ralpress.org

Simply stunning.

Just been poking around the Flickr stats pages. Here’s a chart of camera brands arranged by popularity among Flickr users (the stats are derived from image EXIF data). The two manufacturers in the top two spots are predictable enough, but it’s interesting to see Sony* ahead of more traditional still camera brands such as Olympus, Kodak and FujiFilm. Also interesting to see Nokia camera phones ahead of other famous camera industry names such as Pentax, Konica-Minolta and Ricoh. Flickr’s statisticians point out that camera phones are the most likely to create images without EXIF data, and thus, as a general class, are under-represented in the stats.

Canon and Nikon hold the the number one and number two spots respectively, with Canon’s EOS Digital Rebel XT model, used to take 52,320,730 (sic!) of the images that have been uploaded to Flickr. The most popular Nikon model (in terms of images uploaded) is the D50 which has been used to shoot 25,368,091 of Flickr’s images. There are some staggering numbers involved here!

Equally interesting is the “Activity Factor” figure which Flickr describes as “… a number derived from photos, members and a model’s rank to indicate cameras that are used ‘a lot’“. In other words, the Activity Factor shows the cameras that are used the most by their owners.

My own camera (now a discontinued model) is responsible for 15,285,064 Flickr images, with 27,162 of those being uploaded yesterday.

(*Before you consider buying a Sony DSLR, you might want to check out the reviews on DPReview.com. It seems that Sony have some way to go before its still cameras match the quality of its video cameras.)

Back in March I posted about Fixation’s DSLR sensor cleaning service in London (incidentally, the third most visited post on this blog) and I’ve finally gotten around to availing myself of the service. I had my Canon 30D’s CMOS sensor cleaned yesterday and the results are pretty good.

Newly cleaned Canon 30D CMOS sensor

Only one major dust speck remains (upper right edge of image - view the original upload to see) with a less serious speck near the centre of the image and a few other faint ones scattered around. Some of these specks may be on the rear of the lens. I haven’t changed the lens since the clean so I can’t yet verify that. Either way, they will hopefully be invisible in most images.

Whatever one may think of the results, it’s a significantly better job than that once performed by Canon UK’s own cleaning service on the sensor of my old D60 (back in the days when Canon were in public denial about the susceptability of their DSLR’s to dust). I once managed to clean my sensor to a similar standard as that pictured here, but it took me an entire day of painstaking work with isopropyl alcohol, microfibre cloths draped over a rubber-tipped “art chisel” and lots of hand washing, regular wiping down of nearby surfaces and wash-dry cycles for the cleaning cloths.

Fixation thoroughly cleaned both the front and back of my attached lens as well as the sensor (so leave your favourite lens attached when you drop off your camera). The lens used to take this shot was the one attached when I took it for cleaning yesterday and it has not been removed from the camera body since, so this image is an accurate portrayal of the results of the cleaning.

It would have been instructive to do a before-after comparison, but I seem to have deleted the images I took of the state of the sensor before the clean. You’ll just have to take my word for the fact that it wasn’t pretty. The last time I checked, there were at least half a dozen specks similar to the one that remains at the top-right edge of the image, two or three of those significantly larger that the one that remains. Indeed, they were large enough to be annoyingly visible in ordinary images. In addition there was a plethora of fainter, light grey spots all over the sensor. It was a bit of a mess.

The image is of a large piece of (blemish free) white card, taken out of focus with a 50mm prime at F/22 and scaled to half its original size. I used the GIMP’s auto levels function to highlight the remaining blemishes on the sensor-lens combination.

The rather drastic light-fall off in the top corners is due to the lighting on the subject and not the lens (or the camera). Other shots I took had similar fall-off but in other sections of the image. I couldn’t seem to take a shot that didn’t have similar fall-off somewhere in the image - I think this might be something to do with the maths behind the GIMP’s auto-levels function (?) as I don’t recall the same tool in Photoshop creating results with this degree of light fall-off when used on similar types of image. But don’t worry, I took other images which were not underexposed in the top corners and which confirmed that no significant blemishes or dust specks are lurking there.

Fixation is surprisingly anonymous from the street and you could easily pass it by without seeing any clue that the company is there. The staff I encountered were pleasant and helpful and when the woman who dealt with me saw Jake (my three year old son) she recommended Vauxhall City Farm, a small, free to visit urban farm a block away. It was a great place to keep the little one amused while I waited for the clean to be done. (It’s clearly run by people who love what they do and is well worth a visit, especially if you have kids - and don’t forget to make a small donation on your way out, their money box is guaranteed to make you smile).

One final note, if you travel by car to Fixation you can get there without having to pay the congestion charge if you thread your way carefully. They have a limited number of parking bays in front of their offices, but if they are full you can always park in Tesco’s car park, barely a hundred metres away. (Tesco allows its customers a three hour stay. Fixation staff no doubt provide Tesco with a constant source of revenue, so it’s probably not that much of a liberty to park there).

A couple of weeks back I bought an Asustek Eee PC 900 in order to obtain some seriously mobile networking and computing (see screenshots). I considered waiting for the 901 but really wanted to take the device on a trip down to Cornwall so took the plunge. I opted for the black model; no kitchenware here thank you ;-). It sports 20 GB of storage split across 4 and 16 GB solid state drives, 1GB of DDR II SDRAM, a 900MHz Celeron Mobile CPU and a 1024 x 600 resolution display on a 22.7cm (8.9″) diagonal TFT panel.

Ubuntu Hardy Heron running on my Asus Eee PC 900

Like many other Eee PC owners it seems, it didn’t take long (about 5 minutes) before I decided that the pre-installed version of Xandros had to go. Functionally, there’s nothing wrong with the default OS but it probably won’t appeal to those who are used to the standard GNOME, KDE or other more conventional Linux desktops. Eee PC’s Xandros eschews common Linux desktop configurations in favour of a simple tab-based application launcher style interface. It also presents you with licensing agreements which must be agreed to for both the customised operating system and at least one of the tools. Nevertheless, it possesses a capable set of applications that will take care of the essential tasks such as web browsing, email, word processing, Skype, music, basic photo and video editing, and more besides. It uses tried and trusted free software favourites such as OpenOffice, Firefox, and Thunderbird, together with some of the most popular KDE applications providing the rest of the core functionality and features. Importantly, the Eee’s version of Xandros has (naturally enough) been tailored specifically for the machine and everything works perfectly with the hardware. However, it doesn’t feel much like Linux, not least because neither a terminal window nor package manager is available by default.

If you want a “more kosher” Linux desktop, the full KDE-based Xandros desktop can be installed, but I decided to go for Ubuntu instead. I initially tried a version of Xubuntu 7.10, eeeXubuntu, customized for the original EEE 701, but the wi-fi packed up after I downloaded the first batch of updates and a couple of hours of diddling about with suggested solutions failed to restore it. I was concerned that a  GNOME installation might be a bit much for the Eee’s relatively limited hardware resources, but was attracted to the idea that the Eee would be running the same apps and tools as my desktop machine and so gave Ubuntu EEE a whirl, a customised version of Ubuntu 8.04.

The install went pretty well and most, if not all, of the Eee’s hardware worked without further tweaks. There were some issues such as the audio volume adjustment keys and the wireless adapter power switch failing to work (although the former can always be adjusted via the standard software controls and the latter can be switched on and off in the BIOS). In addition, shutting down Ubuntu from within a GNOME session fails to subsequently power off the hardware (requiring instead a five second push of the power button), whereas choosing the shutdown option from the login screen shuts the machine down correctly. There are a few other minor glitches as well, although workarounds and fixes, which I haven’t yet got around to trying, exist for many of these. (The fact that I haven’t got around to trying them yet indicates how relatively trivial the issues are.)

Multiple GNOME windows running on the Eee PC 900

Glitches notwithstanding, the core features worked fine. Ubuntu installed without fuss and the Wi-Fi worked out of the box. I was immediately able to connect to my WPA2 network and most of Ubuntu’s tools and standard functionality seem to be present and correct. My Wi-Fi connectivity survived the initial surge of updates. I needn’t have worried about performance, it’s perfectly acceptable thanks in part to the full gigabyte of DDR II RAM. The only performance irritation so far has been the boot time which is a rather agricultural 53 seconds. I didn’t record the Xandros start up times, but I think the boot up time was well under half the time that Ubuntu EEE is taking (there are scripts available that claim to reduce Ubuntu Eee’s start up time, I’ll try them out and report back). Not all of that start up time is taken booting Ubuntu, it takes around 6-7 seconds for the Eee’s BIOS splash screen to appear and the grub menu displays at around the 10 second mark.

I admit that I got a thrill when I first saw the Ubuntu logo together with the “circle of friends” on the black boot screen on such a small device. The ability to carry Linux around on a device weighing around a kilo, that is smaller than a typical computing book, is pretty cool. The key differences between Ubuntu EEE and the standard version of 8.04 seem to be driver installation scripts, smaller fonts, Evolution, GIMP and some other apps not installed by default, and the creators have tried to banish the brown and orange hues, replacing them with a conventional blue, Xubuntu-style desktop (based on the Clear Looks theme, I think). I wasn’t having any of that and restored the Human theme and Ubuntu’s default browns and oranges A blue background still appears momentarily somewhere between logging in and the session desktop appearing. I guess the Ubuntu EEE authors are in the “lose the brown!” camp.

This picture should give you a pretty good idea of the dimensions of the Eee PC 900

OK. I’m out of time this evening. I’ll probably create a Part II to this post in the next few days.

Screenshots

NB. The Eee PC’s screen pixels are physically smaller than those on most desktop LCD panels. For example, when viewed on standard 17″ and 20″ TFT panels the “original size” (1024×600) screenshots in my Flickr set will appear around 30% larger than the actual screen size on the EEE.

Of course there is intelligent extraterrestrial life. The clues are all around us. Hell, they even report the results of their Earth exploration on Wikipedia:

“Furthermore, humans have developed culture and technology (such as music, telephone and radio) that allows them to generate, record, transmit, and broadcast sound.”

Fantastic. UK taxpayers are paying to be spied upon. Technology will make slaves of us yet; this is not how it was intended to be.

UK.gov dishes out £19m for comms snoop data silos

Lawrence Lessig predicting that the US government will use the pretext of an “Internet 9/11” to clamp down on the freedom of the network (the following is my own incomplete transcript):

“I had dinner once with with Richard [?] … and I said to him, is there an equivalent, is there a ‘Patriot Act’ or an ‘i-Patriot Act’ just sitting waiting for some substantial event that will become an excuse for radically changing the way the Internet works? And he said, ‘Of course there is!’ and I swear that this is what he said, and I quote, ‘And Vint Cerf is not going to like it very much’. So this is the big terror, right? They’re just sitting waiting for the invevitable to happen, and then SLAM!”

Source video

It’s even more alarming given the information in the first half of the clip about how the Patriot Act was conceived and introduced. This is Lawrence Lessig, a Professor of Law at Harvard, not some hysterical conspiracy nut. We’ve been warned people!

If you have any attachment whatsoever to the rule of reason, to the notion of humanistic individual rights and to the essential equality of men and women, then you really need to read this post. There exists the very real possibility that the concept of “defamation of religions” becomes an international legal norm. Write to your MP, join an organization that defends secular principles (here’s one I recently joined), agitate, demonstrate on the streets, write a blog post (it’ll be indexed by Google, at least), do whatever you can. This is a grave matter and the process is occurring now, not “might occur”, it is occurring, and as the quotes in the post linked to point out, has been for the last 10 years or so, with minimal opposition so far. Don’t be a frog in the boiling pot and then merely shrug when it becomes fait accompli, do something now to make your voice heard.

This matter has serious ramifications for those who do not want the rules that govern their lives to be dictated by theocrats and superstitious belief. As Ophelia Benson remarks:

If the right to free speech can be trumped by a subjective perception of insult, then there is no right to free speech. That’s it. All over.

(I don’t like writing such strident posts, but this is important, isn’t it?)

Following the discovery that Comcast has been interferring with its customers’ ability to download torrents over its network, the Electronic Frontier Foundation has released the Switzerland packet testing tool to enable you to check to see if your ISP appears to be forging packets with the intent of interferring with your ability to share files.

I’ve been running Switzerland this afternoon (the client, not the country), and after exchanging many thousands of packets with other clients on the Switzerland network, the EFF server found no evidence of foul play (well, three malformed packets, but Switzerland reported that these were most likely re-written by my NAT router).

The major UK ISPs have only recently signed their memorandum of understanding with the BPI and the government, and it is not yet clear whether or not they will be attempting Comcast-style interferrence. BT has been contacting us like crazy recently trying to persuade us to move to a different broadband deal, something which will no doubt include a new contract and terms and conditions. It might be interesting to see how those T&Cs differ from our existing set, perhaps containing some kind of opt-out or loopholes to tacitly allow traffic monitoring and bandwidth restriction. Next time they phone, I’ll ask to see a copy before I’ll be prepared to discuss any change to our existing contract with their sales-people.

I will be running tests with Switzerland on a regular basis to keep a check on BT.

PS.

• In addition to core Python, you’ll need python-scapy and psycop, plus a running NTP daemon (all just a Synaptic away on Ubuntu) or Switzerland will moan.
• You’ll need to sudo mkdir /var/log/switzerland-pcaps
• You’ll need to ensure your firewall doesn’t interfere with the testing
• I had to re-start Switzerland several times before it was happy that my computer’s time settings were in accordance with those of my timeserver (but hey, this is version 0.0).

Additional links:

EFF “Switzerland” packet monitor tool looks for ISP meddling

——————————

UPDATE: In the light of the following article the above is probably largely academic:- BT slams bandwidth brakes on all subscribers. BT don’t need to re-write packets, they simply block the packets they don’t like. This probably explains why my torrent download speeds were around eight to nine times slower than those of HTTP the other night. Zen Internet seems to be getting a good press. I’ve seen user reports claiming that it has no Phorm nonsense, doesn’t discriminate against non-HTTP and non-email traffic, and has good tech support. Sounds like a real Internet Service Provider to me. They do have monthly bandwidth caps, but they’re completely open about them and as far as I’m concerned, caps are a practical matter, not an ethical one. As things stand now, I’m looking to move to Zen, unless I can find an even better deal.

Turning the tables on unsolicited commercial callers, in style.