It wasn't me. You can't prove anything.


2006-12-02

Written 2006-11-30
Brain Drain

I did some more yum work today. This time I was trying to get the client working. I downloaded one rpm after another and finally had most of the pieces I needed. I installed and gathered all the modules in the correct places. For some reason (my configuration I bet) yum wants to install all over the machine in stead of the directories it is supposed to.
After I got things working on the client side, I just couldn't think any more. The rest of the day, my brain was at half power. It just plane wore me out getting that stupid thing to work.
They just plane shouldn't try to mix architectures. they should have just bitten the bullet and made 64 bit systems incompatible with 32 bit systems. This half way stuff is killing me. Can you even get a 64 bit windows application to work in native 64 bit?

Today
Friday afternoon I got the RHEL repository to work. It ran through an update and did nothing. That is because the files are not newer versions of the files already on the system. I kicked off a repository download that I really hope works. I'm using CentOS repositories. I would love to use CentOS repositories in stead of Red Hat simply because of the RHN login that I have to deal with office politics to get.
It is amazing that this has taken this long. I'm trying to figure out how to upgrade Fedora Core 5 to Fedora Core 6 via Yum. I know it is possible, I just can't get the magic link that points me to that guy who did it and bothered to write it down. The world needs more of those folks. I'm trying to be one myself.

1 comment:

Anonymous said...

Kelly, I must say that you really struggled to get a plain system to work. I took the fc6 dvd, went through the install, and where it gave me the install or update, I chose the latter. Worked just fine.

Consider that with the newer technologies such as multi-core, xen, etc. that complexities are going to expand by a power of two.
As example, when there was 100 interactions, then we had to look at 100 by 100 possibilities of bugs (10000).
When we introduce xen or multi=core, and that number goes to 200 by 200, then we are at 40,000 possible interactions.
Can we generate code that can handle systems that are expanding in complexity?

That power of 2 example gives credence to your comment that we should seperate 64bit systems from 32 bit systems, anywhere where we can generate isolated islands of logic with few interfaces. That concept, is similar to the one door in and one door out philosophy.