Wednesday, 26 September 2007

Gunning for ARM

I'm a fan of ARM Processors - it's the most successful CPU on the planet today, by a factor of 3 or more. ARM processors go inside pretty much every mobile phone these days (it's possible, though unlikely, that some low-end phones don't have one); hundreds of millions of gadgets have them (from handheld games consoles to Network servers); they're so ubiquitous that it's worth using them for ultra-low-end microcontrollers[1][2] and it's popping an ARM processor in pretty much any device: the Marvell WLAN used inside the OLPC uses one for example.

ARM CPUs are successful, because they are actually a good design: Lean and Clean; they do more with transistors than x86 ever can; spanning a range of CPUs from <100k transistors and up to several hundred MHz with the Arm11.

So, I was really surprised to find OLPC failing to use and surprised further that Intel sold off it's ARM-based XScale CPU range last year. But now it's becoming clear, Intel are now anti-ARM and want to push their inherently inefficient x86 CPUs to ultramobiles and even PDAs.

With Moorestown Intel want to use x86 where ARM would make more sense. But probably the most annoying thing is their justification for it: "Second, if you want to get a great experience surfing the Internet, you need an IA-32 Intel architecture, because the Internet is written around it" This is just complete absolutely untrue FUD - the internet is written around being CPU independent - and that's how it should be.

People are gunning for ARM, we need to be more active when advocating it.

Friday, 21 September 2007

Performa 400 Paradise!


After I started re-using my 14 year-old Macintosh Performa 400, I just can't help raving about it! If you've never heard of one, then it's worth taking a step back in time to 1993 for a recap.


I started using Macs in 1986, 512K monochrome machines, but I couldn't afford to buy one until a year into my first proper job, in 1993. The best Mac I could sensibly get hold of was a Performa 400 - a 16MHz 68LC030 Mac with 4Mb Ram and a 40Mb HD, running System 7.1. It cost me about £1280 with a printer. Although slow even by those days it was fantastically versatile, I produce decently edited brochures; bought Think C and wrote some Mac Programs; I upgraded the memory to 8Mb; the Hard disk to 270Mb and I added a SCSI Zip drive and CD Rom. This was the machine that got me onto the internet in late 1995, early 1996 with Netscape 1.18 and Eudora 2.x for email.

Since then I've gone through a number of Macs: A PowerMac 4400, a PowerBook 5300, a tangerine iBook, a 600MHz iceBook and now my two main Macs are the 12" PowerBook G4 I'm typing this on and an iMac G5.

But I still love the P400, which I reacquired from my sister when they upgraded. It's great! For a start... it still WORKS! And with it's ethernet PDS card I can use an old web browser and also ftp. The neatest thing though is still Think C - it's about 17 years old, but as a programming environment, still much more productive than XCode. It's 1Mb of pure genius; efficient and highly usable for (really) basic stuff! And to think it's all being done on a 16MHz 68030, a cpu roughly 1000times slower than today's computers (over 200x slower than my G4), a cpu so primitive they wouldn't even put it in today's crummiest mobile phone!

It's like a trip to computing paradise - they don't make computers like that anymore :-)

Friday, 14 September 2007

Mysterious Mobile Phone Episode#1

Just a short post this. There's the feeling that the UK is sleepwalking into a Police state. Usually I take all this with a pinch of salt - or rather a hefty bag of the stuff, but now I'm not quite so sure!

A friend of mine was at the recent protests at Heathrow. Lots of fun I hear, no real bust ups between the protesters and police, just everyone playing their part. During the course of this my friend had a number of stop-and-searches, for things like, cycling. The normal thing is to not give name and address unless they have a particular reason. But on one occasion they took the IMEI number from his cheapo Nokia 1112 phone. Pretty wierd that, they had a laptop somehow connected to to the phone network or something when they were doing it. We thought it was probably just to keep track on him.

But this week, his phone has been mysteriously security locked (which they only do if it has been reported stolen, which it hasn't been). So he's now phone-less! Anyone know why the phone has been locked?

Thursday, 13 September 2007

Arctic Sea Ice Extent

I’ve finally gotten around to creating a blog, so I need to start it off somehow with a post, hopefully not a completely purile one. Since I obsess a bit about global warming and climate change I’ll start with a thrilling post about the Arctic sea ice extent.

If you didn’t know, the Arctic sea ice extent has alread
y broken the previous record of 5.32M Km2 set in 2005 by 1.08M Km2 - it’s now 4.24M Km2 and could easily fall for another 2 or 3 weeks (although it may stop any day now). It’s not just that it’s smaller than ever, but that it’s breaking records of all kinds:
  • It’s the largest single fall recorded for a single year.
  • Highest temperatures Recorded.
  • It’s the first time the Northwest Passage has opened.
  • It’s melted perennial sea ice the size of California (Ice that’s been around every day of every year since at least 1979).
  • Ocean warming itself is now contributing to ice melt.
The thing for me though is that most people are still predicting the Arctic ice caps will be free of ice in the summer by 2030, but simply looking at their graph makes it look like (at the current rate over 2005-2007) the ice caps will be free of ice by the summer of 2015. The calculation’s simple: 4.24M Km2 / 0.55M Km2/Yr => 8 years.

But how bogus is my guess? Well, I’m not a climate scientist, so probably very! But it’s worth noting that my estimate is based on a current rate of decline when it’s fairly obvious that the rate of decline is accelerating. Secondly, not only is the rate increasing, but it’s becoming more consistent. Year to year increases since 2001 haven’t exceeded 0.45M Km2 (average 0.26M Km2), which prior to 2001 averaged 0.68M Km2. Thirdly, there's a good reason to believe ice melt will increase and be more consistent - namely that the arctic ocean is playing a greater rĂ´le, absorbing more summer heat, because the cap is shrinking. And the ocean will act as a massive heat exchange, preventing the cap from fully recovering in the winter, and by regulating heat release during the yar.

Scary isn’t it?