Particularly as this redesign is a response to the mixed reception that the 2.5 Happy Cog redesign received.

It’s a bit unfortunate that at the moment I sense an unwillingness for people to say negative things against WordPress in any shape or form – thus my fanboys comment in the post – but I’d guess things will get more discussable pretty quickly when it moves out of beta into mainstream.

I particularly feel a little pissed that I’ve spent some time drawing up documentation that I’ll now have to replace, and retrain non technical users that I only just retrained. In their eyes I can seem less than competent at what I do… how else would they judge.

That being said I’m pretty much tied into WordPress for the next few years via exisiting work and clients so for me it’s a big deal… all I can really do is grit my teeth and rewrite the docs etc…

It would be interesting to see someone like Gez Lemon or Bruce Lawson or Mike Cherim do an accessibility audit of the new WordPress 2.7.

I have been testing 2.7 since the release of Beta 2. IMO, it is far less usable and is also considerably less accessible.

The top-level menu items are not intuitive. Clicking on posts or pages launches the screen for adding a new page or post. Until a user hovers over the area to the right of the top-level menu item there is no indication that this can be expanded. For people who have difficulty using a mouse the area in which to select the drop-down will cause problems (it causes problems for people who expect to work quickly too – I’ve lost track of how often I miss the button!). This is also not keyboard accessible so those who cannot use a mouse will find 2.7 much harder to work with.

The interface is somewhat usable with JavaScript off, but only in a convoluted manner. I am finding increasing numbers of business users have JavaScript disabled at the firewall. With JavaScript off the drop-down button does not exist so the only way to get to the list of posts or pages is by navigating to the new post screen, whereupon the expanded menu magically appears. I cannot see how anyone using assistive technologies, such as screen readers, will be able to find their way around the new interface.

This, to me, calls into question the value of the usability testing that was undertaken. The earlier interfaces may not have been perfect but they did at least work for the majority of people. 2.7 does not.

Jane is correct in one way – entering new posts is faster than with 2.6.x. However, I find that all other tasks are slower and after using 2.7 beta extensively over the past few weeks I am inclined to stick with 2.6.3. The new nested comments is just not outweighing the disadvantages of the UI.

I still have to consider the viability of commercially using a software product that is asking my clients to learn yet another admin interface layout. Regardless of anything you wrote. A usability study does not extract the bumpy move to 2.6 out of the equation and people like me in the middle are going to have to re-explain how everything works.

Would you appreciate if the interfaces were regularly changing in Photoshop – 2 interface designs in six months? Or your Word Processing application? It’s not that I don’t like change so much as I am commercially having to pick up after these changes and it’s costly in the real world, as opposed to the theoretical one at Ball State University’s Center for Media Design.

Now if 2.6 had actually been UNusable, which it isn’t actually… there may be a case. I think Matt will survive this, as will WordPress, but it’s a rubbish way to develop software and a rubbish way to treat the community, regardless of the high moral ground of the user experience. Users, most of all, require a consistent experience IMO.

We’ll have to agree to disagree I expect. But thanks for explaining the process and testing regime. I’ll give 2.7 a whirl when I get the chance.

After identifying all the issues with 2.5 through the testing (we used Morae and eye tracking for data collection), the Crazyhorse prototype was developed to find ways to address about half of those issues (we were limited by time). We gave the test group access to Crazyhorse for a couple of days, then brought them back. We had them do the same tasks they’d done in the 2.5 test, and asked for their opinions. The users were all faster at completing tasks using Crazyhorse, and all said they would choose it over the current (2.5) interface. Even though it was a hacked-together prototype that wasn’t pretty, they still preferred it. That’s the reason behind the decision to change the design so soon, not a movement by developers. For the record, a lot of developers told me they hated Crazyhorse. The switch to side navigation caused plenty of developer complaints on the hackers list.

Many people don’t like change. Some people have trouble adapting to change. But if our test group was faster using the prototype after only two days, compared with using their own 2.5 blogs for 3 months (and previous versions beforehand for most of them), then it seems like the learning curve for the new design won’t be too big an obstacle for most people. We’ll be releasing new help resources around launch time as well, to assist those people who require a walkthrough of the revised interface.

If you were using both 2.7 and 2.6 like I have been lately (2.7 for my test blog, 2.6 for my wordpress.com blog) you might find yourself comparing the ease of tasks on each version. I have to say, I find it much easier to get things done quickly in 2.7.

How many people were really really happy when Microsoft Office 2007 came out and it took a month to figure out where all the ordinary functionality went? Not many. It may have, on the surface, been an improvement in design. However, it messed up pre-existing mental models of existing users in the wild.

WordPress. Please stop changing those admin interfaces…