Y Safle

No one could be so stupid as to try and pass off a stolen cheque for the same amount as the whole of the European cybercrime market could they?

Well yes, actually. Apparently if you are not too good with figures, and you have just stolen a cheque, then you may just believe that it is possible to cash a cheque in someone else’s name for US$ 360 thousand million! I suppose adding 10 noughts after the 36 just looked so much more pretty!

But here is the scoop. This story was brought to you by the Dallas Morning News. But at the time of writing this article, if you searched their site for “360 check” you would not turn up this story but instead a spreadsheet of “disdchecks” for July containing some 3000 names and payments with many personal details included on them.

Oops!

I can only suppose there is something in the water in Dallas that is affecting concentration and mental function!

have you ever heard of the Happy Endings Foundation? Well neither had I until Sunday morning when I saw on the BBC news an article about their campaign to ban books with sad endings for children.

The BBC paid for two experts - a child psychologist and someone else - to come into the studio and pontificate on how this campagn was misguided. What a pity though that these experts, and the BBC news researchers (and the Daily Mail, who were also taken in) were not so expert in the realm of critical thinking.

The first clues that all was not well could be found in the web site itself. Rewrite Lemony Snicket? Are you allowed to do that? Why would you want to? Also the BBC admitted they could not actually contact anyone from this web site to come on to their programme.

Also, what an odd list of books they were saying had happy endings.

But one skill that should be taught to school children up and down the country when we teach them basic IT skills is how to find out who pubilshed a web page. This is not actually very hard. Point your web browser at any of the whois services, but I particularly like this one:

http://www.whois.sc/

Look down the page for the registrant details. In this case we have a registrant as follows:

Registrant Name:Peter Rope Registrant Organization:ArtScience

Usually this is enough (and it is here, if you know who ArtScience are - a promotional company trying to promote Lemony Snicket!) but this whois search has a great feature. It will do a reverse domain lookup to find out what other web sites are hosted on the same web server as this one. In this case we find:

Artscience.net Artscienceclick.com Charlie-bone.com

And a quick click on any of these links will quickly show you that these people are in the business of marketing children’s books.

The BBC was duped by marketeers. I hope an apology will follow for using license payers money to advertise someone’s books.

The Daily Mail was also duped - but that is par for the course.

But the BBC really should know better.

Not my usual fayre for this blog, but there is an excellent issue of the Journal: “Library Philosophy and Practice” available. Nearly the whole issue speaks about how Google can be coupled with traditional library services in a manner that benefits both (i.e., none of the hit and miss nature of Google, whilst enhancing the work of librarians).

There is also a very enthusiastic article about the benefits of Open Source and Open Access Journals to libraries. Wel worth a look

The BBC writes:

One in 10 web pages scrutinised by search giant Google contained malicious code that could infect a user’s PC.

That may seem a worryingly high proportion! Fortunately it is also nonsense.

Looking at the actual paper, it seems that Google in fact analyzed several billion pages, and sifted these with a preliminary analysis tool called MapReduce

“MapReduce processed all the crawled web pages for properties indicative of exploits”. As it says in the paper, “MapReduce allows us to prune several billion URLs into a few million”. This process left around 4.5 million pages that were likely candidates. Out of those, they found 450,000 pages that they were confident were correctly identified as malicious.

So it is not 1 in 10 pages. It is perhaps 450,000 pages in several billion. It’s more like 1 in 10,000.

The BBC failed to spot this - although perhaps that is not surprising, considering some examples of previous sloppy scientific reporting by the corporation (and most other media corporations of course).

The media is working itself up into a feeding frenzy again. This week’s issue: the alleged dangers of WiFi radiation. First BBC Radio news, then Newsnight, now the Daily Telegraph have all had articles from quoting from people who believe that there might be health risks from WiFi. The BBC programmes at least balanced the sillier claims by noting that WiFi power output is far lower than that from mobile phones.

The Daily Telegraph, on the other hand, had this nonsense:

However it is believed that a classroom containing 20 laptops and two routers could combine and be equivalent to the emission from a mobile phone.

Notice the use of passive voice: “it is believed”. Who believes this?

It turns out that no-one does, although the Telegraph quotes from the “Powerwatch” web site. On this site, there is a calculation designed to show that a room full of 20 computers and an access point (or two) is approaching the power output of a mobile phone call.

But this is arrant nonsense. It is nonsense because of the silly assumptions being made about average distance from antennae. It is nonsense because they first make an assumption of low power output from the phone, and it is especially nonsense because it makes no difference whether you have one computer or fifty in a room. The maximum power output is the same - capped in the UK at 100mw (but usually less than this).

Let me explain: WiFi stations are receiving a radio signal when they receive data. But if two stations transmit a signal at the same time then what occurs is a collision - the signal is garbled and not received. The solution is that WiFi uses a collision avoidance scheme to try to ensure that packets are never sent out whilst another station is already transmitting. A couple of carrier sense mechanisms are used to do this, and the result is a very low collision rate. Essentially, if one station is talking, all the other stations are quiet. Click on the thumbnail image to see how this looks in practice. Each wireless station waits a random amount of time before transmitting, and if another station transmits first, the wireless station will quietly defer to it.

Therefore the maximum power output from a room of 20 computers remains capped at 100mw. What is more, no one station is transmitting all the time. For the majority of the time, even on a congested network, each station is actually transmitting nothing.

And that brings us to the next bit of nonsense - the assumption of the amount of traffic in transit. The powerwatch page assumes more than 100% usage for periods of two hours or more in its calculations. Note that if a single wireless station on an IEEE 802.11g network were downloading constantly (and that the downstream network could maintain the same throughput), then even assuming protocol overheads reduce the actual throughput to a mere 30 Mbps, over a course of two hours, that station would download some 26 Gigabytes of data.

To put this in perspective, fairly heavy Internet users download about this quantity of data in a month. Light users (people just browsing the web, using email and similar) would not get close to this figure.

26 Gigabytes is the equivalent of about 44 CD images, or several thousand podcasts. It is a guge quantity of data - not the kind of thing people will be downloading in one sitting - and certainly not regularly.

Ah, you say - but spread between 20 computers? What classroom environment is encouraging pupils to download over a gigabyte of data per computer? None!

So the assumption of 100% utilisation is nonsense.

Next problem: The assumption is made that people are - on average - a metre from the transmitting antenna.

This is just silly. The writers seem to know that magnetic fields decrease by the square of the distance from the antenna, but they assume that on average, people are a metre from the emitted radiation. Based on the fact this is a classroom, and 20 stations are variously transmitting, the fact that some of these transmitters will be many metres away only really makes this assumption if the pupils have swallowed the transmitter.

Admittedly, I wouldn’t put this past some children - but if they have done so, they probably have worse problems to deal with than the emitted signal from the station!

No, in fact the vast majority of the transmissions come from the access point. (Remember, we download much morethan we upload) this may not even be in the class, but if it is, it will probably be several metres from even the nearest child. The law of squares tells us that these signals will be far lower than the silly assumption made on the powerwatch page and repeated uncritically by the Daily Telegraph.

As one commentator on the Telegraph page says:

“Can there ever be a better example of why the current decline of Physics teaching in schools and universities is so worrying?”

Let’s give the last word to Mike Clark, senior spokesperson for the Health Protection Agency. He has run the figures rather more intelligently than the Powerwatch pressure group, and tells us that 1 year of exposure to WiFi radiation in a classroom is equivalent to 20 minutes on a mobile phone.

Maybe we could quibble and bring that down to a month or so. So what? The worry with mobile phones is that the radiation causes excitation of water molecules in the brain near where the phone is held to the head. This causes a slight warming effect which could theoretically be a cause for concern, although there is no proof of harmful effects.

WiFi radiation - if it causes any warming at all - is so slight as to be unmeasurable.

There is no risk here.

There is a new search engine around that offers a human guide to help you with your search. The company is betting that revenue from the advertisements on the site will pay for these guides… I’m not convinced, but it must be worth a look.

Take a look at the ChaCha site. If you have a difficult search to make (and you are frustrated with all those Wikipedia links that keep obscuring the good stuff), then ChaCha may be the search engine for you.

For other stuff about search engines, look at my Really Useful Search Engines post.

I was reminded recently of the Geek Code.

This is quite old, but I have always thought it has a major deficiency: any true geek would not use such a cumbersome code. For anyone who has written a text editor in one line of APL, and a dancing sig in four lines of perl, the ineligance of the geek code speaks volumes.

For each attribute we are told to specify the extent of our geekiness by modifying the attribute with +, or ++ or - or - -. However, this requires two unicode characters (32 bits), or at least two ascii characters (16 bits) to represent something that can be expressed in just two bits.

As the attributes are all text based, we can used the unused bits of the attribute and XOR the modifier bits to them. This needs some care, as we don’t want to return any non printable characters, but a true geek code would use just one byte of data for each attribute with modifiers.

Of course, quiche eaters would have none of this. A quiche eater would probably want to structure the code something like this:

begin occupation_code 
  attribute.x666.type = "Geek of Computer Science" ; 
  attribute.x666.modifier = modifier_struct.values.positive ; 
  attribute.x666.strength = 2 
end occupation_code 

(They would probably also write it in XML or provide a published schema under LDAP).

To accomodate both real geeks and quiche eaters, we could perhaps have a compact but human readable code, where the modifier is written as an integer, and also acts as a separator (as all modifiers are integers, and all attributes are [a..z] or [A..Z]). This is similar to the scheme that works behind the scenes in the DNS. A scheme written by geeks and for geeks.

Thus we could have a Geek code that would look like this:

GCS3M2R5h7q9MMOUSE4PRES0i-1

I have just noticed that since I wrote my XML code samples, with some rather long lines, Internet Explorer has started floating the right hand sidebar of this site down to the foot of the page.

Not strictly a bug. It must be a font size thing. But for best results, use any browser except Internet Explorer

It must be something about vehicle manufacturers, because they seem to be some of the worse offenders for this kind of nonsense:

Upgrade your browser This website does not support your current Browser version. For the best experience of this website we recommend you to use Microsoft Internet Explorer 5.0 and above.

Last time I had one of these, it was when looking up information about Renault cars before buying a new vehicle. We did not buy Renault. We did not even bother go to a Renault show room.

This time it is Scania (don’t ask!)

Now I browse the Internet with either Safari or Firefox. Both are far better browsers than Internet Explorer. Occasionally I use text browsers or browsers on handheld devices, and these stupid sites break with these browsers too.

Now if everybody wrote sites with an eye on web standards, then we would all be much better off, and we would not have brain dead retailers trying to persuade us to infect our machines with Microsoft’s dangerous products, telling us we must “upgrade” to this rubbish before they are willing to sell us their products.

In case you were wondering how the Internet works, US Senator Ted Stevens, representing Alaska and legislating to destroy the Internet experience for Americans, gave us this wonderful definition:

The Internet is not something you just dump something on. It’s not a truck. It’s a series of tubes.

So now you know.