I happened to be looking through my awstats output for this month when I noticed something I found a little odd, or at least worth mentioning. I’ve been examining these types of logs for years now and there’s so much data there that perhaps I’ve missed it before, but in the list of search engine bot statistics, I noticed an interesting relationship. So far for the month of March, Googlebot has hit my site 409 times for a total bandwidth usage of 3.10MB. Yahoo has hit my site 719 times for a total bandwidth of 6.39MB. Last, but certainly not least is the MSNBot coming in with only 590 hits, but a whopping bandwidth consumption of 11.89MB! Now, there are a number of others in my list, that each consume far less than these big three. Looking back over the past several months’ data, Google and Yahoo appear to be rather even in this measurement, but MSNBot is the clear leader with sometimes 15 times the bandwidth consumption! I don’t exactly know for certain if it’s fair to interpret these numbers this way, but it occurred to me that Microsoft is trying extra hard to beat Google, and perhaps this increased bot activity is evidence of it. Anyway, just makes you wonder how well it’s going for them. ;)

I’d be interested to know if anyone else has seen this activity or has a better explanation for it.

I’ve long been interested in parallel computing, having become intrigued by the subject when reading about the first “Beowulf” clusters in use years ago. I began experimenting with PVM on my Linux and FreeBSD machines at home years ago and was pleased to build a “cluster” of two machines which could process POV scenes more quickly than either one by itself. It was a nice achievement, but I recognized that parallel computing did not hold much promise for my day-to-day computing experience.

As time went on, I discovered MOSIX. I began playing with MOSIX and was thrilled to have a cluster that did not require special apps compiled against PVM or MPI libraries. MOSIX made parallel computing transparent to the applications running on top of it. While there are still places where the PVM or message-passing models may work better, MOSIX greatly simplifies this parallel computing process and opens it up to every Unix binary. (that can be forked)

As a Mac OS X user / administrator today, I have been playing with Xgrid since Apple released it’s technology preview. Based on the initial marketing hype and my experience with MOSIX, I was greatly anticipating an amazing advance forward in parallel computing in a system with the advantages of MOSIX and the signature ease-of-use that Apple products bring. I have since discovered that while Xgrid simplifies some things, it is certainly not as trivial as I first expected. In addition, there doesn’t seem to be a whole lot of examples for new uses for Xgrid. Doing a search for Xgrid applications, I discovered the same basic POVRay example that I used to test my PVM-based cluster years ago.

Now I’m not saying that Xgrid is bad. In fact, I’m very glad to see Apple working on such technology, and I’m certain that it will continue to improve. Perhaps we’ll see advances in this area for Leopard. In the meantime, wouldn’t it be great to have MOSIX ported to Darwin? I’ve never even glanced at the MOSIX code, so I have no idea what this would entail, but the MOSIX team says it has been ported to several Unices, and they seem to indicate that it is fairly portable code. So, let’s do it. Are there any other Mac OS X / parallel computing enthusiasts interested in such a project? Am I way off here? Maybe Xgrid is more powerful and flexible than it appears to me. What are some of you doing with Xgrid? I’d love to hear about it.

In the meantime, I’ll continue to use Xgrid for POV and Xcode’s distributed build technology for compiling large projects. But if I could speed up certain tasks on my older G3s and G4s by tapping the idle CPU power on my newer boxes, that’d be marvelous.

I just added a new “Trail Log” feature to replace my old, very simple one. Since moving my entire site’s content into the WordPress engine using static pages, I needed a new solution for a trail log. In this case, I decided to use the WordPress blog engine itself to record my trail log entries. I established new template files to handle the displaying of this new blog category. The result is a very functional, attractive list of trails I have hiked. It will be much easier to keep up to date and I look forward to filling it up with info. You can click the “Trail Log” link on the right, or click here.

With the recent upgrade of my cell phone from the Sanyo 4900 (One of the most reliable and functional cell phones I’ve used) to a new PalmOne Treo 650, I tested my BibleReaderME application on it using the IBM J2ME JVM for Palm OS. I’m happy to report that the application functions very well on the Treo. I’m happy to add another confirmed cell phone model to the list.