On Saturday, 23 August 2003, at 10:53 am, Jason Stubbs wrote:
On Saturday 23 August 2003 18:38, Stroller wrote:
I think he's saying Gentoo doesn't exist in terms of outsourceable o/s
support contacts. You can buy support from RedHat (and presumably SuSE)
that you can't from Gentoo. Gentoo's non-static nature & dynamic
updates also, I believe, make it difficult to justify in the enterprise
marketplace.
Actually, the devs are working on at least two systems to make Gentoo more
attractive to the enterprise. I subscribed to the gentoo-dev ML a couple of
days ago and already have read about a system to help admins with security
issues as well as a talk of "releases". Once that's done, Gentoo will become
viable for the enterprise and outsourcableness (is that a word?) will soon
follow.
If you'd been signed up to -dev for 3 days, you'd see that I'm also subscribed, because I posted to the group on the 21st.
;-]
I don't think anyone (or at least many folks) here would argue with the statement that Gentoo is a wonderful distro, and it's certainly getting better all the time. But I don't believe that GLEP #14 addresses the moving-target nature of Gentoo - if you `emerge sync` and find that you need a single security update, I believe it's possible that you will need to update more than one package; this surely implies the potential for a lengthy & involved etc-update process. Whilst this isn't a problem for me, I think it makes Gentoo unsuitable for the enterprise.
Also important to enterprise is support for commercial software, which is usually supplied as pre-compiled binaries. These binaries are, as I understand it, compiled against particular versions of libraries, and may be incompatible with distros using other versions of said libraries. Whilst manufacturers are prepared to say "we support Mathematica on Linux for RedHat versions 7.1 & 8" (and perhaps also for SuSE), they cannot do so for Gentoo, as they do not know which libraries (or optimisations?) you have compiled on your systems. This may or may not be covered by another GLEP, as I am aware of that Gentoo "release trees" are often discussed as desirable, but certainly Gentoo would have to do a lot of catching-up with RedHat in the enterprise to have enough market-share to justify many commercial software houses supporting it.
Perhaps the most significant factor is that Gentoo as a commercial entity isn't big enough for enterprise, and even claims to be going non-profit. I can see a good argument for choosing RedHat because they're big enough & wealthy enough to sue: although one might be unlikely to sue them, their size & commercial status mean they have a financial interest in ensuring reliability & software-integrity, and they have a corporate structure in place to manage their liability. If Daniel Robbins or a community of Linux hackers b0rk up your enterprise-scale systems, there's not much point in suing them, so they have no accountability to you.
At the end of the day, however, I feel that little of this is important. It would be nice for me if Gentoo was found suitable for use by smaller businesses, consultants and system admins responsible for a few (or a few tens of) boxes. But the enterprise market is something else entirely, and undoubtedly shares few of my needs - I suspect that were Gentoo to go chasing this market it would risk becoming unsuitable for my purposes. Right now, it is perfect.
Stroller.
-- [EMAIL PROTECTED] mailing list
