Gartner findings on open source not new news on ebizQ - Open Source Software Up the Stack

« Don't believe "No open source down under," "No closed source on deck" | Main | OSS April 8 Podcast--Talking to... Mark Radcliff of the Open Source Initiative »

April 07, 2008
Gartner findings on open source not new news on ebizQ

I posted recently about the tendency of the open source software (OSS) blogosphere to take newspaper stories and analytical findings out of context... and then to really misinterpret them.

That recent post was about the many opinions on the web about the Australian government banning OSS and the U.S. Navy mandating it, neither of which was true. This week's nominee for most mixed-up open source message concerns a Gartner "finding" that OSS will take over the world on July 22, 2011 (exaggeration for effect!).

My first reaction to all of the breathless comments by others in the blogosphere about the future of the OSS movement was "I think I heard this before." And the reason was because Mark Driver of Gartner said basically the same thing right here on ebizQ on August 9, 2007. You can still listen to Mark's webinar at this link.

For a movement with source in its name, no one in open source seems to ever go back to the source. So, as I did with the claim that the "Navy just announced a bold IT policy," I went back to the source. Not surprising, Mark Driver was one of the authors of the "findings." The "findings" appears to be more a promotional piece. It links off the Gartner home page and links in turn to three or four dozen Gartner subscription reports written by more than 50 Gartner analysts, which I assume contain the meat behind the analysis. I think it is some sort of summary of all Gartner's open source related research in 2007. You can download the piece as a .pdf but I recommend you look at it online; although tedious, each of the three or four dozen links contains a mini-abstract that contains many more interesting opinions about OSS in 2008.

What makes the blogosphere fun, of course, is that half the commentators argued with Gartner over its analysts' "findings." So here are my disagreements with Gartner but I agree with most of what its analysts say:
-- Gartner makes a distinction between open and closed source vendors; as explained in my January lookback report, I don't believe that distinciton applies any longer (Microsoft was the last holdout and it got religion during 2007)
-- Gartner feels TCO on Linux-based systems might pass other systems in the future. The underlying subscription report probably has more detail but I don't believe you can break a TCO analysis down simply by the operating system that underlies the solution set being analyzed.
---------------Linux is UNIX; its TCO is going to track the TCO of all other applications to which you typically apply UNIX
---------------Linux is not Windows, it's UNIX; don't use Linux where Windows will suffice
(I know Microsoft will argue with me that Windows can do anything UNIX can do but that's not true yet and may never be true.)
--Gartner claims that 90% of enterprises will use open source "in direct or embedded form" by 2012. It would depend on how Gartner measures it but I would guess the number is either already north of 90% or that it might approach 50% by about 2020. It all depends where the AS/400 base moves and, if it moves to Wintel, how much open source Microsoft puts in its products. It also depends on whether you mean "free as in air" open source (called free software by the religous) or simply open source that is open source because of a business tactic.

Here's how I read the Gartner report:
1. Open source is not faster, cheaper, of higher quality, of different functionality, and so forth than any other software you might use. Software is software (and SaaS is not software but a delivery model for software).
2. Open source is a development model and a set of terms and conditions, some of which put limitations on your software's use that you might find unfamiliar and too restrictive.

There is no doubt that the OSS development model, which lead to an increased amount of open source code (I don't think Gartner mentioned Apache anywhere?) starting in the mid 1990s, will continue for a decade or more. If you're concerned about license restrictions (they are not in all OSS licenses), read the fine print.