Satnavs in the spotlight: Satnav tells driver to drive off cliff

(excerpt from today's Independent newspaper (UK) here)

A driver whose car was left teetering on the edge of a cliff after following his satnav was ordered to pay nearly £900 for driving without due care and attention.

Robert Jones, 43, nearly plunged down the 100ft cliff in his BMW after obeying instructions which sent him along a steep, narrow path, in Todmorden, West Yorkshire, on 22 March.

[...]Speaking after the incident earlier this year, Jones, who works as a driver, said he trusted his satnav system and relied on it for his job.

He said he continued to follow the instructions when it told him the footpath he was driving on was a road.

Police officers were called to the scene after reports that a BMW was hanging off the edge of a cliff off Bacup Road.

Snatch wars... absolutely hilarious

Rediscovered this after a long time: got to be the funniest movie on YouTube...

Does Apple have rather large cloud ambitions?

Cloud computing is definitely the buzzword of the moment, extremely necessary to many software businesses and their delivery of easy-to-use software services over the internet, but also rather overhyped of late. It was refreshing to read an interesting piece in the CultOfMac blogsite from a few weeks about an amazingly large computing facility that Apple is constructing in North Carolina in the US. This story has now resurfaced among the bloggers in the last few hours and gaining visibility.

The facility, which apparently will be one of the largest such facilities in the world, as large as Microsoft's new Chicago facility, does seem to mean that Apple is looking to cloudize at least some of its services in the future. Rumors aside, this does show how the larger players seem to be gearing up to a future that's probably more cloud-based than not. Will Microsoft play catch up with Google in offering cloud-based Windows OS in this manner with Azure? Probably. Will Apple do the same? Known for their secrecy and protectiveness of brand and product development integrity, probably not to the same degree.

But the facility sizes mentioned do start to illustrate the amount of investment required to really provide economies of scale in this space. Those without the deep pockets - and the robust business models to power those deep pockets - will increasingly have to pack off part of their infrastructure to 3rd party cloud providers, and retain the core assets that would be most at risk in case of an (unlikely but finite risk of) catastrophic systemwide provider failure.

I mean, if even Google grid/cloud computing maestros can have increasingly common cloud blackouts as demand increases for their GMail/GoogleApps cloud services, then this scenario should be in the standard contingency handbook for the rest of us, and not to be dismissed just because it says "on-demand" and "unlimited" in the marketing blurb.

It seems that beyond the hype, the reality of cloud/on-demand computing is already here. It is no longer a heated debate over if, nor over when, but over the rather more pragmatic "what".

FT somehow misses the DPJ triumvirate

According to the FT today, new Japanese premier Hatoyama's position was "weakened" due to his nomination of Ozawa to the post of DPJ secretary (not too dissimilar to the vice-presidential post in the US). Unfortunately for the FT, Japanese politics does not work in blacks and whites, and is a dense forest of greys, where one grey seeks pole position over other greys. And the Ozawa nomination was the best thing Hatoyama could do.

The new ruling party of Japan, the DPJ, is run by a triumvirate of 3 men, none exactly what the electorate want, but that between them pull together the opposition to the Japan-one-party-system of the LDP. Hatoyama, Ozawa and Kan are the triumvirate and no-one can convince me otherwise. Kan wrested power from Hatoyama a couple of years ago, then got embroiled in a scandal and had to step down, but the wounds of Kan's backstabbing were too fresh at the time for Hatoyama to take over and Okada, a midweight party leader took the interim job. It is possible Hatoyama engineered the merger with the Liberal Party chief Ozawa that led to the current enlarged DPJ. In hindsight, this was a smart move, but wasn't seen as such at the time.

After the merger, Ozawa became top-dog at the DPJ but subsequently got embroiled in a scandal and had to step down. Only guy left of the triumvirate was Hatoyama who had licked his wounds and waited in the wings patiently as gen-secretary. He was thrust into the position of leader, and as ex-PM Aso took the rap for the worst economic recession since the war, Hatoyama's squeaky clean position made him the new PM of Japan, even though he is not charismatic, displays no particular affection for showbiz politics, and can't really connect with the average Japanese. But by default he has been chosen as the guy to lead Japan out of the recession.

Back in the day, Ozawa was a LDP new-kid-on-the-block whose meteoritic rise led to resentment within the ruling party, and all his various political machinations eventually led him to be shut out of the party and his new and depleted Liberal Party a minor party kowtowing to the LDP with no real gain. Having managed to create unhealable rifts within the LDP that led to its demise in the general elections this week, he is a great political mover and shaker in Japanese politics, but has never created a lasting legacy. You could say he is one of the pillars of the new DPJ, but it is difficult to brand him as a "heavyweight", either in the ascendant or the descendant... he is there full stop and until he gives away his sphere of influence to the open-source community under GPL license, he's going to be around. At the end of the day, Hatoyama created the DPJ in 1998 with his family funds, and seems determined to get even well-known opportunists with large spheres of anti-LDP influence to consolidate his party's still-tenuous hold on power.

Thinking about it for a moment, it is almost the opposite gameplan to the Obama campaign: build up the grassroots, win the election, leverage the new grassroots infrastructure to lead the people. The Hatoyama campaign has been: leverage the LDP's infrastructure that the electorate has handed to them on a plate, win the election, and build up grassroots support for the future.

To say that Hatoyama's position is weakened by Ozawa's nomination is superficial and the best thing Hatoyama could do. By cementing a Machiavelli squarely behind him in a kingmaker position, Hatoyama can work on upgrading the LDP's grassroots political infrastructure to serve his party and turn his party's historic win into lasting action.

Peridot short on Garmin's GPS future

The investment firm Peridot Capital Mgmt's blog - the Peridot Capitalist Market Blog - has made clear in a post today that it is short on Garmin and its upcoming nuvifone smartphone introduction.

In the medium term, as GPS handsets will become (even more) ultra-competitive and even more hardware players enter the already crowded field, the blog suggests to think of moving off already highly-priced Garmin stock and into other, more revenue-generating and less risky, investments.

Food for thought for the other GPS players out there, both in the mobile and auto industries. Quite a tough market out there considering Garmin has a zero-debt/high-cash position, has a reasonably high PE ratio, and is still getting this kind of rap...

JBOWS, cloud computing and security

Just been reading up on cloud computing, and came across the JBOWS (Just a Bunch Of Web Services) idea, that was more or less derided here by Joe McKendrick at ZDNET for being too ad-hoc as compared to SOA. As the sheer complexity of SOA goes the way of SGML (and the dodo), I think that on the contrary, ad-hoc stringing together of web services to usher in a new age of web-driven services is a good thing for organisations learning to internalise the skills of this new paradigm/environment.

The problem with turnkey solutions is that you do not get the benefit of experimentation or immersion in building organisational knowledge of the technology. In a short-term world of quarterly expectations, you want to focus on what you're good at and what makes you different. If cloud computing and software as a service is not your core business, and is not likely to be going forward, I agree one shouldn't think about it too much and go ahead with SOA and other initiatives.

But if your company is going to offer software as a service as part of the portfolio going forward, you need this JBOWS attitude to gain long-term expertise about cloud computing and web-driven online software services understanding more generally. It's a longer term play but that boosts organisational knowledge, expertise, confidence and capability, leading to market-leading innovation and distinctiveness.

The other thing I have been reading about are the drawbacks to cloud computing, such as the lack of control over your own IP and risk of decreased security for your cloudised assets. I would argue neither of these problems are issues that are dependent on putting applications, source code and other IP into the cloud as opposed to keeping it behind the company firewall, but a good example of mitigation for the risks of cloud computing for SAAS/PAAS/IAAS players is HP's new Cloud Assure.

This service offers cloud computing assurance services against the main 3 concerns of cloud computing: cloud security, cloud performance and cloud availability. It will be interesting to see the uptake of such services by blue chips or IT companies that have already made the leap. Maybe it is too early to tell or maybe we will need more major outages a-la-Gmail, over at Amazon (Amazon Web Services/Marketplace), Microsoft (Windows Azure) and Salesforce.com and for prolonged periods of time to highlight the need to pay for these services. This movement would help establish industry standards for cloud assurance, and ensure safety, consistency and compliance for the upcoming mass-scale adoption of cloud computing services predicted over the next 5 years. So cloud computing assurance looks more like an emerging market within the next 5 years, but with limited revenues until then.

Peer-to-peer network-based location on the mobile phone in the mid-term?

According to an article in the MIT Technology Review yesterday, Intel Research is working on a peer-to-peer protocol for exchanging data between handsets. There is already something called Bluetooth that could do this, but people are loathe to switch them on, although if this is something offered at the platform OS level, this could be very interesting.

My take on this: this could help devices position each other with greater accuracy and possibly transition to a more amorphous location-capture mechanism. As opposed to every single mobile device querying the mobile network/GPS satellites etc as they do now, you could have certain numbers of handsets pulling the data down and then broadcasting it to others in the vicinity. The broadcaster handset would perform triangulation on the fly at the ground-level, based on mobile phone density in the area. This could be sensed by measuring the ambient mobile phone signal strength in the area, or could be network-triggered.

Once the broadcast is finished, the mobile device would be released. The setup would be essentially similar to the DHT model used by the (much maligned) peer-to-peer network clients such as BitTorrent, Vuze, MuTorrent and others, where downloaders are much more frequent than uploaders. In this setup, every client is both a recipient of data, and a source of received data. The more clients are around, the faster the downloads. And you can switch it off at anytime, which apparently most people tend to do when they're done on the PC. If it was an OS feature, this would be more optimal as people would forget / have difficulty turning it off. Granted, this setup would probably work best in high-density urban areas with high mobile phone penetration rates.

An important element to this software feature would be anonymised & encoded handshakes so that although the data remains on the client, device-to-device anonymity would be preserved if hackers compromised a device to spread spam. Again, this is an issue that has already been looked at and taken care of to some degree with online peer-to-peer networks, provided tunnelling and other security features has been switched on.

This is mid- to long-term research here, but another example where the convergence of mobile and PC is allowing the porting of common software on the PC to better the mobile experience.

This may also help kickstart DARPA's attempts at building an interplanetary internet infrastructure to guide the numerous satellites and probes we have floating around in space right now, that are all guided by earth-centric - read centralised - communications. This is less than optimal as there is no failover mechanism, and an operational risk in case of temporary failure of data transmission from earth or "accidental" jamming from competing / unfriendly probes. DARPA's emerging DTN protocol, that I wrote about on my blog 6 months ago, may benefit from working with these groups as opposed to living in a box with the federal slushfund and creating stuff that is strictly limited to military applications.