Let's say that if Lotus did a Notes client port, that they'd convert 10 million of those clients to Linux in the first year. Sounds good? Not to Lotus it isn't, because those 10 million have already paid for the Notes client license, so there's no new revenue. Maybe 1 million of those users might be lost if there were no Linux client. Lotus will only do the client if doing it brings in new client licenses or stops the loss of licenses to someone else. They won't do it if at the end of it they have the same number of licenses as they would have had if they didn't do the port. So the port has to be seen as marginal to Lotus.

If the client port is marginal, then so is the Designer port; it's probably the same amount of effort all over again, and for what? It would save few lost Designer users. It might create new users of Linux servers, because they can then create Domino Web applications without any Windows. Lotus will be interested in that, but how many new server and Designer client licenses will they sell, and how many will they stop the loss of? Not enough, in my mind, to warrant a Designer port.

Florin writes back
Florin responded with some more comments, first on the costs question:

The infrastructure costs are of two sorts: upgrade licenses and work done for upgrading. While the work costs may be similar, maybe slightly higher for Linux, the license costs are clearly smaller for Linux. I think that the combination of the two sorts of costs yields an only slightly higher cost for moving to Linux than for upgrading Windows.

End user training is indeed a significant difference. But not if your upgrade team manages to make the Linux installations behave and look a lot more Windows-like than out-of-the-box Linux installations do, which is possible.

The tricks with the file system are not a must, as long as the new workflow engineered by the switch/upgrade team doesn't require such things, in my opinion.

From my own experience, switching to a Linux desktop takes no longer than a few weeks to any Windows power user. And after that it becomes a delight! As for non-power users, they will find their way around in just a few days, given only a few hours of training, if the Linux box is properly configured.

Mick takes the floor
I think the costs issue really will come down to the specific installation, but I still think that they are huge. Florin says that switching to a Linux desktop takes a few weeks for a power user and a few days for less demanding users, at the cost of spending time on the setup. To my mind that's a big issue. Power users to most companies are people who need to keep productive; they tend to be the leaders on the ground, the people who make things happen, and who create and innovate. To most of these people, their computer is an extension of their fingertips and their brain, and messing with that is a serious business.

Remember, most of them don't work in the IT department and are not IT specialists by profession. Those who are IT specialists are likely to see a change from Windows to Linux as an opportunity, but those who aren't will see it as a real barrier to progress. I think that such a loss of productivity among these people is a serious issue. It's a significant hidden cost, but one that should be attended to. And add to that the IT department time that has to be taken to create a customized non-out-of-the-box Linux that looks and behaves like Windows, as Florin says is possible, and remember to add all of that into your helpdesk training and support budgets. For it to be effective, it really does need to behave exactly like Windows, right down to all the keyboard shortcuts.