Thanks for 4KSTACK explanation

Paul A Houle ph18 at cornell.edu
Fri Aug 5 13:06:59 UTC 2005


    I'd like to thank Dave and a few others for a good explanation of 
the 4KSTACKS move.  Anything that improves reliability for my server 
apps is welcome.

    I apologize for being less than tactful sometimes.  If I'm critical 
of RH,  it's because I like Linux,  Fedora and Red Hat and want to seem 
them succeed.  If I was commited to,  say,  Suse,  FreeBSD or Open 
Solaris,  I'd be making waves on their mailing lists,  not this one.

    Some people are concerned that there's a lot of negative content on 
this mailing list,  and I think that's correct.  We've seen arguments 
that boil on for weeks without resolution.  I think the problem,  
however,  is that there isn't a system in place to produce positive 
propaganda for Linux and Fedora.  Red Hat Magazine is pretty good:

http://www.redhat.com/magazine/

    but

http://www.kerneltraffic.org/

    is being updated less infrequently than it was and I wish that

http://www.kerneltrap.org/

    had 3x has much content as it does.  I went looking online for 
information on 4KSTACKS,  and by the limitations of current browsing and 
searching interfaces,  I found mostly people complaining about how 
4KSTACKS broke this or that,  and nothing that made it clear that 
4KSTACKS has the real benefits that it does.

    Fighting a headwind,  some Sun employees write good documentation 
for OpenSolaris in blog entries,  such as

http://blogs.sun.com/mws/

    Sun's blog aggregator incorporates a lot of material that's off topic:

http://www.opensolaris.org/os/blogs/

    but if it had some editorial effort,  I could see it becoming a good 
system that combines documentation at the user level,  kernel developer 
level,  and technical-friendly marketing in one place.

    LKML is great,  but it's like drinking from a firehose for people 
who have other job responsibilities.  There's a lot that can be done to 
cure the impedance mismatch between the information that's available and 
the information that linux admins need.

    Finally,  there is an impedance mismatch between Fedora and RHEL -- 
the connection between them is more intimate than a lot of people 
admit.  I had bad experiences with RHEL 3 on an x86 and had a lot of 
skepticism about the direction RH was going when the Fedora project 
started.  We was so impressed with FC 3 that we chose RHEL 4 for an 
x86_64 server we brought online.  If it hadn't been for Fedora,  we 
might have switched to a different distribution or OS.

    (And yes,  it's really a miracle that FC and RHEL are as good as 
they are on a platform as new as x86_64!)

    With better marketing,  ISVs might take Fedora a little more 
seriously and start testing products on it.  We knew well in advance 
that RHEL 4 was going to be based on FC 3 and ISVs could have had better 
products qualified for RHEL 4 faster if they'd done beta testing against 
FC 3.  Framing the Fedora discussion from a "half full" rather than 
"half empty" perspective could help with this.

   




More information about the devel mailing list