Archive for July, 2007

Accessible - just because you say it is doesn’t make it so

Wednesday, July 25th, 2007

Been checking out portal “best practice” recently and had a look at Directgov, the UK government portal. The site has a very large, clear well placed Accessibility link at the top right of the page (page top is preferred to page bottom - where it is often relegated) so they must be onto it right?

First glaring observation is the colour scheme. Red and orange backgrounds and red links - really tough to read regardless of how good your eyesight is.

Hyperlink fashion - is black without underline the new blue with underline? (red underline or red without underline??) Bad luck for colour blind users who could use the underline as a cue to presence of a hyperlink.

You can go to the Accessibility page and then click on the link to read about “Information on how we tried to make this site accessible and easy to use for everyone”. Where’s the link? You’ll figure it out real quick because you’re a smart sophisticated user of the web (that’s how come you’re here) - but the audiences of a public service website aren’t all like that. So now you’re there you can read all about how much effort they made to make the site accessible (people with disabilities really love to read this stuff - not!).

The clincher - the image of a presumably disabled user has no alt text. Presence of alt text is a WAI Priority 1 checkpoint.

This on the page that states “The site’s layout takes into account users who are blind or visually impaired. It is fully compatible with popular screen reading software.”

Another rail crossing death leads us into Rail Safety Week

Sunday, July 22nd, 2007

This month there have been five level crossing deaths in New Zealand. Based on value of statistical life that’s a loss of $12.5m at current valuation or $20m in proposed revaluation in just one month.
Apparently, OnTrack only has sufficient resources to install or upgrade seven new alarms per year (PDF 39kB). Not only are they costly but take specialist resources and time to develop.

Rail Safety Awareness Week is supported by Land Transport New Zealand and we are going to get “a series of hard-hitting television advertisements will screen throughout the week”. Why? “… to drive home the messages that tracks are for trains, and even a moment’s inattention around the rail network can lead to tragedy.” In other words - if you get killed at a rail crossing - it’s your fault. I also hope that the initiators of these “hard-hitting” advertising campaigns are mindful of habituation and consider infrequent showings of these ads which will provide greater long-term benefit than a couple of weeks saturation. The cynical might suggest that a saturation advertising campaign is just there to show that something is being done - but we remain ever optimistic.

Over at the New Zealand Safety Council they have taken a close look at the dangers inherent at level crossings and found a number of safety issues. For example, “Alarms were inaudible at 25 metres inside a vehicle with the engine turned off.” Good luck to Jonah Lomu listening to his car stereo.

Where has “safety” gone from LTSA?

Put “LTSA” into Google and in the search results you’ll see “The Land Transport Safety Authority is a stand alone Crown entity charged with promoting safety in land transport at reasonable cost.” (my emphasis) Unfortunately that’s out of date - the current website has sparse metadata, so Google is using the old (2004) stuff.

Looking at the Land Transport LTSA website archive page, the LTSA was disestablished in December 2004 and merged with Transfund to form Land Transport New Zealand. The Performance Agreement between the Minister of Transport and the LTSA for 2004/2005 (PDF 450kB) set some Outcome Targets. These include: the 2004 total social cost of road accidents (injuries, fatalities, property, emergency services, road users,…) should not exceed $2.75 billion, rail level crossing injury accidents should go from average 21.3 per year to 16 per year in 2010 calculated on 7-year moving average.

Take a look at Land Transport’s 2006 Annual Report and see if you can find performance data on road and rail accidents and injury. The various Output Classes do not seem to include any actual safety statistics.

This stands in stark contrast to the LTSA Cessation Report for the five months ended 30 November 2004 (PDF 1.1MB) in which the first statistics you find on page 9 are “Road safety outcomes and targets” or the last Annual Report 2004 (PDF 600kB) where on page 11 you’ll see the “Report on Safety Outcomes”.

Given road crashes are a $2.75+ billion problem, where has the focus gone?

Add two more to the 80+ rail crossing fatalities

Monday, July 16th, 2007

We have 2 deaths at a rail crossing on State Highway 1 at Ohingaiti - yet another case of tragic repetitive failure.

To quote Lex Henry from Ontrack, “But I don’t think every time there’s a fatality at a particular crossing, that on a cost-benefit analysis, we need to go off and install [barrier] arms.”

So - let’s look at the cost-benefit analysis. LTSA’s estimate of “statistical value of life” is $2.5m (proposed $4m (PDF 32kb)) and barrier arms cost $150,000 to install. At that rate a barrier arm has to save one life over a period of approximately ten years to be cost-justified.

The Ohingaiti crossing isn’t just some low traffic frequency crossing - it’s on State Highway 1. The driver was a recent immigrant to New Zealand - possibly unfamiliar with our conditions - and we have thousands of tourists self-driving in cars and campervans every year. Reportedly, there has been no “rail collisions” at that crossing for at least ten years - however, it would be interesting to find out its ALCAM (see below) risk score.

Lex Henry is reported as saying that it is not practical to install $150,000 barriers at the 1400 rail crossings in New Zealand - and also states, “At the end of the day there’s thousands of crossings at many places in New Zealand. From a practical point, unless they are high-risk sites… what is the real benefit-cost analysis?”.

Yes, some of these crossings will only be used by farmers and local people accessing their own properties and are comparatively low risk. A rail crossing combined with a tricky and somewhat restricted vision S-bend on State Highway 1 is a completely different matter. Yes, let’s deal with the high-risk sites.

The bigger picture - over 80 deaths in 10 years. In 2005, David George, CE of Ontrack, said “We know from the Australian experience - where they have a programme of routinely replacing these active flashing light signals with barrier signals - the reduction in fatalities and accidents has been 95%.” It looks like barrier signals do work pretty effectively and the Aussies have been addressing the issue. It seems that “driver error” is a high factor in crossing accidents - so, if drivers are predisposed to error then surely it is incumbent to take steps to reduce or prevent error (which may have been the cause of these latest fatalities).

80 deaths at $2.5m a death is $200m loss over a ten year period. At $150,000 per crossing - that’s equivalent to the cost of about 1300 crossings. Perhaps we should revisit the “real benefit-cost analysis”.

In the meantime, it does appear that Transit New Zealand has been examining this problem and “are targetting 2006/07 as the year to apply ALCAM.” ALCAM is the “Australian Level Crossing Assessment Model” and is also mentioned on the Road Controlling Authorities Forum web site.

What has actually been done with ALCAM here is somewhat harder to find.

A new angle on customer-centric

Thursday, July 12th, 2007

Reading “customer-centric” strategy at 90 degreesSo what sort of customer is this “customer-centric” document to be read by?

Not a major problem - I just clicked the button to rotate clockwise 90 degrees - and Acrobat Reader crashed…

Time for alarm!

Monday, July 9th, 2007

LG phone - alarms display shows timeLooks like the designers at LG (unlike Motorola) realise that people who want an alarm function like to see what time each alarm is set for.

Given the position in the display, they even assess that the alarm set time is more important than the given name - right again!

No time for alarm

Friday, July 6th, 2007

Motorola phone - display of alarms - doesn’t show alarm set timeIs there still a market for alarm clocks now that cellphones have alarms built in? Can’t see why I’d use, let alone buy one - especially since my pretty cellphone has alarms for Africa.

What is the most important thing you want to know about an alarm? Pretty much the same answer as what is the most important thing you want to know (once you know where it is) about a timebomb.

I can enable or disable an alarm with one key press - but to find out when it’s set, I have to 1 press the menu key, then 2 scroll down one step to the Edit option and then 3 click Select to see the time  and then 4 Cancel to get back to where I can enable/disable.

It can’t be for lack of room on the brilliant screen that they don’t show the set time.

So I get a free alarm clock with my cellphone - but it’s missing just one tiny vital piece of information - and that detracts from my experience.

Model meeting practice?

Thursday, July 5th, 2007

Roll call, agenda, minute taking - must be a great meeting.

Ambiguous at a glance

Thursday, July 5th, 2007

Had to reboot the machine that doubles as our internal web server. It runs two web servers OmniSecure (port 80) and Apache (port 8080) for different purposes.

A quick look in the system tray and Apache’s running it’s services - no worries.

Apache - system tray message - Running none of 1 Apache services

So why won’t our web applications coming from the Apache server work?