Google recently launched the Google Blogsearch. A first I thought this was a typically good Google product: fast, up-to-date and reliable. Then some spam started to appear in their index and it became rapidly less useful.

Now it is approaching uselessness for searches ordered by date (and what other kind of search would you want for a blog search anyway?)

For instance, compare this search for email howto mac on Google Blogsearch, IceRocket, Feedster, Blogdigger and Technorati.

Right now 7 out of the first 10 results on Google are obvious spam, with no pretence of being anything else. IceRocket and Feedster have a couple of spam items on their first page, while Blogdigger and Technorati seem to have results that at least look spam free. Those results actually makes Google look better than it actually is: it looks like roughly 40 out of the first 50 items Google are showing are spam, and all come from the the same couple of sites.

I aren't normally one to criticize Google; I think that the Google search engine itself is a lot better than the competition. I have to wonder, though: with all those highly qualified PhD's why is that they can't detect that a site called 1001-the-most-complete-broadband.info which contains nothing but alphabetically ordered lists of linked keywords is a spam blog?

Most commercial GPS recievers stop operating above 1,000 nautical miles/hour. That's fair enough I guess – it makes it more difficult to build your own cruise missile (found via a Nitrous Oxide Hybrid rocket picture).

Congratulations to England on winning the Ashes (I never though I'd live to have to write that!).

The jokes have already started:

Q: Why didn't the Aussies need vaccinations when they came to England?

A: Everyone knows Aussies can't catch anything.

At least we beat them in the Soccer.

I've had to pull back from my involvement in the Apache Harmony project because of pressures from other things.

However, Sun have posted a very disappointing document called Java SE IP Issues, which I have to post about.

This document is quite possibly the worst example of FUD I have seen from Sun. While it says it was written to address common IP questions around creating an independent implementation of Java SE, especially around the Apache Harmony project it does no such thing. For example, instead of discussing the copyright, trademarks or patents, it says:

No where in the document does it explain exactly what this “Intellectual Property” is, nor how it is protected. It is clear that this “Intellectual Property” is NOT implementation techniques, since that is dealt with elsewhere (This implementation-specific IP is not licensed as part of the Java SE IP which is licensed through passing the TCK.)

There are clear warnings that are aimed at open source Java implementations that have been careful not to call themselves Java (eg, GCJ):

Note that an incompatible implementation is not entitled to use the Java SE IP. This applies even if the incompatible implementation avoids use of any Java brands or Java trademarks. Having an incompatible implementation call itself “fred” does not magically resolve the IP issues!

This document is not going to win Sun any friends in the open source community, and will only increase suspicions that Sun do not really understand the community they are trying to engage.

My last 6 months work went live this weekend. It's a large metadata repository and associated services which drive a dozen or so Govenment and NGO websites.

It wasn't exactly the most satisfying project I've ever worked on, but it is always good to go live and have things work.

Good luck with that whole Katrina thing.

From the point of view of an Australian, it has been shocking to realise that the reaction to the Boxing-Day Tsunami in places like Indonesia, Sri Lanka and India have been a lot better organized and run than in the US.

It has also been disappointing to see that unnamed US authorities have stopped Australian consular staff from entering New Orleans to help Australians. That has played quite badly on Australian television, especially compared to the response of Indonesia, which – despite not always having close relations with Australia – had no problems letting Australian officals go where ever required after the Tsunami – even to actual war zones like Ache.

Omar Shahine has a great post about FireAnt, Microsoft's AJAX toolkit which they are using for the Hotmail rewrite.

It sounds fairly similar to DWR in the Javaworld, which I have raved over before.

Speaking of AJAX, I came to the sudden realisation this week that AJAX programming can actually be easier than conventional web programming. It's true that tool support for AJAX still sucks, but that is sometimes compensated for by the fact you get to ignore the traditional problem with webapps: the lack fo state.

Transitions between the two programming models are the tricky bit – for instance if you want an AJAX control to be in a particular state when a page loads then you either need to setup defaults for it (which is nasty if the defaults have to change depending on the state of the application), or have an page-load event to trigger updates to the control. Neither option is particularly elegant.

PWNtcha stands for “Pretend We’re Not a Turing Computer but a Human Antagonist”, as well as PWN capTCHAs. This project’s goal is to demonstrate the inefficiency of many captcha implementations.

http://sam.zoy.org/pwntcha/, via BoingBoing.

It still can't solve the Yahoo captchas, though; as it says the Yahoo captcha is “A very good captcha, but not always human-solvable.”

While the complaints against Captcha systens are valid it is unfortunate that there doesn't seem to be a better spam solution at the moment.

The new version of Google Desktop is a beautiful piece of software.

While the features are nice, it is the little things that make it a joy to use. For instance, the animation it uses when scrolling though the extended displays of mail messages or news items is the kind of thing one would have previouly expected from Apple.

The included aggregator is Google's best effort yet (better than www.google.com/ig, anyway – I don't have the GMail aggregator turned on), although I think I'll still be sticking with Bloglines (at least until Findory get its Feed Reader to import subscriptions).

One hint: if you think using it as a sidebar will take too much room set it to AutoHide. Then you get the benefits of both the sidebar and the extra space.

Google Earth has finally added high resolution pictures of Sydney Harbour. They haven't propagated to Google Maps yet, though. You can zoom in quite a lot more than in the picture below – close enough to see the shadows of people out the front of the opera house.