On this week’s show: Paul talks about taking a brand online, Marcus gives some advice about reviewing your information architecture and Ian Lloyd introduces us to the challenges of designing for screen readers.
Paul’s personal news
Just a bit of personal news before I get into the industry related stories. I want to let everybody know I have signed a contract to write a book. The book is going to be primarily for website owners rather than web designers, however to be honest I think it could appeal equally to both. I intend to look at what “client’s need to know about building and running a website” so hopefully it should show by example how best to communicate and work with clients.
The most exciting thing about this book from my point of view, is the fact that I want to write it as a collaborative process with you the boagworld community. I am going to release chapters for you to see in advance of publication and also blog on various aspects of what I am writing. I really want to encourage you to share your thoughts and make suggestions as we go along through comments and the forum. I have already set up a forum thread dedicated to book ideas as well as an initial blog post on the book.
Obviously writing a book is a really slow process, but hopefully it is something that we can all get excited about.
News and events
Building for the iphone
Unsurprisingly there is a lot of information appearing relating to building web applications for the iphone. There is an iphone gallery consisting of hundreds of screenshots of the iphone. This is great if you want to mirror the look and feel of the iphone as closely as possible. There is also the iphone developers guide from Apple which provides loads of great advice. Finally there is iphoney, a piece of software that replicates some of the iphone’s web browsing functionality and lets you see what your application will finally look like.
Of course whether it is worth developing for the iphone at this stage is another matter. I guess if you are trying to reach the tech-savy audience who are iphone owners then maybe. Otherwise it might be better to wait until the iphone becomes more mainstream or other phones start offering the same level of web experience.
I was gutted to miss @media this year. Well, I say gutted, I was actually on a really pleasant family holiday, so I cant complain. However, I did miss a great line up of speakers talking about some amazing subjects. I was particularly depressed to have missed Jesse James Garrett’s keynote on “Beyond AJAX” and “Diabolical Design: The Devil is in the Details” by Jason Santa Maria.
Fortunately the recordings of the @media sessions are beginning to filter out for me to download and listen to. However, note that I don’t call them a podcast. There is no feed that I can find which is extremely frustrating.
Setting that little moan aside, it is great to be able to listen to these speakers even though I did not attend the conference and I would strongly encourage you to download and listen to a few yourselves.
Common mistakes in web copy
Although we would prefer to avoid it, the reality is that as web designers we write far more copy than we would like to admit. As for those of us who are website owners, a substantial part of our responsibility is writing good web copy.
We have talked on the show before about writing good copy but our focus has mainly been on style rather than technical detail. This week, I came across a post about common grammatical mistakes. However what I liked about this post is that it wasn’t focusing on the silly details of grammar that don’t really apply particularly well to the conversational tone of the web. Instead it looked at errors such as when to use “me, myself or I” and the difference between “i.e. and e.g”.
If you ever have to write copy then spend a few minutes to check it out. It only covers the worse offenders so doesn’t take long to read.
A department dedicated to the web
Jeffrey Zeldman has written a post entitled “let there be web divisions“. If you are responsible for deciding who should manage your corporate website then you simply must read this. If you are a mere foot solider then it might not be as relevant but it is still a good read.
Basically Jeffrey proposes that a company website should not sit under IT or marketing (as is traditional) but should be a division in its own right. I am not going to repeat all of Zeldman’s logic, but I have to say I wholeheartedly agree with it.
Websites are simply too multi disciplined to sit comfortably under either department and too important to be caught in an endless tug-of-war.
Paul’s corner: Taking a brand online
About a week ago, I had to give a presentation to a board of directors ,explaining the process we went through to develop a new design for their website. A large proportion of that presentation focused on the issue of brand identity. This organisation had a very well developed style guide and we spent a lot of time and effort getting that guide to work online. My presentation talked about the various steps involved and it occurred to me this might make an interesting podcast section.
I have also put together a blog post on the subject of “taking a brand online” and it is this that I cover on the show.
Marcus’ bit: Information architecture review
I am currently in the process of carrying out an information architecture review for a new Headscape client. I have done a fair amount of IA work over the years but I have found myself particularly enjoying this one so I thought I’d waffle on about what I’ve been doing.
We have covered the various aspects of IA work in previous podcasts – Expert Review, Stakeholder Interviews, Card Sorting and Wire Frame testing. This section is looking at the first of these, expert review, in a bit more detail.
I think it’s worth explaining what I mean by Expert Review. When we carry out an Expert Review we are effectively analysing a client’s existing site content, site structure and naming conventions with a view to creating a new IA based on our experience of using and developing websites. This is a collaborative process with the client – it has to be; we can make logical, usability based decisions but cannot claim to be experts in the client’s particular field.
First things first
I make sure that I have a good grasp of a number of things prior to carrying out an IA review. At the kick off meeting make sure the following are covered:
- Target audience – this is crucial for the development of the IA. It may be that the existing site caters for one group well but another poorly.
- Site aims – is there a stepped process that the client wants their users to go through.
- Design – things like horizontal over vertical navigation can affect the IA.
- Homepage requirements – find out what the killer apps and content are as these will need to be linked to from the homepage.
- Finally, have a general discussion about content and site structure. See what the client thinks is important and what’s not.
Map out the existing site
The first thing I do is map out the existing site’s IA. This is a fairly slow and laborious task but it is the best way to not only learn about a site’s content and structure but also to understand what they do and what they offer.
Be logical, captain
Usually, the goal of this type of exercise is to streamline content into groups and name those groups so that users will understand what’s inside them.
Site’s that have grown organically over a period of time tend to spread content all over the place. It is usually fairly easy, though time consuming, to group content together. There are various methods for doing this; I tend to print out the existing site IA (that I usually create in Excel unless it’s a particularly small site, then I might use Visio) and scribble all over the printout until I’m getting somewhere. Some people like to use cloud/cluster diagrams (either on paper or using software) or there is always the age old method of creating ‘cards’ where each page name is written onto a scrap of paper. This is a bit like doing card sorting on your own where you group the cards into piles and give names to each pile.
We come from the ‘it does exactly what it says on the tin’ school of page/section naming. Marketing departments often don’t! A good example of this is the trend to verbs as section names over nouns. I remember one client wanting to call a site section ‘Enjoy’ when the section covered ‘Leisure Activities’. No prizes for guessing what we recommended!
Labels should be as descriptive as possible. Sometimes this can be difficult when:
- there isn’t much space, for example, ‘How to register for our newsletter’ won’t fit on the average button, even ‘Newsletter Registration’ would probably be too much for a top level, so I would go for just ‘Newsletter’. It’s fairly obvious that the content underneath will relate to the organisation’s newsletter and should logically include registration, whereas ‘Register’ leaves the user asking ‘register for what?’
- Sometimes sites are so big that main sections can include too much differing content to be labelled descriptively. In this case, I would recommend either shortcuts on the homepage replicating the main sections that include descriptive words or create drop down navigation that displays the lower level links.
This should follow some sort of desired path through the site. For example, the client may want users to get a bit of background, followed by an understanding of what the organisation offers, followed by some examples of previous work with a view to finally making contact. This would translate to:
About Us | Services | Case Studies | Contact Us
Users don’t want to have to think (that sounds familiar!); they want to look and understand straight away. Following conventions helps with this process. For example, many sites include an About Us section as the first main section. This usually includes some history, annual reports, job vacancies and contact details. Users looking for this type of information don’t want to have guess that this information might be under, for example, ‘Company Background’ which is located at the far right of a horizontal navigation.
Collaborate – to a point
When you have created your first draft it then needs to be reviewed by the client, discussed and iterated until everyone is happy. Take on board any changes that are based on your lack of understanding of what the client does but be prepared to stand your ground on issues relating to web conventions and usability – after all, they’re paying for your expertise.
Ask an expert: Ian Lloyd on screen readers
On this week’s show we have Ian Lloyd giving us an introduction to the world of screen readers. I vividly remember the first time I heard a screen reader being used. I was gob-smacked by how painful an experience it was and I am still amazed that anybody manages to use them effectively.
It struck me that many of you listening to this show might not have heard a screen reader before. Hearing what blind people have to work with really makes you take their needs seriously and so I thought I would get Ian on the show to give you a taster.
In his segment, Ian takes us through some classic problems that screen reader users experience. Unfortunately to best understand what is going on in some of the examples you need to see what he is doing. In order to get around this problem Ian has made a screencast to accompany the audio. There was too much detail to make it available online or via your video pod but you can download the screen reader .mov file here.
What follows is a transcript of Ian’s section of the show…
Hello Paul, Hello Marcus and hello to listeners of Boagworld. This is the ‘Ask the Expert’ section and today I’m going to be talking about screen readers.
Now, I don’t actually qualify [meant to say classify!] myself as an expert screen reader user simply because I don;t use one on a day-to-day basis, because I’m not forced to; I do have good vision. As such, the way that I would use a screen reader would be different from someone who has to use it on a day-to-day basis. That said, I still think it’s useful to demonstrate to people what a screen reader sounds like. And the reason for this is that as far as I am aware on your podcast although you’ve talked about accessibility a lot and mentioned screen readers I don’t believe we’ve ever had a demonstration of what they actually are like for people when pages are built incorrectly.
So, today I’m going to be showing a few problems using a screen reader. I’m also going to be doing this as a video, so this is a screencast. I understand that at the end of this you will be providing a URL for listeners so that they can access this and view what’s happening on screen. Because of course it’s all well and good to listen to this stuff but to get the full context it would be good to actually see the video as well. I will try my best to describe what’s happening on screen throughout this podcast though.
Now the first example we’re going to look at is Amazon dot com. And somewhat cheekily I’ve brought up the page for my own book on Amazon. And, er, just having a look around at what I can find on the screen and there are some issues there. So, let’s have a look at this.
[Screen reader reads out page graphic correctly ‘Build your own website the right way using HTML and CSS, Link graphic’]
Oh, so that’s not too bad. I’ve just found an image there and it’s announced it correctly because it’s found a suitable alt attribute but underneath there are a couple of thumbnail images which, if I want to access those, it gives me a whole different … well, hear for yourself:
[Screen reader announces: ‘See larger image, Link’ then moves to next link, the thumbnail image and reads an unintelligible string of characters – numbers letters and underscores – out to the listener].
Mmm, doesn’t make an awful lot of sense does it? Let’s try the next image:
[Screen reader reads out more unintelligible characters and takes almost 8 seconds to read it out]
So, what’s happening there? Well, it’s quite simple: there’s no alt attribute defined for that image and so JAWS tries to fill in the gap and, er … oh I didn’t mention earlier that JAWS is the name of the screen reader that I’m using. So it tries to fill in the gaps because it doesn’t have an alt attribute it uses the file name instead and the filename, as is often the case on Amazon, is a right load of old gobbledegook! So it doesn’t give it any useful information about that image.
Here’s another example of the same thing.
[Screen reader reads out an image gallery as ‘thumbs slash zero, thumbs slash one, thumbs slash two’ etc]
So this is actually a photo gallery, erm, with a bunch of thumbnail images hence it’s reading ‘thumbs’ because that’s the folder where the thumbnail [image] is actually in and it’s reading them sequentially as well. It doesn’t sound quite as painful as the Amazon example but it still doesn’t tell you any useful information about the images on the page.
[Screen reader announces more examples from the same page]
So let’s listen to a slightly improved version of that:
[Screen reader announces the same images but with appropriate alt attributes, e.g. ‘The Mystery Machine, driven by Scooby’ for a photo of a camper van that is painted like the Mystery Machine from the cartoon Scooby Doo]
If we were to look at that on the video I’m showing that page with the style sheet disabled and the alt attributes displaying inline next to the image. As you could hear in the second example it was far more usable – you could actually understand what the image was about (as long as you understood some of the VW terminology used in there), whereas in the first example none of the images actually had alt attributes so it was just trying to read out the location of the file.
So let’s look at another example.
[Screen reader announces content of new page ‘Page has no links’ and then starts reading subsequent page content before I stop it]
What I’m looking at on screen is a page that seems to have a page full of links. But if you were listening carefully to the beginning of that, the screen reader thought otherwise. I’ll just try to find that again for you.
[I scrub back in the video clip to find the part where the screen reader says no links]
So, you could have a page that’s full of links that say ‘click here’ but of course that’s another problem all in itself. Let’s have a listen to that:
[Screen reader reads ‘Click here to view’ repeatedly as I tab through the links on the page]
Yes, so … the problem there is that it doesn’t give you any information at all. And this is actually still quite common. In fact just yesterday I was looking at Facebook dot com (for my sins) and, er, I was quite shocked to find that they were using a lot of this where the link phrase was ‘click here’ as opposed to the phrase that you would really want to have, so for example instead of saying ‘click here for more information’ and having ‘click here’ as the link phrase you would have ‘for more information about our products’. That would be the link phrase. Erm, but if you just use ‘click here’ and you’ve got a whole page of links that reads ‘click here’ this is what you get:
[Screen reader once again reads ‘Click here to view’ repeatedly as I tab through the links on the page]
Basically, completely unusable.
Now the next example I have is of a form, and in this example, er, the form has been laid out using a table. Thankfully, these days, tables are being used less for layout and people are using CSS for page layouts. However, for forms it’s still not uncommon to see someone put a table in there. And, er …
[screen reader interrupts as page loads]
OK, so in this example what I’m looking at on screen is what appears to be, um, well … four text inputs, and then there is a radio button and it’s basically asking for some personal information, first name, surname, your age, place of birth and then a question ‘Do you have a nut alergy’, the answers being ‘no’, ‘yes’ or ‘don’t know’. So let’s see what the screen reader makes of this.
[screen reader says ‘table with two columns and four rows’. I tab to the first input and it reads ‘surname/family name – edit’]
Already we’re hitting a problem. Because the first field that I tab to I can see on screen is *actually* [the one for the] the first name . But the screen reader believed that to be the surname.
So I’ve now tabbed to the second field which is the surname and it didn’t announce anything. So let’s tab to the next field:
[screen reader announces field as ‘town/city – edit’]
Again it’s getting it wrong. I’ve actually tabbed to the field that says ‘Age next birthday’
[tab to the next field, screen reader announces ‘tab – edit’]
And *now* I’m in the ‘town/city of birth’ field and it hasn’t told me anything.
[screen reader announces ‘yes – radio button’, then ‘don’t know’, reading the radio button choices]
This is all a bit confusing here. OK, so it’s asking me the question ‘Do I have a nut allergy?’.
[I tab to the next field, screen reader announces ‘Yes – radio button – unchecked’]
OK, so … that thinks I’m at the yes radio button but I’m looking at it on screen and it says ‘no’. So, what’s going on here? Now this is going to be a difficult one to explain on the podcast; this is one of the sections where you really need to see the video. But what’s actually happening here is we’ve got a table to lay out the page and the text sits above the text input, so for example where we’re asking for first name, the text that says first name is in the first column and the input that relates to that is in a column underneath, sorry, I mean a table cell underneath it in the next table row. Now the reason this is causing a problem is because if you were to actually linearize that table, in other words look at it in the order of the source code you get a very different view of it. And this is what happens with the screen reader. So if I were to look at this form and read it out in a linear fashion, it goes like this:
First Name [text]
Form input for First name
Form input for Surname
Town/City of birth [text]
Form input for Age
Form input for Town/City of birth
And so on. The problem is that the screen reader expects the text for that input to appear before that input, and because of the way this has been laid out it really really gets things confused. As I said, this is quite a difficult one to explain on the podcast but if you look at the video clip you’ll see why this is causing a problem.
[screen reader blurts a few things out as I try to manipulate it … poorly]
Sorry about that, that didn’t add anything useful at all. Hopefully Paul can edit that out!
OK, so …. the big problem here is that you may be asking a question as we have here that says ‘Do you have a nut allergy?’ and the answers are ‘no’, ‘yes’ and ‘don’t know’. But if you do put the form elements in the wrong order you’re gonna have a problem. And the reason is obviously that with a nut allergy that could be a life or death situation. You could be filling out a form as a blind user and you select what you think is the ‘yes’ radio button but because the form has been poorly laid out and doesn’t have
OK, one final example. Now everyone’s talking about AJAX, it;s the buzzword of the moment. Unfortunately it’s also not very good for screen reader users. And the reason for this is that, er, anything that is updated on the page after page load is very very problematic to pass on to the screen readers. now the example I’m going to give here is a fairly simple one, and it’s the Google Suggest page. What Google Suggest does is let you type in your search phrase and as you type it’s calling back to the server, finding suggestions for you which it then populates in a list underneath the search input. So let’s have a listen to that.
[screen reader announces: ‘google search – edit, type and text’ then reads each letter of search term ‘this is a test’ as I type]
So I’ve just typed ‘this is a test’ and on screen underneath that is a whole bunch of suggestions that it has found. But if I try and actually access any of those using the keyboard:
[screen reader announces ‘Google search – edit, Google search – edit, Google search – edit, Google search – edit’ with each keypress on the down arrow]
It’s actually doing nothing. On screen I can actually see that it’s going up and down the options but the screen reader, it’s getting nothing back at all, nothing useful at all.
[more screen reader confusion]
Well thankfully with Google Suggest this is something tat you can opt out of – you don’t have to use Google Suggest, it’s not enforced on you. But it’s a very simple example and it just goes to show that a very simple technique like this can be, basically, completely unusable for someone using a screen reader.
So, that was just a few examples. Hopefully you’ve had an indication of how a poorly built website or web page can affect a user. the bottom line is, keep listening to the podcast, keep doing things right, keep using good markup and, if you can, test your own web pages or web sites using a demo version of JAWS. It really does pay dividends to find out how this works – or doesn’t work. So, thank you very much, I hope this has been useful, and I look forward to the next podcast, Paul. Thanks guys.
No show next week
Unfortunately there will be no show next week as I am away speaking at the Institutional Web Management Workshop. However we will be back the week of the 23rd July.