OSI enforcement?

Tzeng, Nigel H. Nigel.Tzeng at jhuapl.edu
Thu Jan 10 11:27:37 UTC 2008


Rick Moen [mailto:rick at linuxmafia.com] wrote:
>Quoting Tzeng, Nigel H. (Nigel.Tzeng at jhuapl.edu):

>> Except that you assert that the legally equivalent binary that is made
>> available via CentOS is not exactly the same set of bits and therefore
>> is disqualified as a replacement in your scenario.
>> 
>> Which is picking really tiny nits.

>Not to many enterprise users.  Try telling them that CentOS is exactly
>the same as RHEL without the latter's bundled support, except compiled
>in a vaguely similar manner in a vaguely similar build environment, and
>you will be thrown out of the corner office onto your ear -- and that
>includes situations where the company has _no_ desire to ever invoke
>bundled support.

Unfortunately I don't have anything besides anecdotal evidence but SFGate
did choose CentOS over RHEL.  I've also been part of OS trade studies 
where deployment costs trumped RHEL support and other Linux servers
distros won.  This was for highly mission critical systems where I was a RHEL 
advocate.  Although I would also have advocated CentOS instead had it been 
an option (for whatever reason it didn't make the downselect) since it IS 
functionally equivalent, equally zero cost to other alternatives with exactly 
the same RHEL advantages except support.
 
Within the DoD community, not picking RHEL/CentOS or SLES is sub-optimal
90% of the time.

>> Except you can.  There's a reason that CentOS ranks above RHEL on
>> distrowatch and google search trends. 

>There's a reason Red Hat continues to sell copies at full retail even to
>firms that have no desire to use its support services.
 
But they don't do it for $20K per copy like I was paying for Solaris with 
Silver support.  Neither does Sun.
 
And...um...anyone who pays full retail for RHEL for any kind of significant
deployment isn't even bothering to pick up the phone.  Yeah, if you only
need one or two copies you might pay full retail.  Although if you only need
one or two copies you often don't NEED enterprise level support.

>> I am also unsure of your point.

>OK.

>My point is that "copylefted commercial software is something only
>stupid people pay for, unless they're paying for bundled services"
>(paraphrased) is factually inaccurate.

Well...I try not to use the word stupid in even semi-professional discussions.
 
At least not on public mailing lists. :)
 
Your statement may be correct in the general sense given folks will pay for a
commercial license for things like MySQL because they don't want to open up
their source but I don't think that RHEL is a particularly good example.  Especially 
since nobody wants to buy RHEL 5 Update 1 gcc as a standalone product and 
wouldn't likely pay any money for it even if RH offered it as a standalone product
like Sun does for it's compiler suite.
 
However, given that gcc is free even Sun Studio now has an "express" version for
free.  If you want "premium" support it's $1200.
 
I wonder how many folks pay $1200 for Sun Studio 12?  More importantly, how many
folks pay $1200 per seat for Sun Studio out of their own pocket or project budget?
If it's someone else's pot of money I can see getting the support as a hedge or maybe
one copy out of 20 I might get support for on a mid-sized project.  I don't see dropping
$24K to get 20 seats though like we used to have to do.
 
And its not copylefted nor, as far as I rememeber, open sourced but Sun's effective
sale price for that piece of software is now zero.
 
Nigel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20080110/a91e9686/attachment.html>


More information about the License-discuss mailing list