Yusuf Goolamabbas

The blogosphere is abuzz about the news that Apple plans to ditch IBM and move to Intel . WWDC just got a whole lot more interesting.

I am still ambivalent about this news. Whilst part of me wants to believe that Apple will “switch”, this type of news has come up far too often that it would seem like a ploy to get Apple some leverage with IBM.

Will need to seek some rumor sites who are going to transcribe the keynote in real-time.

My previous post made me think about some sacred cows in the community project I am associated with. A friend identified 4 options

1. Maintain status quo
2. Build a new team to learn infrastructure
3. Don’t be hung up on the existing toolchains, if need be completely change the toolchain if people are available with skill sets in the new toolchain
4. Outsource the infrastructure

• Option 1 is a non-starter since its obvious that status quo is not working
• Option 2 is only possible if the new team hits the ground running and doesn’t need to take any resources from existing players.
  Training the new team is not an option since that itself takes a massive time commitment
  Anybody who says otherwise has never worked in the trenches
• Option 3 is conceptually similar to Option 4. People who like to do backseat driving would feel uncomfortable with this since it represents a change but its obvious that they have no clue to the cost of option 1 and 2. For them, I can only say
  

  Change is the only constant

However, I would like to challenge a “sacred cow” which is

Should this community project exist at all today ?

The circumstances in which the project was initiated and the circumstances today are very different. Its possible that resources of the ‘community’ are better utilised by disbanding the current project and having the alumni of the project move to other projects which have more ground support and mindshare.

Thanks to the cool Dashboard in the WordPress 1.5 series, I found out about the 1.5.1.2 security release. The fix was very quick though I would really like the WordPress team to change their release mechanism. The tarball should be named
wordpress-version.tar.gz and it should untar to wordpress-version. Currently they use latest.tar.gz and it untars to a ‘wordpress’ subdir

Also, via Richard Brown of VIA marketing I got to know about the VIA C7 (Esther core) launch. There has been a lot of print on its low power,low cost and how VIA plan to make a big buzz in laptops and price sensitive countries. In my humble opinion, I think VIA should highlight the work OpenBSD has done on supporting the crypto hardware onboard the C7. In June 2004, Theo De Raadt of OpenBSD fame mentioned that he was getting 800 RSA signs/sec on what I guess would be pre-production hardware.

VIA should look at the availability of the ‘NX’ bit and the fast crypto command to create buzz for both WindowsXP support for NX and reduced risk of viruses as well target the infrastructure operator and suggest that OpenBSD+C7 boxen can become SSL proxy boxes for different services.

Ali informed me that he was not able to trackback to my last post. A bit of googling indicated that there was some issues with trackback particularly on hosts which disable fopen which is what my webhost Dreamhost seems to be doing.

So a few more google searches later, I figure that it shouldn’t be that difficult to upgrade WordPress to version 1.5.1.1. I plunged ahead and voila, a shiny new look which I kinda like, trackback works and Ali says that I generate cache-friendly headers.

A big thank you to the WordPress gang

For a long time I have assisted a community in building out some infrastructure so that they would have an online prescence. Over time, focus has shifted to generating content and publishing that whilst the number of volunteers working in the trenches has diminished substantially. It could be my fault that I was not able to attract a core of people who believe in the value of infrastructure. I’ll accept it as my weakness but unfortunately it puts a lot of burden on me and I no longer find the motivation to work on this.

Some of this lack of motivation has to do with differing views with various key players, some volunteers who have great heart but don’t bring anything substantial in expertise to the table, others who want to morph the vision so that it is “politically correct” . A lot also has to do with the nature of communication, it’s very rare to meet other volunteers face to face and timezone differences make it difficult to do voice communication though I have tried often to call people instead of emailing them. Sadly this isn’t done by other old-timers.

It seems that there is an illusion that the infrastructure is more capable than it appears and since things have been working for a while, people’s attitude have been “No point in working in the trenches, It’s not something end-user visible”

It’s a struggle working with people who know the value of everything but the cost of nothing. I just don’t have the energy to do this anymore.

He has said it in a way that just says “wow”

Been busy with work related things hence the quietness on the blog front. I need to spend some time figuring out the best way to upgrade my WordPress installation.

I had access to a Sun v40z Quad Opteron box in the office for some time. Very nicely engineered box, noisy as hell (12 fans I think). Ran some of my workloads on CentOS 4.0 and Fedora Core 3. Wanted to install Solaris 10 but ran out of time. I also have a socket 754 AMD64 box (MSI K8MM-V). The mobo is very nice and compact and runs very quiet. It’s running FC3 for now. I plan to install FC4 on it soon and see if there is any perf improvement with the toolchain compiled with gcc4

The AMD64 box with 1GB RAM and 2x120GB SATA cost around HK$ 3500 (US$ 450).

After a long break, got back into playing with Cherokee . Alvaro’s been doing a great job on it. He’s got event ports support now, epoll works more reliably now and I am very grateful that he took my suggestion and enabled O_NOATIME support in it. I think Cherokee needs some online documentation and then some evangalism to get more users using it.