mechtopia

A new breed...

Season change denial by ohp

Posted at 10:36PM Nov 23, 2009 by ohp in Comic Strips  |  Comments[0]

Referrer Spam: what do you do?

Back in the early noughties, web traffic was increasing to the point where you had to create summaries of the traffic for analysis. Enter the web stats package.

One thing I noticed at the time was that many web stats packages published their referrers, and that many of the people had (often unintentionally) made their stats public.

Google had become the biggest web search engine on the planet, and this posed some ideas. What if you searched for web stats pages on google, then created a link to that page. you would instantly register a link back to your own site. In fact, why bother with a link? What if you just faked the referrer? you could push your link to the top of their referral chart very quickly with an automated script, and google would register another link to your site, increasing the pagerank.

No matter how appealing the idea, google was already by then battling against underhand techniques, and although at the time I was sure I was the first person to think of the idea (something which, in hindsight is ridiculous) I figured that any new underhand techniques would be factored in to the indexing algorithm, and would come back to bite you on the rear.

I'd forgotten completely about my underhand scheme until the 8th of this month, when I examined my logs to find that I was the victim of this technique; every 50 minutes a host was requesting the root document from my web server, with a referring url that on inspection, did not contain a link to my site. I considered my options. Do I:

1. Ignore it?
2. Block the IP address on the firewall?
3. Install a comprehensive web threat management layer to filter out such requests?
4. Put an exclusion rule for that referrer in the module that generates the referrers?
5. Write a snotty email to the admin contact for the whois entry on the source IP, and cc in the admin contact for the referrer domain?
6. Create a rewrite rule that directs requests from that specific IP to a range of large downloads, in an effort to swamp their broadband connection, or crash the script?
7. Create a script that spams their domain with a referrer of wwww.stopspammingmysite.com?

What would you do? answers on a postcard please. Next week I'll tell you what I did, and what happened.

Posted at 11:44PM Nov 18, 2009 by ohp in General  |  Comments[0]

What if a medium recieved very specific information from the other side?

Well, it just wouldn't work would it...

Posted at 06:43PM Nov 08, 2009 by ohp in Comic Strips  |  Comments[1]

I have a confession to make...

It's true. I've tried to hide it all of these years, but the truth will out: I was a competitor on Robot wars.

Sporting a terrible goatee beard, I presented a robot that was eliminated in the first round.

Posted at 11:15PM Nov 06, 2009 by ohp in General  |  Comments[0]

The dowse-o-matic 1000

Recent visitors to my blog may have read the post about the busk-o-matic 3000. buoyed with enthusiasm for constructing computer controlled mechanical devices, I have been pondering the creation of a dowsing device.

The aim is to allow people to test their dowsing skill in an automated fashion. My preliminary thoughts are to have a pair of reservoirs linked by a pair of tubes. A computer will randomly select one of of two solenoids that will allow water to flow from the upper to lower reservoir via one of two paths, and the dowser will press a button to indicate which of the two tubes contain the flowing water.

This process will be repeated until a statistically significant number of tests have been performed.

I don't for one moment think that this apparatus will fulfill every dowsers requirements for a test... I see it more as a fun device to make scientific testing more accesible to the masses.

So.... what do you think? Do you have any suggestions?

Posted at 11:40AM Nov 02, 2009 by ohp in Projects  |  Comments[0]

Mundane conspiracies

It strikes me that only large, newsworthy events tend to be subjected to the scrutiny of conspiracy theorists. How would they react to everyday pictures?

Posted at 04:28PM Oct 30, 2009 by ohp in Comic Strips  |  Comments[0]

Dr Perry Orford: Futurologist

A few years back, I created a blog for a ficticious futurologist. At the time there seemed to be a trend for people to make fantastical claims about the future that didn't stand up to scrutiny, and I felt that they were ripe for parody. It was very low key at the time, and I only really promoted it amongst my friends.

If you read the last blog post, made a few years back, you'll notice that Dr Orford predicted the widespread use of augmented reality :-)

Posted at 12:03PM Oct 29, 2009 by ohp in General  |  Comments[0]

DLP: The end of the rainbow?

It's generally acknowledged that DLP projectors have the edge over LCD in many respects, with great colour reproduction, great contrast ratios and the ability to project 3D movies with shutter glasses, but there's one thing I believe that is preventing many DLP projectors from winning all the awards, and that is the colour wheel.

The colour wheel is a way of processing all of the colour information with a single light path (LCD projectors, for example, need to create a separate beam of light for each colour, then recombine them) by spinning a coloured wheel - this projects each colour rapidly in sequence.

Having one path means you only need one DLP chip, and this lowers costs, but it has two side effects: The first is the sound of the colour wheel : LCD projectors have been engineered to produce virtually no noise in use, but DLP projectors have two sources of noise to get rid of - both the fan, and the colour wheel. The second issue is one that affects me personally, and that's the rainbow effect. Despite the fact that the colour may be switching hundreds of times a second, this is still slow enough for many people to detect rainbows as their eyes flit around their scene. This is not only annoying, it gives me a headache after around 15 minutes of viewing.

Even if you don't suffer from this yourself, it might affect one of your friends.

Three chip DLP projectors use a chip for each colour, combining the best of both worlds. Unfortunately they're still very expensive. Only when this happens might it might usher in the end of the LCD projector.

Posted at 01:56PM Oct 26, 2009 by ohp in General  |  Comments[0]

Using pendulum dowsing to fix computers

I created this video last year.

I currently have another three videos on the pipeline, I'll keep you posted!

Posted at 08:18PM Oct 25, 2009 by ohp in General  |  Comments[0]

Psychics

Ages ago, I created a few strips on stripgenerator. As the blog has got rather quiet, I thought I'd fill it in with repeats :-)

Dead People by ohp

Posted at 03:05PM Oct 19, 2009 by ohp in Comic Strips  |  Comments[0]

My T-Shirt site

Everyone and their dog is selling T shirts. There are a number of sites that allow you to design them online, and then set up shop in a matter of minutes.

You can purchase these T shirts at unteeshirt.spreadshirt.net

Bandwagontastic!

Posted at 11:44AM Oct 09, 2009 by ohp in General  |  Comments[0]

The Busk-o-matic 3000

Ever since attending my first karaoke night, I've wondered if I can harness the power of the wailing and catterwalling masses for the good of mankind.

So I created a device, which I have called the busk-o-matic 3000

Here's how you use it:

• Each singer is given an RFID chip that uniquely identifies tham as a singer.
• The singer places an RFID chip on the back of the unit at the beginning of the song.
• As the song progresses, the audience puts money in one of two slots, depending on their positive, or negative reaction to the act.
• A running balance of positive and negative contributions is indicated by the arrow.
• If the arrow should move into the red zone before the end of the song, it will be ended prematurely.
• At the end of the night, prizes can be awarded to the best singer, the singer that raised the most money etc..

Here's how it works:

Coin counting is done by a pair of Mars Electronics Cashflow 126 coin mechanisms. These are nice mechs in that they have a nice simple parallel interface mechanism, which is pretty cheap and easy to interface with. I've used a phidgets USB 16/16/0 interface board to interface with the computer.

The RFID reader is a phidgets USB RFID board, which reads the unique 125khz em4102 tags

The arrow is moved by a standard rc servo, via a usb phidgets servo controller.

The brains of the operation is a java program that maintains an internal database of all of the deposits made. This can be queried for winner reports.

The unit has been tested in a real environment. In an hour and a half, 11 singers raised 67 pounds for charity.

First Test Run

Here are some charts generated by the Busk-o-matic's internal database after the first test run.

Improvements to be made:

• Improve coin validation. About 30% of the coins were rejected. Some coin path modifications would address this.
• An audible and alarm when the singer has reached the low threshold
• Make better use of the coin sorting mechanism offered by the mechs.
• The unit could be made more visible in dark lighting enviroments
• Nicer looking case
• Enclosed arrow, so that people don't try to break it off!
• An embedded motherboard. I have plenty of spare ITX boards on which I can run Linux

Watch this space... coming soon, the busk-o-matic 4000!

Posted at 10:57PM Oct 06, 2009 by ohp in Projects  |  Comments[0]

The great HDMI Scam

It always annoys me when people are sold unnecessary electroncs. Often people are bamboozled into purchasing components that are way overspecified for the task in hand. One recent example that I have been noticing is as new high definition equipment is coming onto the market, people are being sold overpriced HDMI cables, and are being told that it improves the picture quality

In the early days of video cables, you’d need to link up using RGB, component, or VGA: all analogue formats, and it was important to preserve the signal quality by using good cables for long cable runs.

HDMI is a newer cable designed for consumer video, and supports high definition video. Basically the designers took DVI-D (a system for connecting computers to monitors), implemented content protection (HDCP), added audio channels and squeezed it all into a compact connector.

So what’s the point?

Well, HDMI is a purely digital connector. This means that whatever interference is picked up by the cable en route, the receiving device can separate out the digital signal, and reconstruct it, exactly as it was before it was sent down the line.

Will buying a more expensive cable make the picture quality better?

Well, do you think this web site would look better with an expensive ethernet cable? Do you think that if you had a gold plated modem lead, then online photos would look crisper?

Digital transmission changes the way in which failure occurs. If your cable was bad enough to break up the signal it would break up entirely, like Digital TV, or when you're on a GSM mobile phone, and you travel through a tunnel. It wouldn't slowly degrade like in an analogue system, where you'd experience hiss as the signal degraded. Even really cheap HDMI cables will carry a signal way above the threshold required for a perfect image.

My advice is, don't spend any more than £5 on a short HDMI cable (in the order of 1m) and no more than £30 on any cable under 10m without built in electronics (some really long cables will use repeaters, for example).

Posted at 12:20PM Sep 30, 2009 by ohp in Skepticism  |  Comments[1]

Brianspredictions.com - did he predict the UK lottery numbers?

This article was originally posted to the jref foundation forum in February 2007. Additions have been made in red

I was googling for something entirely different (boring techy stuff) when I came across this site:

http://briansprediction.com/

Well, I thought I'd have a laugh, so I dug about a bit. There's a small bit of text at the top that says "More than 4000 confirmed dream predictions"

Uh oh, I thought. This is the old scattergun approach, where you write down a load of gibberish, and wait until some of it matches. Ultimately much of it seems to be the case, but I came across this bit:

http://briansprediction.com/dd/5050.htm

Wow. He predicted the outcome of the lottery draw on the 3rd of January, in a dream on the 2nd! Not only that, but he guessed the bonus ball too.

Now, this is a bit of a statistical nightmare to work out, but the chances of winning the national lottery are 1 in (49!/(43!*6!)) = 1 in 14 million. Add the bonus ball prediction, and I believe that comes to 1 in 588 million.

Watch out Mr Randi... your million dollars might be at stake! Oh, wait, anyone that can predict the national lottery numbers doesn't need a million dollars!

In light of this amazing power, I thought I'd poke around a bit more...

The first thing that is odd is the order of numbers. The plain text of the prediction (oh, and there's a silly face there too) was

"394613174749 48 UK SH"

When you compare this with the actual results:

13,17,39,46,47,49 Bonus Ball: 48

You can see that the difference is that the middle two numbers had been switched to the front in the prediction. So what? you might say, perhaps they were drawn in that order...

well, they were drawn in this order:

17,46,49,13,47,39 bonus ball: 48

You'd expect a psychic to read them out in the order of the draw wouldn't you? Well, perhaps I'm making too many assumptions.

OK, so what about the time of the prediction?

There's a link above the image with floatover text that reads

"This documents creation date has been confirmed"

This takes you to a site called ymlp.com, which appears to be a mailing list provider. There is an archive of posts. the one we're interested in is:

http://www.ymlp161.com/pubarchive_show_message.php?brian2179+733

If you take it on face value, these predictions were posted at "2.3.2007 1:28 PM EDT USA" which is, if I'm not mistaken, is 5.28pm GMT (UK time) on 3rd February.

So when was the draw? The UK national lottery site doesn't have the actual time of the draw, so I called them up to find out. They could only tell me the usual time, even after referring me to their line manager. They suggested that I call the BBC. When I called the BBC all they could tell me was the draw was between 7.50pm and 8.35pm on the 3rd February. (the draw occurs during a panel game style quiz show)

Wow. This posting was made around 2 hours and 20 minutes before the lottery draw!

As one final check, I checked to see that the image files are part of the mail itself, as an attachment, so I know that the image files were mailed out at the same time as the post itself. The crucial one in question has the following URL:

http://briansprediction.com/pri/FEB20097.jpg

Ah. Looks like it's pointing back at Brian's site. Well, it's just an innocent mistake isn't it? we all know this image was created at the same time as the post, don't we? How can we tell?

There's a thing with http called the last-modified: header that most web servers will take from the file date stamp on the filesystem. When you download a file with the wget utility, it will preserve this date stamp on the file you downloaded. Now, it's not foolproof, but if it shows a date before the predicted event, then either this bloke is a bit clever with web sites, or he's psychic. If it's after, then he's either a faker, his server's clock is wrong, or he didn't upload the premonition until after the lottery results, which is odd, as the mailing list post that references it was posted at 5.30pm.

I checked the clock was correct from a server header response.. yes, it shows the current time, so the server clock isn't wrong.

So what's the time stamp of the image file I hear you ask?

Drum roll please....

3rd February at 10.26pm, GMT, around 2 hours after the national lottery draw.

Posted at 11:06AM Sep 30, 2009 by ohp in Skepticism  |  Comments[1]

A New Blog

Well, After years of half-hearted attempts, and posting to other people's blogs, and anonymous blogs, and forums, I thought I'd stick it all in one place. I don't plan on regurgitating other people's stuff... this should hopefully all be original material.

Posted at 02:21PM Sep 27, 2009 by ohp in General  |  Comments[0]