Google Chrome: first looks

I've been playing around with Google Chrome this morning on my work (Windows) machine. This is Google's browser software, and I have to say I'm impressed. I had expected something a fair bit less polished, but instead we've got a near-usable browser already! Something that still hasn't happened with Apple's Safari on Windows. What's more, this will (soon) be available on all platforms.

The new browser is fast, stable, slick and smooth. The chrome-less interface is very impressive, and makes perfect sense when you consider the browser the centre of your computer interface, as Google intends. Memory use is very modest, with each tab separated out into its own process. In total it's using 175 megs of memory, versus 152 megs for Firefox for similar activity. Importantly though, it doesn't seem to bloat out with repeated use.

JavaScript works fine with Google applications, less well with some other sites. It's certainly zippy though! Gmail and Google Analytics zoom along nicely.

Developer tools look interesting so far, with a DOM inspector hidden under the confusingly-named JavaScript console -- there's a console there, but a whole lot more too! It would be hard to have FireBug-level functionality in this first release, but there's a good chunk of it there. I've got some work to do making sure everything works in this new browser, so the development tools (or lack thereof) will either make life easier or (as with IE) very hard.

The only thing I can see preventing me from making this my main browser is the lack of an ad blocker, but I'm sure one will be along shortly. Once it's available for Linux, with an ad blocker, I suspect I'll be making the switch.

Google discovers solvent abuse


We all thought the folks at Google were on crack. Turns out they've been on something quite a bit cheaper. Sometime "tomorrow" (though mentioned as "Tuesday", but probably Mountain View-centric time), Google Chrome will be released for Windows. It's Google's attempt at a browser, as leaked by Google Blogoscoped.

JavaScript process separation

It's well worth looking through the promotional comic about the design process and what they're aiming for. Complete process separation of browser tabs can only be a good thing, though I wonder how much of the web it will break. How will things like the iframe hacks Google themselves use to get around browser scoping remains to be seen. Though breaking them wouldn't be a terrible thing, provided there are other, more secure ways to achieve the same thing.

All in all, it looks very promising. More players in the web browser market can only force more innovation, so more power to them. Of course for those of us trying to make browsers work, it makes life a little more difficult for us. One more test target. Then again, the more browsers the more standards become important.

Windows only at launch, according to the official blog, with other versions following soon after. Hopefully not too far after!

Key signing in Sydney on 12th September, 2008

Geo: -33.85967,151.207025

Udo van den Heuvel is visiting and would like to do a key signing to improve the web of trust. I've suggested The Australian in The Rocks where I tend to take overseas tourists because they have good beer and pizzas topped with both our two national animals, in a good central location.

If you're coming, please bring identification that matches your key and a bunch of little slips of paper with your key details and fingerprint on it, to enable people to verify your identity and key. I'll be following the "informal" method.

Friday 12th September
18:00 for 18:30
key signing, then beer and pizza.
The Australian, 100 Cumberland Street, The Rocks
Probably the back room "ladies" bar. I look roughly like this.

Channel 7 misses the digital point

Channel 7 holds the rights to broadcast the Olympics, but clearly doesn't get the possibilities that digital television brings.

Last night was the first football match, with Australia's mens team playing Serbia. The game started at 19:00 Sydney time, but Seven delayed coverage until 21:00. What's more, they interrupted coverage to cut to the announcement of the flag carrier for the opening ceremony. Then they broke for an ad at the forty minute mark, in other words with only about five minutes to half time.

This kind of thing is completely avoidable with digital television. They could have very easily used spare bandwidth on their transmitter to show the football match live, while keeping their main-channel programming.

The DVB standard, used for digital terrestrial television in Australia and most of the world outside North America, allows for all kinds of dynamic reconfiguration of channels and bandwidth. In Australia the mandate for high definition television reduces the available bandwidth, there are alternatives.

For example, the bandwidth on the high def channel could be reduced during the Olympics. Or Channel 7 could lease spare bandwidth from some of the other broadcasters, perhaps with some revenue sharing -- DVB receivers are very flexible with where the programming comes from.

This kind of setup would be useful throughout the Olympics. There's loads of concurrent stuff going on, so why not show more of it? If the broadcasters want to push the digital switchover, using the technology to its full capacity during such a high-profile event would surely help.

Another alternative would have been to allow SBS, the "complementary" broadcaster for the Olympics, to carry all the football. SBS are the acknowledged home of football in Australia, and we would have ended up with good live coverage and knowledgable presenters. I'm not the only one to suggest this, of course.

As an aside, aren't the Chinese spectators subdued? I've seen more lively funerals than last night's football match!


So the ACCC has launched the new "GroceryWatch" site, GroceryChoice, which gives comparisons of a basket of goods between supermarkets. I wonder how hard it would be to game this system, if you could work out what was in each basket? Dropping the price of a single item could have a big impact.

UK retailers do this with what they call "known value items", such as milk, bread, eggs and bananas. Customers know how much these normally cost, so by dropping the prices on just those items, customers get the illusion of cheap prices and get stung on the prices of goods they can't so easily compare.

The best news from the supermarket inquiry is that the government will bring in mandatory unit pricing reporting. I wrote to the NSW fair trading minister about this a few years ago, with the response that there wasn't demand or need. Unit pricing means the supermarket shelves will tell you the price per standard unit, for example price per 100ml or 100g. Next time you're in a supermarket, compare the prices of 400g and 800g cans of tomatoes. The 800g cans cost more than double the price of a 400g can.

The site itself seems fairly well designed, though the colour scheme isn't ideal. Yellow-on-green isn't really ideal. There's a "latest news" and "subscribe" option but no RSS feeds?

The papers managed to find someone prepared to moan about the site, because he has vision and mobility problems. Sorry Mr Kerr, it's not the web designer's job to show you how to turn on the disability options of your software. They've done everything that they should (though the colour choice isn't helpful) to make it easy for you. Learn where the options for a user-defined stylesheet and minimum font-size are, and use them. Better yet, I bet you're using Internet Exploder. Try Firefox and see the zoom option -- it rocks! (Ctrl + and Ctrl - or Ctrl and the mouse wheel). If you still need more help, there are other tools that will help.

Firefox search in the URL bar

It seemed people appreciated my last Firefox tip about deleting autocomplete entries. Here's another one.

If, like me, Wikipedia is your first point of call for just about any enquiry, here's a quick way to search for what you're after, without faffing about with search sidebars and the like.

What this tip does is allow you to type wp <search term>" in the URL bar and be instantly transported to the result of searching Wikipedia for that particular term. So type wp Eltham Palace and you'll be immediately taken to the page for the art deco masterpiece in South London.

Firefox search bookmark

To do this, create a bookmark with the Location set to Note the %s component, which will be filled with the search term. Put wp into the Keyword field of your bookmark. Now whenever you type wp followed by a space and then your search term, you'll go to the Wikipedia search.

Obviously this can be used for other searches. For example, I search the Internet Movie Database with It's a brilliant way to search the sites you use all the time.

Add a Keyword for this Search in Firefox

Update: Andy Owen points out a much better way of doing this. Right-click on any search box and select Add a Keyword for this Search... which then pops up the bookmark window. Much easier than my method, and something you could explain to mere mortals who don't understand URLs.

XRay for CSS


Very very cool little bookmarklet that shows loads of cool info about the styles of an element you click on a page. Wow!

All CMSes suck

All CMSes suck slide

I have a slide I use when presenting about content management systems to manglement types. Content management systems suck. They're designed that way. The main idea behind them, in a corporate context, is to allow content to be published quickly by people who don't know (much) HTML so that it fits within the existing structure. The trade-off is that it's a whole lot less flexible. Anything you want to do that doesn't fit into the built-in templates and structure is going to require some development.

The reason I use this slide is to explain that there's a middle ground in there. If you're an enormous publishing organisation with daily deadlines and a team of editors, sub-editors and content approvers, then the top-end systems are necessary. For most corporate web sites though, somewhere in the middle makes more sense.

When you demonstrate the idea of approvers and workflow that's built into most CMSes, manglement types start getting ideas. They love the idea of a defined workflow, with things getting signed off as they go through the workflow. Yet I've never worked in an organisation where the workflow features work. You always end up with one or more approvers who constantly forget their logins, forget what they're supposed to do to approve, or do the wrong thing when they're in there. It always ends up reverting to a paper-based system, printing out the content for approval. In most situations, it's a better approach. The workflow systems in most CMSes get turned off pretty quickly in most implementations.

I'm about to start specifying a new content management platform for my employer, so it's got me thinking about the complexity level we're going for. Parts of our site are quite tricky, technical bits of code and integration with other systems. But most of it is just plain content. Our current CMS is a little too far to the right on the axis I've drawn above, with seemingly-simple changes quite hard to make, involving digging through templates, schema definitions and template elements. I need to find a middle ground for the next crack at it.

Looking around at the free software content management systems, it's pretty difficult to compare them in any meaningful way. And more important to me is the company we end up using to build the thing. I'd really rather not have a PHP/MySQL platform, but two off the seemingly best-supported CMSes (Joomla and Drupal) are just that. I love the look approach taken by the Perl+XUL cyclone3, but the only companies supporting it seem to be in Europe.

So, lazyweb, can you recommend any good CMS development companies who are interested in a pretty big project? I'd like a small, nimble, free software-loving group who are well-versed in Agile development. Sydney ideally, but elsewhere in Australia is okay. Let me know.


I've been thinking over the weekend about my post about the iPhone and the crap plans being offered. The problem is the lack of data covered by these plans. I reckon the stories in about a week's time will be all about the $2,000 bills iPhone users have been copping due to going over the data caps and getting the punitive excess data charges.

I poo-pooed Mark and Stilgherrian's idea of forming an MVNO, a virtual mobile phone company. I still think that's way harder than they think, but I have an alternative idea.

How about if you signed up for a mobile data plan and instead use VOIP for the phone part of it? You could use an online service for cheap outbound texting, and inbound texting could come to the mobile SIM's number. You'd have a landline number so people calling you would find it much cheaper. Though you might need a different number for texting.

One potential issue: what's the bet the mobile providers shape down VOIP so it's unusable?


Mark Pesce and Stilgherrian are talking about starting their own telco. More power to 'em, I say, but I suspect they're in for a rude shock. Running a telco isn't as easy as it looks. The big telcos make it look harder than it is through beurocracy, but it's still hard.

They talk about creating an MVNO -- a white-label mobile operator using someone else's network. The problem is that the deals you get as an MVNO without much volume just aren't very good, which is why there aren't any out there offering particularly good plans. I use an MVNO myself (and work for another) run by an ISP who I don't use, becausde their plans just happen to match my usage profile: not many calls, a moderate amount of texts, and lots of short calls and texts between me and my partner, which are free. I can guarantee you Exetel aren't making a lot out of me.

The basic problem with the MVNOs is that the wholesaling operator doesn't want to canibalise their high-profit areas. They'll structure the deal so that you can't offer deals attractive to their most profitable customers.

Mark is annoyed that all the iPhone plans come with piddling amounts of data, and punitive over-quote data rates. But if you get one of their "data" deals, where you plug a USB thingy into your laptop you can get mobile data at much better rates. I'll explain why they do this below, but the big secret they don't want you to know is that those things are just mobile phones. I've heard that with some providers you can actually stick them in a phone and use them, even for voice.

Virgin also offers a "fixed" broadband wireless service over the Optus network with pretty generous voice and data plans. Quite an attractive deal, except for the problems that are being reported with it. You can't use it in a mobile context though, which would massively help them in terms of planning their network upgrades.

So why do telcos have different data prices for handsets and data cards? The reasons are segmentation and mobility. With data cards, they can sell you another service on top of your mobile plan -- you'll notice the most attractive deals require you to also have a mobile with them. Another of the secrets they don't want you to know is that you can use your 3G mobile for internet access on your computer via Bluetooth. It works just fine, except for the punitive data prices on your mobile SIM.

The other issue is mobility. Despite the fact that with a mobile data card, you can travel down the highway using it, most use it in a familiar context. Sitting down somewhere and not moving. This makes it quite a bit easier to plan for, whereas if everyone was whizzing around from cell to cell using prodigious amounts of data, it'd get quite hard for them.

You have to remember, this is an industry used to making 25c out of something that costs them, effectively, nothing. SMS uses spare capacity on the network yet makes them a fortune. They see data as the next cash cow. If they started offering phone+data plans, they think they'll lose all the people using those data cards switching to just using their phone.

Another part of the problem is that users want a "free" or subsidised phone. Where do you think they claw back that money? So the real solution for you would be to buy the phone outright, and sign up for one of the more attractive non-iPhone plans. Possibly even with 3, assuming you're not going to be using data outside the capital cities.

PS: Mark and Stilgherrian, get in touch when you've realised that you can't be a telco without a billing system and customer service setup. And that both such things are very hard. Then I'll clue you into some of the other hard parts. But good luck. If you make it happen, I'll be one of your first customers.