Wow, doesn't online advertising suck?

I haven't seen online ads since about 1995, when HotWired first pioneered annoying, animated banners that detracted from the content. Initially I was using hacks like hosts files and the like, then I started using Craig's Squid-based blocker until these days I use a combination of Adblock Plus, Filterset.G and some Greasemonkey scripts to block some of the more annoying inline ads. Any time an ad gets around these filters, I quickly squish it because I've got used to viewing content without distracting, annoying crap flashing near what I'm trying to read.

Now I'm on the other side of the fence, and we're advertising through some of these ad networks and I've had to test view some of it. First of all, I'm staggered at how intrusive they've become. On some sites like NineMSN, they overlay animations over the actual content! Unbelievable and I'm amazed anyone uses sites that do that kind of thing. Mainstream sites still even use pop-ups and pop-under crap?!?!? I can't believe audiences don't go out of their way to work out how to block all this.

It's quite remarkable how useless the ads are in delivering a valuable audience. I guess it's not particularly surprising. If you're reading about the cyclone relief effort in Burma, you're hardly going to click on an advert for, say, a car. It's a drive-by impression, with similar brand-building impact to a billboard or a display ad in a newspaper.

What's more, the publishers don't help themselves by even further devaluing their impressions. Fairfax sites reload every five minutes, so if someone walks away from their computer, it happily counts impressions with nobody home to see it. That makes me very suspicious of their impression counts. I can't believe the publishers could be so stupid. Google changed the nature of this business by delivering benefits for everyone: advertisers and advertisees, which is why they're make kongbucks. Seems the dead tree purveyors, and those apeing them, haven't worked that out.

So apart from a very cheap (and it would want to be, given the publishers' efforts to devalue the product) brand building exercise, I don't really see the point of display advertising on content sites. It's generally not that relevant, not particularly targetted, and doesn't deliver.

Search engine marketing, however, is gold. I don't filter out ads next to Google searches because they're often actually what I want! If you type in "Marrickville plumber", the ads actually show better results for local plumbers than the search engine results. Everyone wins! If someone clicks on your ad, they're actually very likely to be wanting what you're selling. But it seems the publishers don't understand that to be successful, they need to find a way to make their advertisers successful.

Finally, we're seeing some rather nasty behaviour. The ad network we're using is reporting about 30% more clickthroughs (as recorded by the publishers) than we're seeing actually arrive at the landing page. They claim this is a fairly normal rate. What's going on here? Click fraud?

Disable Same Origin Policy: JavaScript

The Same origin policy ensures that your browser won't load files from sources other than the original source of the loading file. It's vital to help prevent cross-site scripting attacks, but a bit annoying when you're working to draft AJAX responses.

One way around this is to run a web server on your local machine. There's lots of circumstances where this isn't ideal though, such as on an underpowered Windows piece of crap like I use at work.

Fortunately, Firefox has an option that helps get around this. I didn't find it when searching for "Disable same origin policy", so hopefully this blog post will help future seekers.

The Firefox option security.fileuri.strict_origin_policy allows local files to bypass the same origin policy. Yay!

Goodbye eBay

I've used eBay for a very long time. Since it was first launched in Australia, I think. They had got the balance between buyers and sellers right. Recent changes of policy have forced me to think about whether or not I'll use it in future. I've decided that I won't.

They now require that you offer Paypal as a payment option, and will soon make it the only payment option. Paypal is not a bank, so your money is not guaranteed against failure of the institution. Paypal's dispute resolution procedures are opaque and poorly-implemented. Getting Paypal to connect with your bank account is a tedious, multi-day process that often fails. I refuse to use Paypal.

The recent changes to the feedback system also, I believe, skew things in favour of sellers too far.

Online auctions are really useful ways to buy and sell stuff, and you can often pick up real bargains. Conversely, when selling some people will end up paying RRP or more for used goods, which makes it great if you're selling. Now to find an alternative venue.

Any suggestions? I've found OZtion a bit crap.

Tools to stich photos

Map
of existing land use in the County of Cumberland 1945

A very helpful man from the Critical Mass list took a bunch of photos of a map I've been hunting for. The map is Map of the County of Cumberland showing Interim Plan No.2, 1947 from the NSW State Records. I want this map because it demonstrate's how the RTA (then DMR) has been hell-bent on implementing this ancient plan, which was conceived as Sydney's answer to LA's freeway network, before LA realised you can't build your way out of traffic congestion, which is something the RTA hasn't realised yet either.

So he's taken a big swag of photos with a digital camera of this very large canvas-backed map. Of course all the photos are at different angles, with different light and with differing perspectives. I need to stitch these together.

Hunting around online, most of the tools seem to be designed to stitch together linear panoramas, taken from the same vantage point and with very similar perspectives. Not useful for this job.

So, lazyweb, any ideas on the best approach?

PS: The photos are already in the public domain, and my stitched-together version will also be.

Dead pixel

My LCD monitor has a dead pixel. I've tried running some stuck pixel tools for a few hours at a time in an effort to fix it, without success. Next step is to try the more drastic methods. Unfortunately it's right in my field of vision, so can be quite annoying.

IceTV judgement's impact on other data?

In case you hadn't heard, IceTV lost their case in the Federal Court. IceTV provide(d) an electronic programme guide (EPG) service commonly used by people with personal video recorders (PVR) that work like Tivo. In other words, you can say to your recorder "record The Simpsons whenever it is on" and the PVR, knowing the programme guide, can just schedule those recordings. Any time you want to watch some Simpsons, there'll be loads sitting there for you.

Fortunately for PVR users, there's alternative, though I suppose it's legally-dubious in that it uses screen scraping and the like to replicate the EPG. It works brilliantly though.

To my non-lawyerly eyes, this judgement seems rather far-ranging. It seems that basically any collection of data can now be covered by copyright, no matter how you recreate it. The court viewed the act of scheduling programmes on a television channel as a creative act. Having worked extensively with schedulers in a past life, I'd tend to agree, it is something these people agonise over. But I think the impact is going to be quite widespread.

For example, let's say I wanted to set up a web site that allows people to compare phone plans. I use publicly-available information about the various suppliers' pricing to build a database that is looked up to recommend a specific plan from a specific provider. This would now seem to be a breach of copyright. I work with people who design phone plans, and I can tell you it's a very creative process -- though perhaps not for the right reasons.

So this judgement seems to have impacts far beyond the commercial channels' obsession with preventing people form skipping their shitty, all-to-regular ads. It could, in fact, prevent efficient markets as in my example (though telcos go to great lengths there anyway). Is this a desirable outcome?

Okay, now I miss London

Dave Cross just posted a link to the Programme for Opentech 2008 in July in London. While I've long missed all the awesome bands I've been missing in London, this is one of those events I really would love to go to.

There's really cool stuff like the stuff the mySociety guys going on over there. Hackers hacking government to make democracy work better.

I really enjoyed some of the random conferences I went to in London.

London Perl Workshop had really inspirational talks on people doing seriously awesome stuff in Perl.

Sadly defunct NTK put on a really impressive day called XCOM2002 showcasing people doing weird and cool shit with computers.

I'd love to start something like this here in Sydney. London, at the time, had NTK, which gave enormous amounts of publicity to really interesting geek stuff going on, which helped in getting the word out. Not sure there's the critical mass here, but there are certainly loads of people doing really interesting things.

Not sure how to get stuff going. I've certainly never made it to a Dorkbot here, so I'm as bad as everyone else.

Perhaps we need an NTK for Sydney? Who's up for starting one? I'd only be interested if there's some helpers. For geeks who didn't live in London between 1997 and 2006 (i.e., most of you) NTK (Need To Know) was a sarcastic weekly, purposely low-tech newsletter about what was happening in and around London. By and for geeks, particularly the charismatic Dave Green and Danny O'Brien. Wikipedia article.

Firefox 3 looking good

I thought I'd try out Firefox 3 for a bit. Part of my job is to stay up to date with the latest and greatest, and I was hoping the much-touted memory management enhancements would be a nice plus.

So far, I'm well impressed. AJAX- and JavaScript-heavy sites are vastly faster, so for example Gmail snaps open, Google Reader zips along. The CMS I use all day every day also flies. Most importantly, they really do seem to have plugged the memory leaks. By this time of day, I'd expect Firefox to be around 200 megs, having used a few AJAX sites quite heavily. Instead, it's around 115 megs with four tabs and two CMS windows open. It also seems to go down when you close tabs and windows, which is something that didn't happen before. It also remains quite zippy.

Haven't noticed any bugs or rendering weirdness yet, which is a good sign. Only problem so far is that Firebug isn't yet available. There's a version of it for Firefox 3, but it apparently has some issues. If I decide Firefox 3 is stable enough to use all the time, I'll try out the upgraded Firebug. Life without Firebug would be a much reduced life...

Inbox zero: the radical approach

There's been a bit of discussion recently about Inbox Zero, getting your email inbox empty. Well I just discovered a rather radical approach to getting there. I accidentally deleted everything in the inbox. Thought I was in another folder and deleted all.

I've got backups, so if I find there's anything vital I'll be able to recover, but for now it's strangely liberating. Will see how I go.

How much should we prop up the remote regions?

Australia has a long history of propping up the economies of the remote regions. You get tax breaks for living in remote areas, telecommunications providers are forced to provide phone services and we subsidise connecting them to the Internet. Australians have an emotional connection to the bush, despite the fact that most of us don't live there and a large proportion have never even been anywhere particularly remote.

I've been thinking about this recently after the new government's tender for a Fibre To The Node network stipulates that 98 percent of the population should have 12 megabit broadband. 12 megs is a pretty fat pipe, and you've got to wonder why they set the bar there -- was it based on any cost/benefit analysis or did they just pluck a number from the air? And where does 98% come from? What kind of density are those last few percent spread over?

What I've been considering is how much it makes economic sense to prop up all these places, and does it distort rational market incentives? For example, Australia has enormous farms in quite marginal land. Perhaps the best outcome would be to not farm these areas, given their fragility and the marginality of the business?

As another example, I once visited a family friend's enormous farm in South-West Queensland. Very dry, dusty land with a couple of sheep per square kilometre. The farmer had a main house on one side of the property and another small house on the other side of the property that was used only a couple of weeks a year. At the time, Telecom was forced to supply phone service to these remote areas at a maximum cost of $5,000 for installation. The microwave links to provision these services would have cost quite a few orders of magnitude more to supply. Because of this perverse incentive, the farmer had a phone installed at the little-used house, where previously they'd kept in contact with the main house via HF radio. So the taxpayer bore the cost of installing all this, when in reality it wasn't particularly needed.

The same idea comes in with this broadband subsidy. Now I'm sure every farmer out there would love to have metro-equivalent broadband, the ability to watch YouTube and the like. But do they really need it? I'm struggling to think of any business-essential application that would require the bandwidth and low-latency of broadband.

Telemetry from sensors would use very minimal bandwidth, and isn't all that latency-prone, so would be much more cost-effectively carried by other means. Access to weather reports and commodity prices isn't exactly high-bandwidth stuff: a 56k modem would do fine. Even having webcams strewn around a property so the farmer can keep an eye on things when away isn't going to use enormous bandwidth.

So what pressing issue requires a subsidy for this? It seems rather extravagent. The political sums don't add up either. These regions are almost all National voters, and would never elect Labor (despite the agrarian socialist Nats having more in common with the ALP than the reactionary, conservative Libs).

I've long argued that we should stop subsidising farming anything like as lavishly as we do. Farmers are business people like all others, so why do they qualify for such extravagent treatment, apart from some weird sentimentality?