This week, Google took that idea and extended it into an independent online tool, making it available for anyone on any browser. It uses a version of WebKit installed on a server to run the test (via the Chrome version of Page Speed Im guessing) to test your site and suggest improvements.

It seems to work in exactly the same way, and suggests the same kinds of things to improve your score, and has the same issues I pointed out last time, in that it marks you down for incorrect caching on Google Services you include on your page (Analytics in my case).

When I first ran the test on my site, I had dropped to 76/100, this was down to the javascript includes I had slowing things down, and the image caching wasn’t long enough. A couple of tweaks later and I’m now back up to 96/100 which is 3 above where I got to last year.

Read the Google post or start testing your site, please note tho, that as its being hosted on the Google Labs site, this could (as many have done before) disappear without warning. If it does I’ll update this post to reflect that.

• http://developer.android.com/guide/developing/tools/emulator.html
• http://www.opera.com/mini/demo/
• https://www.blackberry.com/Downloads/entry.do?code=060AD92489947D410D897474079C1477
• http://msdn.microsoft.com/en-us/windowsmobile/default.aspx
• http://developer.palm.com/

Yesterday Mozilla announced the build has reached BETA and is ready for the more cautious developer to start playing with it. Firefox 4.0 comes with lots of new bells and whistles, you can find a full feature list here, most notable the new UI design and enhancements in <hmtl5> and CSS3. There are also lots of new developer api’s for us to play with including websockets and local indexeddb’s.

If your already surfing the beta channel wave, you should be able to just “upgrade now” although my work machine didn’t find it so had to install it manually. Those of you running a stable realease or want to install manually can do so via the usual beta release channel.

For more information on please read the Mozilla Announcement on their blog.

This new system basically allows them to push anything Googlebot finds into the main index in seconds rather than weeks.

Previously Googlebot would crawl the web, find a site, then submit its content to another system for processing. This other system would then sort and process the content and work out what its all about. It would then pass it to another system for inclusion in the index. This was a time consuming process and was also done in batches, which means your site would have to wait for all the other sites in the same batch to be processed before being submitted to the index.

The new system is different, rather than doing things in large batches, it will work on smaller portions, more portions at the same time and submit them all to the index straight away, allowing users of the search engine to see information much much quicker, and this is just the start!

Image copyright of Google

There are several rumours floating around about what lies in store next for Google users and their shot of caffeine.  But for now we will have to wait and just enjoy our new “real-time” search results.

You can real the full article on the Google Webmaster Blog

The main reason for this is because, Fx3.7 consisted primarily of “Out of Process Plugins” which as most of you know has been implemented in Fx3.6.4. This has pushed developers to bypass the 3.7 release and focus on pushing out Firefox 4.0, hopefully by November.

A couple of things that jump out to me are that there will be no more modal dialogs and software updates will switch to background tasks. This is to help improve the user experience as they are two of the main pause points in a using Firefox.

The background process updates I can understand, Chrome has shown that this is by far the best way to push out updates and bug fixes to users and ensure that everybody is running the same version across the board. The removal of modal dialog however Im not too sure about.

There are also the expected updates, the new chrome (browser layout, not Google browser) redesign, which has removed many of the less used parts of the interface as found during a Test Pilot back in March and from developer feedback.

Firefox 4.0 UI concept - May 2010

Something Im really excited about are the developer tools. In particular the console. Beltzner described it as a Quake style console, pulled from the top of the browser, as an advanced view source. With the ability to edit css/dom elements, and make other tweaks on the fly. They will continue to support Firebug and will also add a couple of other api’s to allow us to access rendering times and memory usage from within our apps with should help a lot with development and optimisation.

If you using Firefox or a modern web browser that supports fully open HTML video, you can watch Mike Beltzner presentation.
It is almost an hour long but I do recommend watching or at least listening, to what Mozilla believe is the future of Firefox and the direction they are going.

For more on this story, head over to Mike Beltzner blog post, view the slides and watch the presentation.

First thing the post suggest is to check out Site Performance in Webmaster tools, so off I went. When this first appeared in the Labs section of webmaster tools it was updating pretty regularly and was quite useful and as the graph shows helped me tweak a few things and improve the load time of the server. But it hasn’t been updated in a while now so isn’t very relevant or useful to me.

Next thing it suggests is to install Page Speed for Firebug and test my site. So thats what I did and I got Page Speed Score: 93/100! which is pretty impressive. So I thought next, 7% to go, what can I do to get to 100/100. Nothing it seems… All the suggestions the tool gave me were unsolvable and prevent anyone from getting top marks.

The first suggestion is to Leverage browser caching,
the file it advises me need to do this too is “http://www.google-analytics.com/ga.js” which I cant do anything about.

The next suggestion is to Defer loading of JavaScript, “66.4% of the JavaScript loaded by this page had not been invoked by the time the onload handler completed” is smugly informs me.
- http://gavtaylor.co.uk  24 functions uncalled of 25 total functions
- http://www.google-analytics.com/ga.js 154 functions uncalled of 243 total functions.

On closer examination the functions it reports on my site are actually from Firebug, which must be open for the test to run? catch 22 there!
And the analytic code again is from Google and I cant do anything about that.

The final suggestion is to Use efficient CSS selectors,
http://gavtaylor.co.uk/core/styles/styles.css has 12 very inefficient rules, 27 inefficient rules, and 0 potentially inefficient uses of :hover out of 101 total rules.

This I can do something about and when I get a moment I will be re-coding my CSS file. But this wont improve my results more than 1 or 2 points.

So this unreachable target of 100/100 appears to be impossible. I could remove the Google code from my site which would resolve the Google plug-in issue but I cant run the test without having Firebug open so there will always be 24 out of 25 uncalled functions reported keeping me away from that 100/100 score.

So… will I be sitting in that tree, doing things I should not be?… No, not until they fix the issues with their testing tools anyway…

Update 4/4/11:

On Thursday (31st March), Google released an online version of their page speed tool, This new test works in the same way to the Firebug plugin but uses a version of webkit installed on a server. I have written a quick follow up post with links and my initial thoughts here.

With this in mind, this last week or so I have been dipping my toes into the wonderful world of native Android development.I have gone for the setup advised in the SDK documentation, Eclipse with the ADT plugin. Quickly passing over the Hello World app, as a developer its a task  I must complete before doing anything else, I started on my first application. I decided that for my first attempt I would stick to something I know and rebuild a web based tool I help develop for a well known Manchester server hosting company.

Its been a long time since I last did anything in Java so was a little rusty, but after a few lunchtimes and couple of late nights I have managed to cobble together something that resembles and could just about pass for an Android Application.

So what is it you ask, this wonderful app you have been building up (wont shut up about if you follow me on twitter)? Well I can now tell you, its a speedtest. Not for how fast your mobile connects to the internet, I’ll leave that to the guys at speedtest.net, but to test the speed your website can deliver files to your visitors, now I cant provide you with a link to download and install it as its not really finished and as its not an officially sanctioned app, but I can show a couple of screen-shots from the current alpha version.

This is probably as far as this application goes, unless the boss asks me to continue with it, as although its quite basic and there isn’t all that much to it, it has served its purpose and helped me to figure out how an application should work, its activity life-cycle and how to correctly put one together.

The next step is to think of that award winning, must have,  how did we live without it  idea that will make me millionaire… any suggestions?

I couldn’t remember when I first noticed they were missing but when ever I checked a working folder on my laptop the tortoise svn icon overlays were missing. When I edited a file the red cross would appear as expected, but the green tick to say a file was up to date or the question mark for non-versioned files were missing, meaning I didn’t know if a file was up to date with the repo head or an ignored/non-versioned file.

This morning I came across a post on stack overflow that suggested modifying the registry to over-right the usual Microsoft default setting cock-up. In Windows, for memory reasons Im guessing, they have limited the number of allowed Shell Icon overlays  to 11. Now at first this seems like a logical way to stop memory abuse in Explorer. But for some stupid reason (most likely so their overlays come first), they are actioned in alphabetical order, so any overlays in position 12+ are ignored. With M being in the middle of the alphabet this results in the majority of these slots being allocated by Microsoft Products. In this case Tortoise SVN (obviously starting with a T) was being pushed out of the allowed slots.

I found this out by doing a search in the registry for ‘ShellIconOverlayIdentifiers’  (NOTE: only open regedit if your comfortable using it, we don’t need to change anything just taking a peek so you should be ok). This reviled that most of the Shell Icon slots were being taken up by something called Microsoft Groove. Not having a clue what this is I turned to a popular search engine to find out.

Microsoft Groove is the name for their multi-user document collaboration tool, which makes sense that it would need some icon overlays as it is essentially the same thing as Tortoise SVN. I don’t do any online document collaboration so re-searched for how to remove it. The Microsoft knowledge base article advised to remove the feature from Office via the add/remove programs control panel but when I looked I couldn’t find anything that mentioned Groove but after a further search or two I found that it has now been renamed to Microsoft SharePoint Workspace, which was listed in my version of Office. I disabled the feature and restarted my laptop and I now have my overlay icons back.

Until now, you had to verify you owned every sub-domain on a domain before you could use webmaster tools for it, while for most sites this is fine, you verify the www. version, sometimes there are a couple more that need to be done.
for example on my site, I use the non-www domain, also there is a dev. and a api. that also need their own verification meta tag/html file. If I want to use webmaster tools on another sub-domain I have to verify again… this becomes a tiresome process.

Enter the new DNS verification. You now add a TXT record to your dns and it automatically covers any sub-domains you add.
That’s it… done! There is of course the usual dns propagation you have to worry about so it may not be a quick as adding a meta tag, but it defiantly saves time in the long run.

You can find out more by visiting the Google Web master tools or popping over to the Webmaster Central blog.

IE9 is rumoured to include several <html5> and CSS3 elements, making it a real contender in the new range of “modern browsers”.

You can read more about the release on the IE Blog, or jump straight over to the resources and “test drive” the new release.

I will update this post with my thoughts on the preview, once I’ve had a play with it.