Tuesday Jan 02, 2007

Identity mis-management

A couple of things recently have brought home to me the importance of identity, or more specifically proof of identity.

Before Christmas I flew to Scotland for a JavaUK event at the Sun facility in Linlithgow. I arrived at the airport, used an automated check-in machine, passed through security and boarded the plane. At no point was I ever asked to prove my identity. I didn't even need to use a credit card to get my boarding pass, just the e-ticket reference number was enough. Whilst I realise that this was an internal UK flight, it still seems somewhat ludicrous to me with the hightened security to prevent terrorist attacks that I was not required to prove who I was.

The second thing was more serious. Over the Christmas break I received a letter from my credit card company saying they were concerned that some transactions made with my card were fraudulent. It turns out that somewhere my card was skimmed, meaning that my credit card identity was stolen. Here in the UK we've switched to chip-and-pin in an attempt to eliminate this possibility. The idea is that, rather than using a signature, which is written on the card and therefore relatively easy to copy, the PIN is stored in a chip on the card which cannot be retrieved unless using secure technology. Whoever skimmed my credit card realised this and used it in the States where it seems anyone can use a credit card with almost no checks on proof of identity (many, many times I've used a credit card in the US and been given the card back before I've signed the receipt). What impressed me most about this was how quickly the credit card company brought this to my attention. After only two transactions, something must have been flagged as a further three transactions were rejected. The software used to analyse spending patterns must be very sofisticated, as I don't have a 'normal' pattern (recent months would should legitimate transactions in at least nine different countries).

Biometrics should be the answer to this, although at one office where I worked they were installing finger print sensors for door access one of my colleagues raised the question of, "Do they still work if the thumb is no longer attached to the hand?"

Thursday Aug 17, 2006

You know when you're using NetBeans

It took me a while to make the switch to using an IDE for software development, but I now know that my switch from vi and make to NetBeans is comlpete. How do I know this? Well, if I look at the keyboard of my old laptop certain keys are more worn than other. In this case both the left hand shift key and the 'z' key are heavily worn. This is a clear sign of heavy vi usage, since ZZ is the sequence to save and exit a file.

My latest laptop, a shiny red Ferrari 3400 is also showing distinct keyboard wear, but this time the z key is hardly touched. In keeping with the popularity of letters in the English language the 'e' and 't' keys are the most worn looking.

I've been using Matisse for some GUI development and this is a major boon to reducing the amount of typing required to build a GUI, and also saves on paper, since you don't have to draw out all the components and work out what bizarre combination of Panels you'll be using or figuring out the necessary incantations for using GridBagLayout.

Thursday Apr 20, 2006

Appropriate choice of error message

Back in the UK and it's time to do some work in the office. Part of this involves extensive use of one of our internal applications. I won't name names, but anyone at Sun will know what I'm talking about when I say that it's a very frustrating application due to the way the user interface has been 'designed'. Good user interfaces are very difficult to get right, as anyone who's tried to build one will know, but this particular applications suceeds in being really bad.

Here's an example. I mistyped my badge number on the login screen and was presented with the following pop-up error message:

Somehow I don't think this is the best way to explain to the user that they typed in the wrong login number.

Sunday Apr 09, 2006

Viva Brazil!

I arrived in Rio last night after a rather bumpy flight; I kind of like a bit of turbulence, as it does make it feel like you're really flying, rather than just sitting in a long aluminium tube (I had a rather heated debate recently with some of my American colleagues about the correct spelling and pronunciation of alumunium as opposed to aluminum; I suggested, rather cheekily, that the 28th ammendment to the US constitution should be the right to drop vowels).

Brazil is one of my favorite places, the people are so friendly, the food is so great, and all the taxi drivers seem to think that the spirit of Ayrton Senna lives on in them. I managed to get to Marius for dinner last night so I am one churascaria up so far this trip.

Monday Feb 27, 2006

It's nutritional, but...

The weekend before last I went to Sardinia to talk to the Java User Group there. It's trip like this that make me think I have one of the world's best jobs. The presentation was on Saturday morning and the room was packed. This is the thing I find remarkable: the fact that so many people are prepared to give up their free time, including weekends, to come and learn about Java.

The enthusiasm of the attendees was great and there were lots of very good questions relating to the future of Java, the recently released beta of Mustang and the open source app server project, Glassfish. The organisers are really dedicated to helping promote the use of Java and, like the DFJUG group in Brasilia who have the Rybena project they have an open source project, AVIS. This is Java based system for helping alert blood donors via SMS when they're needed urgently.

Aside from the technical talk and questions the JUG leaders extended some wonderful hospitality to me. After the morning event we went for a traditional Sardinian lunch of riccio which is sea urchin. I have to confess I'm not a huge fan of shel fish, but I thought I'd give it a go.

I would have to say that it tastes better than it looks. We also went out for dinner in the evening for another great local meal, shell fish, grilled fish and even snails.

All in all a great weekend.

Friday Nov 18, 2005

Who's driving the car?

I was in Munich yesterday at a suppliers fair organised by BMW. The idea was for the IT department and various suppliers to show the rest of the company what they were up to. I did a presentation on innovation at Sun, talking about our work with the Mars Rover, Sun SPOTs and Team Jefferson in the DARPA Grand Challenge. (It hadn't dawned on me until I arrived that talking about the Mars Rover might stir up some unpleasant memories for the BMW folks).

It was particularly interesting talking to various people about some of the research being done on safety systems for cars that are starting to appear and may appear in the future. We already have drive-by-wire in many cars, where the accelerator is not physically connected to the throttle. Instead a sensor on the pedal determines the position and sends a signal to the engine management unit. If the EMU needs to override this for safety reasons it's very easy to do. It looks like this will continue and we may well soon see cars where the steering wheel doesn't physically turn the wheels. If the car starts to skid and the electronic stability program (ESP) can't cope simply by throttle and brake control, it'll just take over the steering as well. Probably going to be a bit scarey the first time that happens. Several companies already supply adaptive cruise control that will keep the distance between you and the car in front a safe distance, applying the brakes if need be. Ultimately, I guess we'll have cars that drive themselves (although the demonstration I saw of this from Mercedes is still some way off from being a usable feature).

Of course Germany is famous for its Autobahns where there are stretches that have no speed limit. It was amnazing to see the speed that some people went past my taxi (which was travelling at close to 100mph). I guess you can see why many of the big German cars have a 155mph speed restrictor fitted.

Monday Sep 19, 2005

What's wrong with Nigel?

When you buy an iPod online from Apple you can have the back engraved with some sort of inscription to make it harder to resell if it's stolen. Just don't try and have your name engraved if it's Nigel. Truly bizarre, but rather amusing.

Wednesday Jul 20, 2005

At last!

I don't want to condone violence, but...

Somebody has finally provided an outlet for all of us who find the Crazy Frog so irritating. Check this out.

Tuesday Jul 12, 2005

Too esoteric for google

Whenever a question comes up that I can't answer my first thought is google it. Google is now a verb as well as a noun.

This week, however, the power of Google has failed me twice.

I've been using NetBeans 4.1 for a few weeks now and have become a major convert to the power of this tool. It really does make writing code easier. However, I was working on some code that used JMF and ran into a problem: native code libraries for the runtime enviroment. From a shell I can set the LD_LIBRARY_PATH (obviously this is Linux, Windows just isn't an option as a developement platform to me). In NetBeans, there seems to be no way to set LD_LIBRARY_PATH for an application. A bit of searching on Google reminded me that -Djava.library.path should do the same thing. Sadly, this didn't work, either from witin NetBeans or from the command line. This perplexed me and I was at a bit of a loss to see how to get this to work within NetBeans. I consulted a friend and he suggested updating the /etc/ld.so.conf. Adding the appropriate line to this file and running ldconfig and voila! JMF runs from within NetBeans.

The second Google failure was for my shiny new Acer Ferrari laptop. I wanted to have three OSs installed on it, XP (for those rare occasions), Linux and Solaris. XP was pretty straightforward once I'd recreated the partitions. Solaris was also a breeze thanks to all the work done by some of the Solaris engineering team to make this probably the best supported laptop on Solaris. Linux, however, proved to be considerably more tricky. At first I tried a version that had an early 2.6 kernel. This didn't really work too well since the ethernet port and built in USB hub clashed IRQs and neither would work. Then I upgraded to SuSE 9.3 which had a 2.6.11 kernel. Things looked much better as the mouse started working, but the network, despite succesfully getting an IP address via DHCP, stubbornly refused to send packets anywhere. A search on google didn't help and it was once again talking to someone familiar with this that revealed I needed to add apm=off acpi=no irq to the boot options.

It appears that Google is still a long way off from indexing the whole of mankind's knowledge (which may actually be not such a bad thing).




Top Tags
« February 2015