>>> On Thu, Feb 14, 2008 at  9:55 AM, in message
<[EMAIL PROTECTED]>, Sven Schuetz <[EMAIL PROTECTED]>
wrote: 
-snip-
> If that does not work for you I'd be interested in your feedback, as any
> issues can be addressed in upcoming releases. Or is there a problem with
> the Redbooks in general?

There are problems with Redbooks in general, but I don't know that any of them 
can be fixed.  (Having participated in the creation of one directly, and others 
semi-indirectly, I'm all too familiar with most of them.)  They tend to come 
out too late, for several completely legitimate reasons.  They tend to not get 
updated at all, let alone very timely.  Hence, they tend to become obsolete 
quickly.  I'm going to raise a ghost now that will make Mike MacIsaac groan and 
run for cover...  Sorry, Mike.

Linux is evolving rather rapidly.  Mainframe Linux is evolving even faster.  
Writing books to cover topics in such a field is guaranteed to be "too late, 
too soon out of date."  A wiki would be much more appropriate for this field.

We do have a wiki just for mainframe Linux at wiki.linuxvm.org.  It's never 
been announced because we wanted to have a significant amount of "seed 
material" in place so it would be worthwhile for people to visit it, and be 
motivated to contribute.  We could never get more than a couple of people to 
add anything, and it's been languishing ever since.  I'm still puttering with 
it in my copious free time, trying to figure out why people have to create 
accounts and login just to *use* the wiki.  But it's been a low priority for me 
since I know it isn't going to make much difference if I get it working that 
way or not.

I guess I'll hijack this thread even further by taking the opportunity to give 
some credit to Rob van der Heij of Velocity Software.  When we first set up the 
wiki, we were seeing some really horrendous performance.  Rob dug into it 
(using Velocity's ESAxxx tools), and helped us out, but things still weren't 
looking good.  Then he dug even further and discovered a couple of (arguably) 
bugs in a package bundled with the wiki, and IBM's JDK.  It was issuing a bunch 
of 10ms sleeps, keeping the guest in queue  causing z/VM to treat it as batch 
workload, not interactive.  It was a really nice piece of work that will 
improve things for all mainframe Linux customers.


Mark Post

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to