DotP’s Demise; e-government lessons learned

Less than ninety days before it’s 4th birthday, as Dan pointed out, DotP is on its way out. Unlike Dan, I declined to mark the occasion by wearing a t-shirt, but I do think the moment is worth some reflection.

You may not be able to tell the difference between the old directgov and the new one. Here they are, if not side by side, then one above the other.

The one at the bottom is the new site, with a snapshot taken today, the old version is from 10th January. The change of platform, whilst neither here nor there in terms of design, marks the near culmination of a change of strategy agreed 30 months ago when it was decided not to invest further in DotP, the multi-site content management platform. Rather than harp on about the merits or otherwise of this decision, I thought I’d share a couple of fond memories from DotP’s early days.

Lesson 1: Always be clear about your requirements lest they be misunderstood

During the development stage, I was wandering through the offices of the lead delivery company, Sapient, and glanced over the shoulder of one of the developers. He was checking out some screen designs he’d just built. I noticed that the URL had a series of “║” embedded in the text. I asked him what they were doing there and he explained that they were “pipes”. I knew that, but I wanted to know what they’re doing there – I suspect they’re the rarest character in terms of usage on the keyboard of the average user. I told him that was no good, no one would be able to figure out how to type them. I wandered away.

Probably 6 months later, I was looking at the pre-launch version of (it wasn’t yet directgov) and I looked at the URLs again. They were practically 3 lines long – once you got past the home page. I don’t remember the exact comment but I imagine it was a string of expletives followed by a “get rid of those”. I was told that I’d asked for them to be that way “as I didn’t like pipes”. Not quite the way it happened but it seems that the conscientious developer had taken my advice not to use pipes and come up with a different way of handling URLs, but the result was 180 characters of URL per page.

I never again looked over the shoulder of a developer and made an arbitrary comment.

Lesson 2: When you’re right, you’re right, even when no-one agrees with you

One of the things that we decided to day with DotP, right from day one, was not to support inline links (e.g. click here). I don’t remember who made that decision, I do know it wasn’t me; when it was explained to me, I got it right away. We didn’t want inline links for a couple of main reasons

1) It creates a sight that is fundamentally difficult for a screen reader to navigate and we wanted to have the most accessible sites possible. I know that I use embedded links in this blog, but then again, this isn’t designed to be accessible.

2) When links break and need to be removed, whole chunks of text need to be rewritten. In fact, we created a tool that ensured that links internally within DotP (e.g. from directgov to department of health) never broke – if a page was deleted, we killed the link. If a page changed a chunk of content, we flagged up that it needed to be looked at in case the link no longer made sense. Links break all the time and we wanted to make the maintenance job as simple as possible, or even simpler. As it was, external links, surveyed by our external link checker often broke at the rate of a few dozen a day.

When we tried to explain this to other government departments, ie potential customers of our shiny new shared service, no-0ne got it. They didn’t want links to be at the bottom of a paragraph of text, they wanted them embedded in the text because they’d got used to doing it that way.

When we talked about accessibility, few were interested. When we talked about the page maintenance/broken link story, few seemed to get it. We tried all sorts of angles and, every single time that I remember, the final request was always whether we could change the system so that it would support inline links.

We tried this with businesslink, with HM customs, with Cabinet Office, with every department and the answer was always the same. The only exceptions were the folks in the Department of Health and those charged with launching the Child Trust Fund (although there’s another story about a giraffe there that should perhaps remain a dark secret for now).

We were right then, we’re right now. And I expect not everyone gets it still. But I do see that some have adopted the discipline, e.g., but not HMRC.

Anyway, good luck to the new platform for You’ve spent 30 months creating a site that is the same as the old one; I’m looking forward to seeing what new things can be done with the new service that couldn’t be done with the old one.

9 thoughts on “DotP’s Demise; e-government lessons learned

  1. How about left hand sectional nav that stays with you as you descend to article level? Not sure how that one got overlooked in DotP…And anyway, why are inline links so inaccessible? Surely they provide user with important context whether using a screen reader or not?

  2. LOL! Two of my common phrases there.Lesson 1. Tell me what you want, or tell me what to do, but not both.Lesson 2. Everyone tells me I am wrong, but noone tells me I was wrong.I had another interview today. For a technical architect position at a consultancy, (really really close to my home, though the interview was near Fenchurch St ?!?!?.) I was repeatedly asked \”You seem to have done an awful lot of troubleshooting, and business rescue jobs.\” I felt like saying, \”Yes, that\’s because agents will only forward you for the exact same job you\’re doing at the moment, because they have zero understanding of IT, and often can\’t even pronounce the acronym they\’re employing for.Speaking of which, I said to Chris four years ago that DotP was doomed. My reason was that .Net had just come in and all the devs wanted to use that. DotP violated my sixth rule of software engineering, which is \”You can\’t beat the market; people like what they like. In general you can only make people know you do it faster, cheaper and better. \” A famous grassy publishing house I know quite well, has made exactly the same mistake, which was to implement an application framework, the common core platform, over the top of .Net, thus removing the creativity, and hence the staff. In fact they\’ve gone even worse than this, they\’ve moved all their staff (predominantly young men,) from drink and pussyland Soho, to somewhere near where their CIO lives in the styx. Cue huge departure of all their dev staff. The next mistake will be to outsource it to Sapient or the like, whence all control will be lost.Oracle doesn\’t outsource.Microsoft doesn\’t outsource.The government does.Mark you, you know my views on consultancies. (Rule 5. – \”Outsourcing never works; why lose control AND pay someone else\’s dividends?\”)Regards mate.Ian

  3. Mark you, you know my views on consultancies. (Rule 5. – \”Outsourcing never works; why lose control AND pay someone else\’s dividends?\”)never a truer word

  4. How about left hand sectional nav that stays with you as you descend to article level? Not sure how that one got overlooked in DotP…—–looks nice; dotp had a breadcrumb trail that achieved the same thing effectively. i seem to remember it was the \”in thing\” at the time. it looks better the way it is now though,And anyway, why are inline links so inaccessible? Surely they provide user with important context whether using a screen reader or not? —–Try it with a screen reader and see how frustrated you get. better still, get some people in who have to use the things. they\’ll have a view.

  5. Surely the problem lies with inadequate link text, rather than with inline links? \’Click here\’ or the equivalent would be hopeless as screen readers read out a list of links, and in this case the link text tells you nothing about the link destination.But I don\’t see what the problem is with a sentence like \’Visit the BBC Website for more information.\’

  6. Also, I forgot to mention…I know that I use embedded links in this blog, but then again, this isn\’t designed to be accessible.What a bizarre statement, why would you design something to be inaccessible?! Perhaps you just mean that you haven\’t paid special attention to making this blog more accessible, not that you intend to exclude people…

  7. guilty as charged Helen. I haven\’t paid special attention (and I\’m using blogger which I can\’t change much; or, at least, *I* can\’t, others perhaps could). And your point about inadequate text bears further exploration.

  8. Many moons ago I worked on DotP and have since had considerable experience on both commercial and open source Content Management Systems (Vignette, Mambo/Joomla etc.) as well as multiple govenrment body roles.All CMS have their advantages and downsides. As a result, other than saying that DotP was no worse than most and in many respects better than the majority (particularly it\’s approach to accessibility), I won\’t comment on it\’s relative merits.However, whilst all of the comments made here and elsewhere have value I find it interesting that having worked with the users such as DH the main point seems to have been missed.Did the end client (and not the Office of the E-Envoy as was) actually want/need a CMS and why was one written from scratch.Why CMS?If the aim of a CMS is ease in publishing and an authorisation workflow what happens when the authorisation process gets out of hand. If a hard copy needs sending to a third party for authorisation where is your workflow then? A CMS only adds value when it is easier to use and more productive (both for publishing and client use) than the current system for genertaing your website.Why from scratch?With a huge range of CMS offerings on the market why was it felt necessary to build a brand new one? If outsourcing was the route needed then surely working with a supplier (or even using Open Source) and enhancing an existing system would have been more effective and efficient. This is in no way derogatory to the suppliers involved but if you want a customised car on a sensible budget you tend to buy one off the shelf and then modify it. Not build one from scratch.

  9. There were a number of issues with the whole DotP project. And, yes I did have first hand knowledge of the project having worked for a Gov Dept at the time..Too many people at the top,making too many decisions, without knowing what the \’ripple effect\’ was going to be.In reality, it has all been a huge waste of tax payer\’s money, and lots of people\’s time!Shares in Sapient anyone?

Leave a Reply