Permalinks — a guide for the perplexed at Sutton Council

27 September 2008   

Sutton Council launched their long-awaited new website this week and it’s disappointingly dreadful in many ways. Possibly worse than anything in the design or content of the site is the sad fact that the new design has broken all the inbound links to the site, just like it did the last time and the time before that.

What does this mean, why does it matter and what can be done about it?

Links are the lifeblood of the web. Most of the time, we take for granted that we can click on links and arrive at the pages we were expecting. Nothing is more fundamental to the good operation of the web overall and individual sites than links working and continuing to work.

Now imagine a large organisation with several thousand workers. The organisation has a switchboard system and each worker has their own extension number which is available internally and externally. You are tasked with upgrading the switchboard system to a newer model or a different model from another supplier. You can imagine the chaos and costs that would be caused if the new switchboard system required that all the internal extension numbers had to be changed, so your first concern is to ensure that the handover from the old system to the new is seamless and transparent. The old numbers continue to work in the new system as everyone would properly expect. No-one would be so stupid as to expect to have to change all the staff’s phone numbers just because you’re upgrading the switchboard.

The web is no different. Instead of phone extensions, a website has pages and each page has its own address (which we call URLs). We get to a page by knowing its address or clicking on a link that encodes that address, in just the same way that we reach a person via their phone number.

Many bad things happen when a page that used to be at:

http://www.sutton.gov.uk/news/latest/Lottery-winning+library+set+for+rebuild.htm

suddenly moves to:

http://www.sutton.gov.uk/index.aspx?articleid=3531.

The most obvious effect is that clicking on the first link no longer works. You end up on an error page.

These links could be in several different places:

Changing all the URLs of all the web pages on your site means imposing considerable direct and indirect costs on your organisation and on every individual and organisation that deals with you.

People who run websites incur the direct costs of having to change all their links to your site so that they still work. I run Sutton Active and Suttonboro which between them include dozens of links to various pages within the council website. Or at least, they did. It will be several hours work to fix them all so that, say, the link on this page about Ridge Road Library back to Sutton Council’s site works again. (Depending on when you’re reading this, this may now be fixed.) That’s time that would be far better spent improving my own site than covering up for the deficiencies in others’, and until I do it, my own site is both less useful and appears broken, reflecting badly on me. So, find the direct costs by multiplying the number of inbound links to the council’s site by the time spent to fix them. There could be hundreds or even thousands of these and most of them will have to be fixed manually.

The indirect costs are harder to calculate but are most probably far greater. These are the cost incurred when someone clicks on a link to a page within the council’s site and it doesn’t work. They just get an error page. They will then either have to search or browse for the correct page within the council’s site or they’ll go away without being able to read what they were expecting and gaining the benefits from doing so. The cost in terms of time, frustration and missed opportunities could be huge.

This is particularly serious because it affects Google and other search engines. Try searching for sutton council recycling and clicking on the link provided by Google. It’s broken, as are all the other links from search engines. It could be several weeks before the search engines rebuild their links correctly. At least for them it’s an automated process, but it’s little help to web users in the meantime.

The worst thing about the whole sorry mess is that this problem is well understood among web designers and the tools and techniques for avoiding or fixing it are widely and cheaply available. Designers call the problem linkrot and the solution, permalinks — that is, permanent links. Web usability expert Jakob Neilsen wrote about linkrot (and coined the term) in 1998 — a decade ago! — and the term has its own Wikipedia page and an estimated 60,000+ references online. Yet here is Sutton Council repeatedly breaking its connections to the rest of the web with every new website redesign as if it were blithely unaware that it’s even an issue.

I even wrote to the council about this on 15 August:

The last time the council redesigned its website (two years ago?) it cost me about three hours’ work as I had to fix all the links to various council pages on my various sites. I assume that the new design will break all these links again and that I and everyone else who links to your site will have to bear similar costs. Techniques for “permalinking” (creating web page addresses that are effectively permanent) are now well understood and widely used. I hope that your new website will use them and that the council won’t keep breaking everyone else’s outbound links with every new redesign after the one currently in progress.

http://en.wikipedia.org/wiki/Permalink

and volunteered my time for nothing as a tester for the new site but have as yet had no substantial reply. Bit late now, really.

So how do we break out of this cycle of broken links every time we upgrade the website or move to a new content management system? Let’s look at permalinks.

Just in aesthetic terms, which of these looks better?

www.sutton.gov.uk/environment/recycling

or

www.sutton.gov.uk/index.aspx?articleid=691

The second link is actually the one that currently works.

I hope most people would prefer the first. And that’s handy, because not only does it read better for humans but because the link itself is meaningful and independent of any system or technology, it can always serve as the address of the recycling page for as long as it exists (and implicitly, lives within the environment section of the website).

Being “independent of any system or technology” is also important. Web technology changes all the time. Every year there are new content management systems (the specialised databases that run large websites) and even whole new underlying server systems. If we know one thing for certain it’s that we won’t necessarily be using today’s technology in a year’s time, let alone five years or ten years. To tie our web addresses tightly to the underlying technology is a mistake. In the link structure above, the address is specific both to the particulary content management system and to the web server system on which it runs. (In detail, the server is running ASP.NET which is where the aspx comes from. While making no criticism of ASP.NET, do you really want to be tied to it forever? Will it even be around in a few years?)

Even if a system isn’t going to provide meaningful names for the URLs of its pages, it can still construct simple permalinks that are independent of the content management system and the server technology:

www.sutton.gov.uk/page/691

works, as indeed does the date-based system on my own blog site here. It’s less pretty, but at least we can be reasonably sure that we could transparently move the site from one system to another without it forcing us to change the address style and break all the inbound links in the process.

In 2008, almost the only excuse for breaking all one’s links is to move from a legacy non-permalinked system to a permalinked one. You (and others) will go through the pain and expense of breaking those links but at least it will be the last time you’ll (forseeably) have to do it. Had Sutton Council moved to permalinks in this new redesign, I still wouldn’t relish the task of updating all their links but at least I’d be happy that this would be the last time I’d have to do it. (It’s 99% likely I wouldn’t have written this blog post either.) As it is, I’m asking, how many more times is this going to happen before someone at the council understands the issue and actually does something about it?

Sutton Council has externalised the cost of its poor web design onto the rest of the community. Thanks a bunch, not least because we’re paying for the site in the first place. Here’s looking forward to 2010′s redesign, when there’s just a glimmer of hope that we won’t have to go through this whole sorry farce for about the fourth time.

Stop breaking the web every time you upgrade your website.

Some questions for whoever commissioned the website:

Sutton  Usability  Web design

7 comments

  1. Paul Geraghty
    28 September 2008

    Hey ho.

    Nicely written.

    You have identified and described rather nicely one of the outcomes of one of the major local gov “anti-patterns” – SEO Suicide ( http://www.councilsites.co.uk/about/antipatterns.htm )

    I had the dubious fortune to work for a District council whose upstream County council did this trick very frequently, sometimes a couple of times A YEAR.

    Someone fiddled with the setting of their cms/Lotus Notes and broke every inbound link AGAIN.

    This went on for years. Eventually I learned never to link directly to another website, but to leave the link in a database, so I could regularly check it, and to reference the link from pages – so in pages I had “link1234″, which then looked up the real URL and sent users on.

    It was a blessing in disguise actually, because it meant I could record how many ppl clicked which links and when – and feed this back into navigation.

    I am a techy, and I know Sutton can fix all those broken links by redirecting users, or use their webserver’s software to redirect users. They just cant be bothered, why should they, what are you going to do?

    Go to another council?

    I will link to this posting from my site, you really hit the usability nail on the head.

    You may also be interested in reading about CoolURI s from the W3C

    http://www.w3.org/TR/cooluris/

    p.s. Oh, quickly, quickly … I just spotted a new anti-pattern “Useful links” … I will write that one up, its a real howler. :)

  2. Adrian Short
    28 September 2008

    Hi Paul,

    You’re right, doing a comparison between different council sites does reveal many patterns and anti-patterns. I’d prefer to call this plain-old Linkrot as it’s not the user’s direct concern whether someone’s site gets a good SE profile, though it helps. Their main consideration is when they click on a link it should work.

    I have considered setting up a redirect system similar to the one you describe. Sutton used to have a whole set of named redirects like http://www.sutton.gov.uk/haveyoursay which were used extensively in print and elsewhere. All those are now broken.

    I’m just starting to get Google Alerts for pages on the council site that I’ve already read and know about, because Google is starting to index the site again and thinks all those pages with new URLs are new pages. *sigh*

    I assume this has completely broken any continuity in the council’s web stats, too.

    There’s a wonderful irony in the “Useless links” anti-pattern as demonstrated here. As a campaigner for better urban design I’m keen to see pointless clutter and signage removed from streets. Here the Useful Links section is illustrated by a gratuitous slice of badly-Photoshopped clip art of exactly the kind of sign that would probably be clutter in the street. On the web, both the image and the link section itself are clutter. The links wouldn’t be any less useful if they were randomly generated from a large set. If they’re not contextual they’re clutter.

    More broadly, I call a Links section or page within a site a Link Ghetto. I know from the stats of various sites I’ve managed that no-one ever uses them. Some organisations like to maintain them through inertia or due to the misplaced notion that there is some value (either “real world” or in an SEO sense) in this kind of reciprocal linking. There isn’t and those things should just go.

    Always consider the maxim, “Every time you add something you take something away”. If you can do without something, you should. Crap Clip Art (or even just Clip Art, it’s the same) is another anti-pattern exemplified in spades by the new Sutton Council site.

  3. Paul Geraghty
    29 September 2008

    Google Alerts – that’s how I found this site :) Critical piece of kit.

    Link Ghetto. ROFL That is sweet.

    Oh, yummy, how about “Crap art”?

    Yeah, my eye was drawn to that awful visual hack, but mostly because I am totally in awe of the “Useful links” label, and the mentality behind it.

    I especially roared at the way the sign pointed into the centre of the home page, bliss.

    But why “Useful”?

    Listen, we have got a load of links and we thought we would sort them out, useful ones and crap ones.

    It always begs the question, where is its sister sign, “Useless links”? It could balance the page up so nicely couldn’t it?

    Now of course that would be silly, so why don’t they just call it “Links” – sounds too much like golf?

    This all reminds me very much of the first days of the telephone, when it was rare that anyone had a phone. In the village I live in, next to my inlaws in fact, my inlaws are very proud of the fact that their home was #1. And in the local mayors office there is a “phone directory” from the 20s or 30s with all the phone numbers in it. There are less than 10.

    No town keeps its own record of phone numbers any more, there is a discovery system which is well documented and supported called the phone directory.

    Now I remember in 98, when I started in local gov, that we had a page called useful local links – and it had hardly anything on it, just a link to the local am-dram society and the few local councils around us that actually had a website. Now these were useful local links because they were the ONLY local links.

    That was the equivalent of the phone directory in the 20s.

    Useful links now is an anti-pattern caused by ineptitude.

    My definition of Useful links is “External links on a page which can reasonably be judged to be in context with with helping a visitor to complete his internet journey in order to fulfil some likely task. That task can be divined from the title and meta data contained in that webpage.”

    In this case those links should not yet another click away but should appear on the page, managed centrally and clearly marked, and ideally carry some meta data describing what is important about that link.

    This rarely happens because of many forces – some of which I have documented. ;)

    For a variety of reasons, maybe because they are technically inept, they (Sutton) cannot do what they should do, therefore they do what they think is the closest match (“Useful links”), and after all that’s what they do down the road at “Council Y” – et voila, another anti-pattern has been teased out from the digital duvet-cover which is council sites in the UK.

    I only published that anti patterns stuff to get it off my chest, and I had thought of eventually making blog entries out of them.

    You’ve given me new momentum.

    ps Did you ever read Catch 22?

  4. Adrian Short » Blog Archive » Fixing Sutton Council’s usability with Greasemonkey
    29 September 2008

    [...] Adrian Short Design, citizenship and the city « Permalinks — a guide for the perplexed at Sutton Council [...]

  5. Adrian Short
    29 September 2008

    Now if you drop Useful Links and the two Crap Arts from the right hand column on the home page you’re just left with the three links under Have Your Say. That’s not much real meat for a column that takes up 25% of the space of the most important page on the site.

    But once the space is freed up it could be used for actual content (radical, I know). How about listing the current consultations rather than just providing a link to them? Of course, this would help if each consultation actually had its own page or at least an in-page target, which they don’t at the moment.

    Something tells me that this CMS isn’t able to create categories of pages and then automatically generate lists of the pages in those categories. If it can, the consultations page is an obvious place to do it:

    http://www.sutton.gov.uk/index.aspx?articleid=2643

    A couple more anti-patterns for your site:

    1. Clock/Calendar

    Generating the current date/time and displaying it on a page for no reason than to fill space and make the site look current. I’ve written about this today here:

    http://adrianshort.co.uk/2008/09/29/94/

    2. Information Architecture Leftovers

    Inspired by the usefully-titled “Information” section in Sutton’s left-hand nav, IA Leftovers are what you get when you do a weak IA job and end up with stuff that can’t be meaningfully categorised in your hierarchy. So you create categories with vague and meaningless titles to hold the arbitrary and unrelated (according to your analysis) leftovers.

    Or is this section called Information to distinguish it from all the non-information elsewhere?

    You can see a great example of Frankensite if you go here:

    http://www.sutton.gov.uk/index.aspx?articleid=481

    and click through to “Councillors Information” (sic)

    This shows the old site template running on Moderngov, complete with the old navbar replete with broken links to the old URL schema.

  6. Paul Geraghty
    30 September 2008

    ROFL OMG, I better start rewriting that stuff this weekend.

    I was absolutely astounded to see the signpost on their REAL site is still there, I thought it was just on the screenshot on your blog.

    It dawned on me actually, that this “Useful links” is actually all the external sites they feel obliged to link to – just bundled up together.

    I am rethinking my description of the links anti-pattern as a result.

    You see, I suddenly got a flashback.

    I have a feeling that Useful links list is an indicator of a botched attempt to respond to a number of internal pressures.

    “The committee voted that the link for this should be shown on the home page of the website” – is one of the most frequent.

    Very often I found out this was actually not the case, but the deliverer of the message tried to force my arm up my back in order to get a link to some obscure, but precious project.

    Now what is a webadminer to do? Stick them all in Useful links is one way out – its a bit like the link to TransportDirect.info, a fantastic and most laudable site – but instead of finding out how TDi works and how you can use it to allow ppl to find travel plans to the council offices, they just stick a link in Useful links. Box ticked, job done, back to minesweeper.

    The link to direct.gov.uk – now that is mandatory (!)

    What’s a webadminer to do? Do what you are told, or put your foot down and start upsetting people. You cant link to EVERYTHING from the home page, so who gets their own way? The most powerful.

    “the chief exec wants this on the home page now – and she says don’t change the text at all”

    And to which usability class the chief exec go to?

    Which delegation classes did the chief exec go to? perhaps more to the point. Chain of command anybody?

    None of the above is fictional, it happened to me – a lot, but I developed strategies to get out of it.

    So you see how I take bad home page design to be a barometer of poor management style, this isn’t just an organisation with a usability problem – but it could well be an organisation at war with itself.

    Frankenstein site, yeah what a jolt that is, but its not their site, if you look at the URL.

    http://sutton.moderngov.co.uk/mgMemberIndex.asp?bcr=1

    Good job I didn’t use their css text-resizer to change the size of the text, all that would have changed too when I made the ‘seamless’ journey between the two.

    Interestingly this part of their site (?) features parts of the LGNL (Local Gov Navigation List) which seems to have now been abandoned in their “new improved” site.

    A truly questionable decision, probably made by a committee, a committee ignorant of the facts.

    Its like the fuzzy “Do it online”, wtf? That’s another pet hate.

    What, do it? What like a rabbit?

    Do it, go on do it, its like a dare.

    Do friggin what? I’m already online! Really, its just a crock.

    When you click that link, you just end up on another intermediate page with 2 options, then say you pick “pay it” you go to another holding page before passing another Frankenstiein Jolt into:

    https://www.e-paycapita.com/sutton/index.jsp

    Cripes ! Whats happened to my text size?

    I just had that the right size, now its reduced to 8pt again.

    Hang on, where’s the lovely A-Z gone, where’s the search box? What, is plain html too hard to add to a template all of a sudden?

    I have to stop and get some work done. Thanks for replying with all those points.

    You’ve given me loads more ideas. IA Leftovers, hmm, gonna have a think about that one, see if I spot it elsewhere …

    Paul

  7. Adrian Short » Blog Archive » Some pleas to reduce WordPress misery
    13 October 2008

    [...] isn’t a category, it’s information architecture leftovers. Make the app work with no categories and start like that by [...]

Your comments