1991. I don't know for sure, but I am pretty sure, that DLA guidance changes every year and so much of this historical data is irrelevant. Why no one made any updates in 2005, I don't know. Data doesn't necessarily cost anything to keep online of course - but it will cost something when it gets looked at just as someone decides whether or not to migrate it to a central site; and it certainly cost something, both in human and technical spend to get online in the first place]


A central site doesn't have to do all things for all people, it just has to get most of it right and hand over to specialist sites for things it can't do - just like the tiers of operation in a call centre, e.g. 80% of calls by first line, 15% by second line, 5% by third line. If the third line sites were specialist ones for specific local scenarios, wouldn't that make more sense?

[And a slide from a May 2002 deck


Duplicating content tens of thousands of times increases the risk that it's wrong, increases confusion for the customer and reduces the chance of landing in the right place first time, wasting time (for the customer), money (for government) and bandwidth (for everyone) There can be only one, but I'd settle for 50 or a 100 to start with. I'd like there to be another round of comments on this, that would be fun.

[And just to finish, here's a slide from the business case for that was written in 2002


Anyway, to my mind, how many websites a [central] government should have is a tired old debate - the answer is one(ish). And, just to be clear, I'm zeroing in here on how the citizen (you and me) finds out information about government - what it can do for us. I'm not going for the "we need one place where all of our data is and then we can just hand the keys of our lives to government." That's a different topic, and one for another day.

More from the Opposable Thumbnail

This stuff is great


There is a whole new series of management and motivation posters in this blog, waiting to be unleashed on an unsuspecting world

Sunday, August 02, 2009

Seen better days


Offline Tax - Doing It The Hard Way but, for some, The Only Way

200908021910.jpgI've just spent an afternoon doing a paper Self Assessment Tax Return for someone who is partially sighted. He can't read the forms and certainly can't do it online.

The so-called "Short Form" - only 4 pages long with page 1 being "name and address" and page 4 being "bank account details" - comes with a 25 page guide for how to fill it in (ok, 24 pages, one is left, apparently intentionally, blank.

The guide is littered with advice like this gem at left. I guess the good news is that anyone educated in the 1940s and 50s is probably more capable than those being educated today to do the maths - BODMAS rules ok. But, double brackets in a tax guide? And don't get me started on mixing W8 and 5.3.

But, surely it's time to get real:

1) The bulk of the information required in this case was already available to the government; and, in any case, the amounts involved are so small as to be negligible to the government, even with its current restrictions

Surely, in this age of "tell me once", if a government has the data, they should ask for what they don't have - it all feels a little bit like we're trying to be caught out now, some kind of hide and seek game where if a government says nothing, we may somehow trip up and provide new data

2) Paper-based forms are going to need to be around for a long time. For many years ahead, there will be people who aren't online, who are partially sighted or who can't handle complex online transactions. Make it easier. Big fonts, clearer text, balanced colour schemes designed to enhance readability. But, more importantly, drive the need down through comprehensive reviews of what data the government already has and how it can make better user of it.

3) There needs to be a massive redesign. I appreciate that the short form is already a big deal and much better than the old Self Assessment (which I have to do) which is a couple of dozen pages. But the short form still mixes pensioners, employees and self-employed all in one form. That would, on the basis of this single anecdotal example, appear somewhat like trying to fit too many non-correlated categories into one form for the convenience of the government, not of the citizen. I am the last person to say that the plural of anecdote is data but, nonetheless, I somehow think I'm on strong ground here.

It's 2009. Getting on for 2010. Five years after 100% online as a mission statement, we should be doing better than this.