Veitch Speaks: Sun's dream of Jini

You don't have to be an expert on connectivity, Java or the vagaries of look-up services to realise that Sun's plan to make its Jini technology the peripheral connectivity standard is pretty ambitious. Call me a cynic, but anybody who knows anything about the history of IT sees just how tough a target Sun has set itself, and how few of these projects requiring standardisation across a huge variety of vested interests ever reach realisation.

Sun spent last week in San Francisco talking up the technology. The specific promise isn't new but it has certainly never been achieved -- a universal network where you can see and access any connected device, from a home lighting system to a CD-ROM jukebox, by way of phones, copiers, printers, projectors and just about anything else. The technology is based on Java. Makers of this type of equipment build in Java Virtual Machines to devices so that they can advertise their existence and make themselves accessible.

In theory, you could control a projector in the boardroom over the Net - so you could address a London conference from a Sydney hotel room. On a more trivial level, Moulinex could build a Jini-enabled toaster that you could set to medium-brown, no crust from a mobile phone on the Sevenoaks train at the end of a hard day. Techno-enthusiasts should note that not even Jini can assemble the bread and cheese yet.

The concept is as old as networking. IBM and Wang talked about this type of stuff back in the early 1980s, while consortia such as Salutation went so far as to spec the whole thing up. Most famously, Microsoft, at the zenith of its Windows-everywhere fervour, discussed Microsoft at Work, a plan to connect office devices that was to have been followed by Microsoft At Home, which would do ditto for consumer items. Much talk of Windows for toasters followed, printers got a bit easier to use, but not much more was heard.

By coincidence, Microsoft is trying again with something called Universal Plug and Play. Hilariously, this is being billed as an attempt to make connecting PCs to home hardware as easy as it currently is to get a printer or some-such business peripheral up and running. Next time you have problems with the pesky setup of your new iron, at least there may be a wizard to hand.

Well, hooking up a wide variety of devices for sharing across a network would be useful, so it's quite right that everybody should be trying to do it. For vendors, it's big stuff -- 98 percent of chips aren't in desktop computers. But that doesn't mean it's going to happen anytime soon.

That's true even though Sun can point to a decent roster of so-called support: Novell wants to write a tailored directory service; Computer Associates wants to provide system management tools; a range of other blue chip companies, such as Hewlett-Packard, Epson, Toshiba, Oki and Canon, could help make devices; and others, such as Ericsson and Nokia, could help at the transport level.

It's easy to come up with blue chip supporters for such projects. Once, Go had a huge list for its short-lived PenPoint operating system and the OpenDoc object group had Apple, IBM, Novell and WordPerfect behind it. Remember those?

So get excited about the potential of these architectures by all means, but remember that there will not be any deployment this year. And maybe there never will be.