Out of the way

A couple of weeks ago I spent a few days in Ireland, and what did I do? Staying in Parnell Square, I managed to see very little of the town. I wandered a bit around the trendy bit in the centre I went to a lot of meetings, I sat and worked a lot.

Read more...

Thanks mostly to White_Lynx and the Core developers, you can now try out MathML in Opera 9.5 alphas. My major contribution was writing the article that explains it so Chris could publish it while I am at Web Directions South.

Read more...

I spent a week in Boston, where it got really cold. Like the last time I did that. Without luggage. Like the last time I did that. At meetings, like the ...

Read more...

Err, that's meant to be Bangalore, the first place I have ever been in India.

Read more...

We have been working on the accesskey implementation in Opera. Accesskeys are nice things for people who find it hard to drive the computer - for example people with various kinds of "motor disabilities". Parkinsons's disease, cerebral palsy, various types of paralysis, repetitive strain injuries, etc.

The way people use them in a web page is to declare an accesskey for a link or control. For example

<a href="http://my.opera.com/chaals" accesskey="c">My Blog</a>

is a bit of HTML that makes a link to my blog, and gives it the accesskey "c". Somehow, this is then supposed to help you get to that link extra quickly - instead of having to go through the entire list of links using q/a keys (or tab, and also have to stop at form controls and other stuff, if you use another browser), you can get straight there.

The problem is that Internet Explorer followed a "helpful" suggestion added in the W3C specification, to use "alt" plus the character as a way of getting there, and "cmd" (the squiggly thing also called apple) on macintosh. Netscape followed them, and Mozilla followed Netscape.

Any mac user will tell you that cmd-c is copy, on any mac application in any situation any time. Except (perhaps) on my page. Similarly, if I used the accesskey "f", anyone knows that on windows alt-f opens the file menu - one of the most basic functions of the Windows user interface.

Well, Opera solved that some years ago. Instead of having to guess whether standard shortcuts were being overridden, there is a key to trigger accesskey mode. You press it, and the next key is treated as the accesskey. By default it is shift-esc which is perhaps not the easiest one, but I remap it (preferences -> advanced -> keyboard shortcuts) to "." since I don't use that.

So now I just press ". c" to get to my page. But wait! How do I know?

Well, in all browsers except iCab, until recently, there is no way to find out. The site might mark the keys it uses, or might not.

So now we have a popup that tells you what the options are. Grab a weekly build, find a page which uses accesskeys (they are very common in UK pages such as the BBC) and press shift-escape (or whatever you configured it to).

And give us feedback. We realise that this isn't perfect yet, but I think it is now best-of-breed. (Sadly, that hasn't been a hard point to reach, which doesn't reflect well on any browser). There's more to be done on it, but a start is good.

I trust that my good friend John Foliot will have somethign to say about this by and by (he andd I have been discussing accesskeys for many years already), and maybe others. Many thanks to Petter and others for getting it going.

SMIL 2.1 is a recommendation. In many ways this is good, but in one way it is terribly disappointing. I have carried out an ongoing discussion with John Foliot, in particular, through several years, about accesskey.

It is pretty clear that accesskey is broken. There has been a suggested fix going around for about 5 years - since even before SMIL 2.0 was published. It would have been nice, in one of the rare W3C specifications that has actually dealt with accesskey and touched the semantics (not like HTML, which hasn't changed in 7 years - and still people build tools that get it wrong!) to fix accesskey.

Nope.

Hopefully W3C will deal with the problem soon, with some new work (apparently they are not happy to go with the work done over the last few years). In the meantime, hang on folks. You'll be able to get to work productively in a couple more years, if you're lucky. Revising a specification after several years, W3C can make it work for mobile browser vendors, but can't actually come at fixing a simple accessibility problem.

This is partly my responsibility. Most of the time I am in WAI's Protocols and Formats group, who have the responsibility of making sure this doesn't happen. That group has done quite a lot of work on accesskey in particular. But when the SMIL 2.1 last call was out I was doing other things (looking for work, for example) and not in the group. It seems everyone blinked, and it seems they don't have the clout to actually make a difference if it is going to inconvenience anyone.

Meanwhile W3C convinced the National Rehabilitation Center for Persons with Disabilities Research Institute to write a testimonial about it claiming "everything is accessible". It is very hard to see what happens in practice, since for the only SMIL 2.1 player I could find has Minimal user-level documentation is included in the player. It is similar for other SMIL 2 players - if anyone can find documentation of how accesskey works in a SMIL 2 player I would be very interested. In the meantime we simply hope that the goodwill everyone claims whenever accessibility is mentioned gets carried through into practical implementation. History isn't on our side

For other formats you can use Opera - our accesskey implementation, in common with iCab, doesn't follow the "helpful suggestion" written so many years ago and responsible for so totally breaking the usability of accesskeys in Firefox/Mozilla and in Internet Explorer. I hope that developers copy the approach of one of us, rather than perpetuating the broken User Interface models of the latter two browsers in SMIL.

If you do, I suggest re-mapping the accesskey mode activation to a single key from the default of shift-escape. Go into Preferences, choose the Advanced tab, and Shortcuts. Select either default setup under keyboard shortcuts, or your current setup if you already use a modified set. Then the easiest thing to do is type "access" into the search box, which will bring up the option "Enter Accesskey mode | Leave Accesskey mode". (It's under applications, if you're exploring the huge range of options that are available). Pick your key, and away you go...