The BBC’s headline ticker is currently running with:

Experts refuse to rule out long-term mobile phone use causing cancer.

Damn those experts! They simply refuse to make long-term health predictions on recently-developed technology. What is wrong with these people!?

The article itself concentrates on the long-term health worries, all of which are entirely speculative, and is heavy on lung-cancer/smoking comparisons:

He said: “We can’t rule out the possibility at this stage that cancer could appear in a few years’ time.

“With smoking there was no link of any lung cancer until after ten years.”

He said the problem during the study was that there had been very few people using mobile phones for over ten years.

Cancers do not normally appear until ten to 15 years after exposure.

The last sentence is weird. Exposure to what? Radioactive materials? Do you think the reporter is confusing different types of radiation?

And this is with a decent study that’s pretty conclusive in its analysis that using mobile phones for ten years isn’t dangerous, a fact you’d think might be newsworthy.

Science reporting by non-science-reporters always tends towards ’scientists don’t know anything’. If it’s a health study that shows no effect, it’s a tentative conclusion. If it does show an effect, it’s an obvious common-sense result. If it’s new evidence that contradicts previous research, it’s impossible to know what to believe. What do you mean, you can’t win?

The Monsterpod is a camera mount with a base of weird orange goop that can stick rigidly to most non-fabric surfaces. I was given one for my birthday, but it’s been sitting in a corner for a couple of months as my DSLR is unfortunately too heavy for it. I realised today that my new flash is not, so figured I’d give it a try:

I’m sure this will come in handy at some point Sticking it to the door and running across the room to grab my camera was one scary, scary moment, mind.

Attached to the base is an el-cheapo Gadget Infinity wireless trigger V2, which arrived this morning. Initial indications are that it works very well. I’m hoping to have more time to play with it over the next few days; I want to write a proper review, as the 430ex flashgun caused particular problems with the V1s and plenty have been asking about the new version.

I loves me the T-Rex, but sometimes he’s just plain wrong.

It took a while to surface, but over the weekend I finally spotted the obvious possibilities raised by Google Spreadsheet’s new xml-linking features. Abi and I currently enter isbn, title and author data into our bookselling spreadsheet by hand, and this gets old quickly. More so for Abi, who has to check the individual Amazon pages to gauge current prices. I knew of Amazon’s API, and was planning to create something to help speed this up, but the only programming language I was ever any good at was Visual Basic, which isn’t around much any more, and I’m seriously rusty at the little .NET I once knew. So I wasn’t quite sure of the best way to approach things, and GS’ new hotness seemed like an ideal solution, if I could get it to work. Happily it did, and I can now enter an ISBN and have GS download the author, title, amazon URL, amazon price and lowest used price data automatically. Here’s a basic way to set it up:

1. If you haven’t got one, sign up for an Amazon Web Services account and find your Access Key. This’ll let you make 1 request per second, although in practice this is fairly flexible.
2. Take a look at the Amazon ECS API, and determine the appropriate request. It’s a web service, so supplies data after a standard browser request. Here’s the query I’m using, with an example book:
   

   http://webservices.amazon.co.ukonca/xml?Service=AWSECommerceService &SubscriptionId=[your_ID] &ResponseGroup=Medium &SearchIndex=Books &Operation=ItemLookup &IdType=ISBN &ItemId=0593055489

   The most important part of this is the ResponseGroup, which determines how much data Amazon supplies. Medium is a good all-rounder, but there are plenty of more specific request types. Specifying ISBN as the IdType allows the use of 13-digit ISBNs, which a standard ItemLookup doesn’t - Amazon’s documentation claims it doesn’t work on the UK site, but it does.

3. Test out the request in your browser window, and make sure it’s supplying all the necessary data.
4. Now it’s time to figure out the XPath statements you’ll need to extract the relevant info. If, like me, you haven’t the foggiest idea what XPath means, don’t worry. It’s pretty easy, and the W3Schools tutorial takes 10minutes to explain everything you’ll need.
5. In this case, I want to extract the Amazon price, the lowest used price, the URL of the detail page, the (first) author and the title:
   • The price I want is in the FormattedPrice element, underneath the ListPrice node. There’s more than one ‘FormattedPrice’ element in the results, so I need to specify its parent node. The appropriate XPath statement is “//ListPrice/FormattedPrice”
   • The lowest used price is the same element, but under LowestUsedPrice, so: “//LowestUsedPrice/FormattedPrice”
   • The URL is in the “DetailPageURL” element, and there’s just the one of these, so: “//DetailPageURL”
   • Similarly with title: “//Title”
   • There may be multiple authors, and grabbing all of them would mess up the spreadsheet formatting we’ll come to in a minute. So I’ll just grab the first one: “//Author[1]“. The number should, I think, be a zero to adhere to the standards. But, it isn’t - zero doesn’t actually work. Go figure.

   These can all be combined into one XPath statement with | operators, so the final result is:

   “//ListPrice/FormattedPrice | //LowestUsedPrice/FormattedPrice | //DetailPageURL | //Author[1] | //Title”

6. Ok, fire up a spreadsheet. Create a column for ISBN, price, lowest price, url, author, title and finally another for processing.
7. GS uses the ‘importXML’ function to import data. This takes two arguments - the URL and the XPath statement. Enter an example ISBN, then, in the ‘processing’ column:
   

   =importXML(”http://webservices.amazon.co.uk/onca/xml?Service=AWSECommerceService&SubscriptionId=[your_ID]&ResponseGroup=Medium&SearchIndex=Books&Operation=ItemLookup&IdType=ISBN&ItemId=”&B2,”//ListPrice/FormattedPrice | //LowestUsedPrice/FormattedPrice | //DetailPageURL | //Author[1] | //Title”))

   The &B2 at the end of the URL appends the ISBN number, so be sure to change this to the appropriate cell.

8. Google should now go fetch the XML file, process it according to your XPath statement and…dump all the data into the ‘processing’ column. If you click on the cells containing the data you’ll see they have ‘Continue’ formulae. This (currently undocumented) formula simply takes data from the array stored in the first cell. So “=CONTINUE(H2, 5, 1)” shows the data from the 1st child (if it exists) of the 5th item of the results of the XPath processing - the book Title. Copy and past these statements to the appropriate cells on the correct row.
9. We need to stop Google filling the subsequent rows, so enter a random text string in the second row of the processing column, then enter a different ISBN into the first row to refresh the query. Google will again fetch the data, then ask if you want to overwrite existing data. Say no, and it’ll only populate the Continue formulae we just set up on the correct row.
10. CTRL-D the formulae down a few rows, and voila, you’ve got an easy-to-use ISBN query spreadsheet that uses one query per row. If you want to tidy it up, wrap it in an if statement like:
    

    =if(B2=”",”",importXML(”[your_URL]“&B2,”[your_XPath]“))

    isblank() doesn’t seem to work properly with cells that have had content deleted.

Problems with this method:

• Queries are currently limited to 50. And that’s 50 queries present, rather than active - hiding them behind an if statement based on the contents of another cell doesn’t work. This is easy to get around via copy-and-pasting, but could be annoying in some circumstances.
• Queries update every two hours. This is unnecessary for this purpose, and I’d personally like to see unlimited one-time only queries that don’t update every 2 hours; hopefully that’ll come.
• Amazon theoretically have a 1 query-per-second limit. I’ve actually dumped 25 ISBNs at once and had results appear instantly, so I think they must have a flexible policy, but it’s a little risky and I don’t want my account suspended. This is another reason I’d like a one-time-only query option. It’s probably wise to play it safe by cutting-and-pasting the data into another table before closing the sheet.
• If Amazon doesn’t have a particular datum, the XPath query will be in the wrong order. For example, some annuals and readers digest books have no author listed, so the ‘Continue’ statement that should refer to the 4th field (in this case, the author) instead refers to the title. I can’t think of a way around this without multiple queries.

Hopefully I’ll be able to refine this in the future, but it’s not a bad start, and I’m very impressed that the functionality exists at all.

Google Spreadsheets now has an autofill box, meaning I can drag a cell to automatically extend a series. So if I enter ‘Monday’ and ‘Tuesday’ in consecutive boxes, then highlight and drag the node, it’ll fill in the rest of the days of the week. This is a useful feature in offline applications, and will undoubtedly come in handy online, but there’s a hidden payload: press control while dragging and it’ll look for correlations in Google Sets. I entered ‘monkeys’ and ‘cats’, and it extended the third box with ‘iguanas’. ‘Han’ and ‘Leia’ produced ‘Luke Skywalker’, ‘C3PO’ and ‘R2D2′. Now that’s pretty cool.

GS also added support for external data. Online XML, HTML, RSS feeds, and .csv files can all be referenced, with imported data updated every few hours. That’s extremely powerful, especially when you combine it with the ability to publish - I can now link directly to the data in another user’s Google Spreadsheet. Since the code’s already there, I imagine this’ll be extended to Google Docs pretty soon.

Google Reader now finally has a search box, and also maxes out counts at 1000 rather than displaying ‘100+ items unread’. That’s pretty much all the missing functionality I wanted.

Finally, Gmail and ‘Picasa Web Albums’ users can now buy more storage space, with 6gb costing $20/year. 250gb is $500/year. Clearly, nobody needs 250gb for email and pictures…Google drive, anyone?

The BBC’s head of tv news, Peter Horrocks, last week wrote this on his blog:

BBC News certainly does not have a line on climate change, however the weight of our coverage reflects the fact that there is an increasingly strong (although not overwhelming) weight of scientific opinion in favour of the proposition that climate change is happening and is being largely caused by man.

This is good stuff. The media generally fails spectacularly at science coverage because the usually-reasonable journalistic standard of ‘fairness’ requires them to present an opposing viewpoint. In the 90’s the whole of the medical industry was screaming that MMR was safe, yet every reporter felt it necessary to interview one of the very few crazies on the basis that it’s a balanced view, often followed by ‘viewers get to decide’. Unfortunately, this decision is the necessarily based on a misrepresentation of the facts.

It’s the same today with global warming: the vast majority of climatologists think it extremely likely that a) global warming is happening (actually, nobody doubts this) and b) it is very likely that man is causing it, yet the deniers get just as much coverage, if not more. The problem is that, unlike politics, a scientific consensus has genuine authority.

Because the process of science is so ruthless - the job of your colleagues is to destroy your arguments - and disparate - hundreds of countries with thousands of independent organisations and millions of scientists from every part of the political spectrum - a consensus of opinion is genuinely valuable, and isn’t suddenly turned into a 50/50 probability when a lone-hero / complete whackjob (take your pick) starts claiming everybody else is wrong. The scientific method will assess the validity of their arguments, because that’s the only authority with the expertise to do so. The results cannot be published by any ultimate authority, because science has no such authority¹ but must be gleamed from consensus opinion. It may even turn out to be wrong in the long run, but it’s the only way that can possibly work. No journalist can accurately assess the merits of a scientific claim, given their lack of time and expertise, so the only sensible approach is to report in a way correlated with the scientific opinion.

To hear exactly such a conclusion coming from a head of BBC news was very encouraging. Then came this:

The BBC has scrapped plans for Planet Relief, a TV special on climate change.

The decision comes after executives said it was not the BBC’s job to lead opinion on climate change.

(…) But against the backdrop of intense internal debates about impartiality, senior news editors expressed misgivings that Planet Relief was too “campaigning” in nature and would have left the Corporation open to the charge of bias.

“It is absolutely not the BBC’s job to save the planet,” warned Newsnight editor Peter Barron at the Edinburgh Festival last month.

Head of TV news Peter Horrocks, writing in the BBC News website’s editors’ blog, commented: “It is not the BBC’s job to lead opinion or proselytise on this or any other subject.”

The BBC clearly feel happy to present the opinions of climate-change activists in a large way - Live Earth shows this - and to balance their news output according to scientific opinion, but are uncomfortable with organising anything themselves. This almost seems reasonable, but how does it fit with Comic Relief? There are plenty of conservatives who might argue that the suffering of children in other countries is nothing to do with us Brits - how dare the BBC ‘proselytise’? Of course, most people consider this morally unambiguous - of course the BBC should do everything it can to help people who are suffering.

But what’s the difference between campaigning against African suffering, and campaigning against a climate change that will cause similar suffering in the future? Is it the immediate visuals? I doubt it. I think it’s more likely what’s alluded to in the above - the BBC would be left open to a charge of bias. Because climate change is so politicised, and because much of the country thinks, wrongly, that there’s some major scientific debate as to whether it’s man-made, the standing of the BBC probably would suffer if it were to take an active position. It’s not an easy position for them.

I must point out that the BBC have said:

Our audiences tell us they are most receptive to documentary or factual style programming as a means of learning about the issues surrounding this subject, and as part of this learning we have made the decision not to proceed with the Planet Relief event.

Instead we will focus our energies on a range of factual programmes on the important and complex subject of climate change. This decision was not made in light of the recent debate around impartiality.

Which isn’t unreasonable. It’s certainly better than Channel 4’s outright promotion of global warming deniers.

I have no problem with an activist BBC, when it comes to scientific issues. Their news departments may not want to ‘lead opinion or proselytise’, but, providing it’s done according to evidence, I don’t see why the BBC shouldn’t lead the way. They have a huge amount of influence, and even, possibly, a moral duty.

Rather than a Comic Relief-style show, how about an evening of detailed analysis? The BBC have a huge expertise when it comes to presenting knowledge in an accessible way - why not put this into explaining, as clearly as possible, why the climatologists are correct? You could even have a section explaining why the deniers are wrong. I guess people might not watch, but I suspect there are many people with the interest but without the time or knowledge to do any research themselves. Which is perfectly understandable. I wonder whether it could work.

1. enormous UN studies nonwithstanding [↩]

I’m sympathetic to the idea of a compulsory DNA database, but willing to be swayed. Neil has an interesting post on the problems of the current system, so I’ll stick to the theoretical arguments. Here’s how it seems to me:

Advantages:

Crime. A national database would obviously help a massive amount in solving crimes.

This is easily enough of an advantage for me to take the idea seriously.

Disadvantages:

Practical - how do you actually get DNA samples from the entire population?

Technological - how do you create a database that big, and keep it secure?

Abuse - what if Future Evil Government want to use it for Evil Purposes? What if Not So Evil Government agree to let insurance companies access the data?

I think the practical and technological problems could, in theory, be solved. For example, if an open-source system were used it’d be in everyone’s best interests to make it as secure as possible. This might take some doing convincing - governments have a habit of awarding billion-pound contracts to IT firms who, unsurprisingly, make everything proprietary - but, still, it’s possible.

Liberty, interestingly, don’t really go with the ‘freedom’ angle I expected. They’re much more concerned about abuse, and this certainly seems to be the major issue. I’m personally happy for the police to have access to my DNA profile - I don’t see what harm this could actually do - but I’m not so sure about insurance companies / others. I can’t think of a way to ensure this doesn’t happen. But, then, aren’t there many laws like this? Evil Government would just do what they wanted anyway, wouldn’t they?

What else is there?

Errors - what if a mistake means the wrong person is arrested?

I’m not particularly worried about this. Re-testing of the individual involved should prove their innocence, and, again in theory, it’s entirely possible that such errors could be kept to a minimum.

Other than the abuse problem, it’s easy to envisage a system that doesn’t suffer from these flaws. But I get the feeling that even if all the problems were solved, Liberty etc. wouldn’t like it. And that’s what interests me. What’s wrong with the Perfect Theoretical Database? I imagine it would involve some mention of ‘civil liberties’, and I want to understand exactly what this argument involves. The BBC feedback site is full of raving…ravers…and there’s little enlightenment to be found. Is it privacy? Because I don’t see why anybody would object to a system in which only the police can access your details, if your DNA turns up a match. What exactly is it that people don’t like? Being forced to do something?

Having said all this, I suspect people are going to have to adapt to their DNA profile being ‘out there’. Once technology reaches the point where my DNA can be scanned to detect increased risks of future medical problems, I’m in. If this means my profile has to be stored in computerised medical records, where any dirty hacker could potentially snag it, so be it. And what if insurance companies start offering free DNA scans for potential heart problems, say, on the basis that you allow your policy to be adapted accordingly? It’s going to come, it’s just a question of how it’s managed. Do I trust private databases more than a state-run system? I doubt future generations are going to give two hoots about their ‘private’ information being accessible to the world - it’s beneficial far more often than otherwise, if you ask me - but the transitory period could get tricky.

Yesterday I had no immediate plans to fly on Nepal Airlines. Today I have no plans to fly with them, ever:

Nepal’s state-run airline has confirmed that it sacrificed two goats to appease a Hindu god, following technical problems with one of its aircraft.

What, no chickens? The radio news ran this as an amusing ‘and finally’, but it’s not particularly funny for a) the goats b) the passengers.

This morning I rang my bank to find out how my mortgage will change once the fixed-rate term expires. Turns out there’s an 80% increase. Holy crap. So, today rapidly became about finding a better deal.

moneysavingexpert.com is my first destination for anything financial, and turned out to have a downloadable remortgaging guide. It recommends that everybody but the most financially savvy use a broker, so I gave London & Country a call. They apparently make their money in referral fees from the banks/building societies involved, so it seems less risky than phoning a random person from the phone book. They’re also able to link into my current bank’s systems to view offers only available to existing customers etc.. Hopefully they’re ok. They’re currently off deal-hunting and have promised to get back to me today/tomorrow.

Obviously fees are going to go up - my fixed rate was 4.3% and the current BoE rate is 5.75% - but hopefully it won’t be too insane. Might have to try and speed up the job applications process, though…

Or, at least, I found out about it this weekend.
Whenever we sell a book on Amazon I receive a ’sold, dispatch now’ email, or if there’s any problem with payment I get a ’sale pending’ email telling me to reserve the book until payment goes through / three days pass. Three weeks ago I received a dispatch¹ email, quickly followed by a sale pending message. So I reserved the book, but didn’t hear anything so forgot about it. I hadn’t realised quite how it works: ’sold’ emails are definitive, and only arrive once payment has actually gone through. I’d probably have figured it out if the Amazon emails hadn’t arrived in the wrong order - for whatever reason there was a delay on the initial ‘pending’ email. Not their fault, just unfortunate timing.

So on Saturday I had an email asking why the book hadn’t arrived, which confused me for a couple of minutes. Once I realised what had happened I quickly posted the book and, after double-checking with Abi, completely refunded the guy. I’m nervous every time I check our feedback, now. Obviously it’d be reasonable to leave negative feedback, but we’re currently at 100% positive and one problem would drop us to ~97%, or over 1 in every 50 orders. Not good, when you’re selling online.

I knew the spreadsheet was out by a few pounds, so would probably have picked it up once I properly balanced the tables. Shame I didn’t spot it sooner, but at least I won’t make that mistake again.

1. despatch? [↩]

A recent eSkeptic reviewed ‘How Doctors Think’ and, amongst some fairly strong criticisms of the book, mentions ‘decision aids’ - algorithm-based methods of diagnosis:

Most doctors do not like decision aids. They rob them of much of their power and prestige. Why go through medical school and accrue a six-figure debt if you’re simply going to use a computer to make diagnoses? One study famously showed that a successful predictive instrument for acute ischemic heart disease (which reduced the false positive rate from 71% to 0) was, after its use in randomized trials, all but discarded by doctors (only 2.8% of the sample continued to use it).⁵ It is no secret many doctors despise evidence-based medicine. It is impersonal “cookbook medicine.” It is “dehumanizing,” treating people like statistics. Patients do not like it either. They think less of doctors’ abilities who rely on such aids.⁶

The problem is that it is usually in patients’ best interest to be treated like a “statistic.” Doctors cannot outperform mechanical diagnoses because their own diagnoses are inconsistent. An algorithm guarantees the same input results in the same output, and whether one likes this or not, this maximizes accuracy. If the exact same information results in variable and individual output, error will increase. However, the psychological baggage associated with the use of statistics in medicine (doctors’ pride and patients’ insistence on “certainty”) makes this a difficult issue to overcome.

I can see how that would suck, from a doctor’s perspective. But that’s quite the statistic on the heart disease prediction. Maybe this is the way of the future: statistical techniques that apply to the majority, with human oversight to spot the minority cases.

Having said that, I was hearing this evening how current advances in the understanding of disease at the genetic level are going to lead to much more personalised medicine in the next 10-20 years. Which is cool. It was Francis Collins who said it, mind, and his reasoning skills elsewhere in the interview were somewhat suspect, but he nevertheless knows more about it than me

This is lovely:

I’m sure it’s far harder than it looks.

I just received a card made from sheep poo.

On the back it says “sheep droppings are sterilised and washed, then turned into paper”. If, like me, you’re wondering the washing and sterilising doesn’t result in a lack of poo, it’s because “a sheep only digests 50% of everything it eats. The rest is useable fibre.” So now I know that.

Made my day. Thanks Tamsin

iTunes launched their UK TV service today. I think it’s the first UK retailer to sell season 3 of Lost, which at £33 is roughly the same price as the DVD box set. Otherwise, individual episodes are £1.89, which isn’t an unreasonable price. Lost, Desperate Housewives, Grey’s Anatomy and South Park¹ are probably the most interesting of the 28 US shows. Damn. I was happy to pick up the odd not-available-in-the-UK show via bittorrent before this; now I’ll feel guilty if it’s available legally. There’s nothing from UK tv yet, but I’m sure it’s just a matter of time.

I haven’t tried it out yet - I don’t have a video iPod, and there’s nothing I’m itching to see - but it’s a popular service in the US and I’m sure it’ll come in useful. News articles have compared it to the ‘catch-up’ offerings from the BBC and ITV, but it’s actually quite different. I can keep these shows for as long as I want² - the ‘catch-up’ services only let you watch for a week /month after transmission. The commercial channels will surely have to implement something like this; ‘catch-up’ just doesn’t cut it - I’m happy to pay for something without adverts that I can watch whenever I like and keep forever, and it seems like it should be profitable, depending on bandwidth costs. Not sure what the BBC will do, though…

1. 22min episodes of which cost the same as 45min episodes of other shows; worth it for the Scientology episode, though. [↩]
2. admittedly if iTunes folds the DRM would ensure I couldn’t watch them any more, but that’s at least not the point [↩]

Yesterday Abi and I headed over to St. Ives (the easterly one), where the National Waterways Festival was taking place. Dad’s in charge, and I don’t think he’d deny this year’s event has been more stressful than most.

Over 300 narrowboats attend, and two months ago a train derailment closed the main canal route into the area. The shipment spilled into the water, which was bad enough, but the bridge involved was damaged enough to be structurally unsound. Initial estimates said it wouldn’t re-open until early September. Then, a month ago, everything flooded. Canals and rivers were impassable, and, worst of all, the festival site itself was under 0.7m of water. Not long after, a waterlogged bank threatened to slip and block a secondary route.

Thankfully. things worked out: the bridge ended up being completely demolished, meaning canal traffic could resume. The landslip didn’t happen / was easily sorted (not sure of the details here) and the festival site drained enough for setup to begin. Rain last week meant it was extremely muddy, and although this dried out over the weekend, there were still some bad patches:

Happily, after a wet Friday the weather was excellent for the rest of the three-day-weekend, and visitor numbers were high. Given the summer so far, this was a great relief.

Much of the festival consists of trade stands, but there’s a good amount of general crafty business too. Jewellery, ornaments, suspiciously bizarre shoes and general chocolate snackery abounded, and, shortly after raiding the second-hand book stall, I saw this:

Do not mock. Wikipedia cannot be trusted when it comes to chickens, so this is a worthy investment.

It’s been remarked that I complain about the small things, and make them seem larger than they are relative to the entire event. Which is true. On the whole, the festival is a good thing. Nevertheless, I have issues with some of the stalls. I could just about be persuaded that alternative medicine stalls should be allowed, providing they don’t make claims about curing cancer / things that matter, but those that simultaneously promote their own ’set up your own business’ pyramid schemes are just evil. Still, it could be worse.

The festival site was directly next to St. Ives, which made a change from the usual field-in-the-middle-of-nowhere, and it was pleasant to wander around town in the afternoon sun. I couldn’t help noticing that most of the shops were closed, though. On a sunny bank holiday, with a huge festival 100m away. Can’t help wondering whether they’re the same people who complain about chain stores taking over…There was a large market, though, and the town itself was very pretty.

Fun times, and definitely worth the 100min drive there and back.