You can never be too rich or too thin and it's hard not to be impressed by Pano Logic's arguments for a Size Zero desktop. Announced last week, the world's 'first truly virtualised desktop' has no software at all - operating across the network to a server-based instance of Microsoft Vista or XP. This makes it more secure (no local software to get corrupted), cheap (no local software to upgrade) and green (low power consumption). Pano Logic reckons it cuts total cost of ownership by 70 per cent - saving as much as $3,200 per desktop over three years.
In earlier times Pano Logic's Zero desktop might have been called a 3270, a VT100 or an X terminal. And the server carrying the software and the applications would have been called a 'mainframe' or a 'minicomputer'. But that was a long time ago and although memories keep getting bigger, they also tend to forget more quickly. There have, of course, been numerous attempts to re-invent the terminal, from Larry Ellison's Network Computer on down - but none of them have, so far, succeeded in ousting the desktop. If Pano Logic's new version of the terminal works as well as the company claims, it might just be the one that makes the breakthrough.
While it is all very worthy, the Zero desktop raises important issues for software developers. The combination of increased bandwidth in comms networks and the trend towards server virtualisation and thin clients means desktop software looks likely to decline and, perhaps, disappear altogether. The massive industry which, over the last 30 years, has grown around developing and supporting desktop software will inevitably suffer as attention shifts to server-based applications. More importantly, the level of skill required to design and build multi-user, multi-threaded, re-entrant code for server applications is significantly higher than that needed for single user dedicated desktop applications.
If the virtualisation software and the operating system work properly then application developers could be insulated from some of the added technical complexity. But this is a big 'if' and past experience suggests multi-user, server-based applications will be harder to build. They may even be beyond the skills of those who have only worked on single user desktop applications.
It is just 30 years since the late, great Professor Dr Edsger W. Dijkstra noted in his informal (and controversial) presentation to the 1977 IFIPS congress that 'microcomputers are not great'. Not only did he suggest that 'the simplicity provided by a large central store is perhaps even more striking than the simplicity provided by a fast central processor' he also railed against what he called 'computniks' - bad programmers.
"The microprocessor will provide at low material cost a delightful outlet for the uneducated computnik with nothing better to do, but the possibility of mass production of those infernal gadgets carries with it the danger of draining our intellectual powers to an extent that no society can afford."
Dijkstra, like many great men, was not always completely right, but in this case it is difficult to find fault with his argument. An awful lot of time and energy has been spent on building - and attempting to use - rubbish software on PCs in the last 30 years. Perhaps a return to the terminal/central server model of computing will herald a new age of better software and the demise of the 'computnik'. Then again, it was dissatisfaction and frustration with centralised computing which enabled the PC to thrive in the first place... ®