Nice article, except for one line where you say

"For example, hosted applications make more efficient use of hardware than on-premise servers"

I think that's a huge generalisation and certainly not true, just because the hardware is on your premises or not.

You could argue that the efficient use of the hardware is no longer the subscribers problem, but of course it is, because if salesforce do a poor job you'll end up paying more.

The other problem is of course if the company goes tits-up or gets sucked dry by some legal argument you can't easily take your toys and play elsewhere.

Finally, they've got you by the short and curlys once you've migrated to their platform. At least with SAP you could stop paying for support while you migrate away. In fact you could stop paying full stop and fight them in the courts if they upset you enough, while you're application continues to run. With salesforce they can always pull the plug.

We all no they have outages, but the issue is scheduled downtime.

Yes 2am EDT may be fine for a US customer, but no good fo EU customers, who are finding that as they get into work, they have no systmes for a few hours.

Hands up anyone who's responsible for a system with a reasonable number of users that's never, ever had an outage?

Anyone?

Okay, look. I'm biased. I work in the on-demand field (recently moved into the cloud from the on-premise model) and I have to say, the kool-aid is actually pretty damn' tasty.

SaaS/PaaS isn't a panacea; there are certainly situations where it isn't the best approach, but they're the minority. Quite a tiny minority. I also saw Benioff's speech, and the part that struck me most was when he asked for a show of hands from small, medium and large organisations. Roughly the same number of hands showed for each category. How many platforms can you name that work this well for the entire spectrum of client companies from tiny startup to global multinational? That's some serious scalability.

And, on the uptime issue, Salesforce are happy to publish their uptime stats and planned outages (which, as much as possible on a spherical world, are kept outside business hours) for all to see. Probably because they're actually pretty encouraging.

Check 'em out yourself if you like: http://trust.salesforce.com/trust/status/

Tcha.