> And as the Wall Street Journal notes, Windows is
> still ahead of Linux in server adoption.

They are looking at /revenue/, not count. See Savio's post.

http://weblog.infoworld.com/openresource/archives/2007/06/is_money_a_good.html

Red Hat's sales (volume) were _up 41%_ in the last quarters. Are we all about freedom or all about the money?

Zack-

I don't believe it's accurate that Windows is ahead. There is quite poor estimation of Linux's actual footprint, as we know of so many un-paid and un-tracked, un-barcoded instances.

If one goes by money -- revenues -- that's not satisfying to me. I'd like to talk about the Value of the Installed Base viewed as an opportunity cost of some other common system.

I ask you: what is the value of Google's Linux machine? Answer: somewhere close to their market cap. Do that again for all the fast-growers you mentioned, and then for the Fortune 1000 where most of the data center expansion is going to be Linux due to its deployability.

"Linux" needs to do better on tracking & PR. How about paying Gartner?

On the face of it, no-brainer indeed, although claim and counter-claim on the subject is as polarized as ever and the truth as elusive.

But I believe OSS has a more fundamental advantage. The traditional IT project methodologies have a torrid success rate. It is a trivial task to list $100M+ projects which end up not even achieving the core goals. OSS, if taken in the spirit in which it has evolved, involves end-users much earlier in the life cycle and (potentially) avoids killer issues like requirements defects and out-of-band use cases.

These are good points. I think that IDC and most other market research firms continue to measure the wrong data (e.g. license revenues) which is by definition not going to give an accurate picture of open source adoption. But Gartner does understand open source trends and they are giving good advice to their clients, as do other analyst firms like Forrester and newer firms like the451group, RedMonk, etc.

--Zack