The Home Server
The key word he adds to the discussion is database. Home PCs will become personal database servers.
This makes sense from an Always-On perspective. After all, what are medical, personal inventory, and security applications doing other than building databases, often huge databases?
But this brings up a key impediment (today) to Always-On, which is the price of software.
As hardware is mass-produced its price declines with Moore's Law. That's not true with software. In fact, complexity (or the ability to handle complexity) increases the price of software.
Look at the price of a Windows server license, for instance, next to the price of a Windows client license. Or look at the price of an Oracle license against that of any client program.
Martin notes in his piece that he's putting his databases onto a Linux PC, and that's a key point.
I've been noodling over the question of whether Windows or Linux will drive Always-On applications for some months. (Windows does have some advantages.) It's the question of price that is determinative. It's not so much that Linux is free as that the cost of Linux solutions move toward free, without the corporate overhead of Windows applications.
What are software companies delivering, after all, once you get the software? They are delivering support. Linux separates the cost of support from the cost of the software, so the initial cost can drive toward zero (especially in volume). That's a key Clue.
An ISP, or telco (is there any difference), or some other company, can earn itself monthly maintenance fees from the Always-On software in your home server, which they can maintain remotely.
I think that's the model for the future. And you won't get it through Windows.
If I was a betting man – that Linux home server will be running java with struts and Itabis or Turbine, with some coolio basicPortal frameworks – all wrapped inside a social software and personal publishing embrella. All open source.
All designed for talking XML and serving up one's digital lifestyle.
OK – so I was off by 10 years. Maybe building the MediaBar back in 95-96 was too early. maybe spending all my money on research was EXACTLY the right thing to do.
I guess we get to find out now.
PeopleAggregator is back on it's feet and it's sister – WebOutliner – now has accounts. Various pieces of the puzzle – from the Laszlo rich media front-ends, to JahShaka high-end editing tools, to pre-built jukeboxes and photo albums Portal software or knowledge management utilities – are all falling into place. Open standards – like ATOM and FOAF – are taking off. This train is leaving the station. [Marc's Voice]