Emil's Wicked Cool Blog

So smaller files are better off on flash memory? Small files:

• Source code
• Most web stuff (HTML, CSS, JS, etc., etc.)
• Lots of miscellaneous docs (word, powerpoint, exel)
  

Sounds to me like the best way to work on these things is to use flash memory as your primary work area and copy the files to hard disk only as backup. At least until Hybrid hard drives solve the problem...

There's a new performance feature in Windows Vista that has piqued my curiosity. The feature is based on pre-fetching (where programs are loaded into memory or in an area on your disk before you run it, because the system knows you usually run the program).

There are some other enhancements to the pre-fetch feature, but the interesting one to me has been coined as ReadyBoost. This feature uses flash memory (like a USB drive or compact flash card), and loads the pre-fetch information on the memory. The concept is that this will make your system faster.

Knowing that USB drives are slower than hard drives, I found the claim curious. It turns out that the way they access the data must be in smaller chunks. While flash memory is slower, the access time is much faster. There are several assumptions that vary the analysis fairly significantly, but my most pessimistic assumptions say that any file smaller than 100k is better off on flash memory. My most optimistic assumptions say that any file smaller than 500k is better off. Over those sizes, a hard drive will be faster (assuming that you keep your disk defragmented).

So I made the plunge and purchased a NAS. I got a Buffalo Technology product, but not the one that I previously mentioned. The difference between the Linkstation Home Server and the Gigabit Linkstation is DLNA certification for media streaming to a very few certified players, which seems to add about $75 to the cost. In 5 years I may regret it, but right now, I'm pretty happy with the HD-HG300LAN. I'm also somewhat future-proofed with the Gigabit option.

I moved all the data from my external USB drive and scheduled backups onto the USB drive (my external USB is now strictly a backup device). So far, so good. The downside is that the data transfer had to occur over the network since the NAS web interface doesn't allow any file manipulation. The firmware (v1.42) is pretty rough around the edges, and setting the security up hasn't worked so well, but everything else has worked as advertised, and pictures of Kathryn are now backed up.

I had seen a website that contained a list of login credentials for many of the "free but registration required" web sites. After more googling than I really would care for, I found it. www.bugmenot.com

In my last post I mentioned bethanycommunity.net web site. In getting it working, I had to get dotnetnuke installed and running on the root directory of the site. Doing this on godaddy was quite an interesting feat and required some changes to the dotnetnuke source code. I hope their latest version helps fix that problem. There are some instructions on doing this here. I found they got me close, but the installer was still a problem, and hence the (temporary) source code changes to skip over some of DNN's more dangerous install behavior.

Some people think that this is GoDaddy's problem and people should be able to create very insecure sites, but I side with GoDaddy that normally applications should not need to update their core configuration files and place files in their production directories.

I've volunteered some time to get a website up and running for a local organization. The Bethany Coalition was renamed to the Bethany Community Network, and was originally formed to help out with planning to accomodate more people moving into the area. Portland has an urban growth boundary that helps to confine population and prevent urban sprawl. In our area, the boundary was expanded to enable more people, but that means new infrastructure needs to be in place. You can read more about it on their (my) web site. Feel free to comment or email me directly with suggestions for improvement.

Why is it that after roughly 18 years of work on the Windows NT family of operating systems, Windows XP still forgets my desktop background picture?

As I had deleted the original file, I had to go spelunking today to figure out that it's copied to C:\Documents and Settings\[username]\Local Settings\Application Data\Microsoft\Wallpaper 1.bmp. The most frustrating part is that I could see the picture, but only when logging out/shutting down.

I got a chance to check out Ruby on Rails the other day. That basically explains it. Rails is a set of libraries that use the programming language Ruby to make web programming very, very productive. I've heard of defections to RoR from .NET and Java. I'm sure there are some, but the stories appear to be anecdotal. Anyway, it looks promising, and I hope the competition helps improve the productivity of all languages...

One of the more exciting things in Visual Studio 2005 for me was the idea of an object data source for grids. Well, I'm not doing a lot of hands on programming right now, and I finally got around to using it. I'm not that thrilled. Here's a snippet from a help file that is my current source of frustration:

The ObjectDataSource control will create an instance of the source object, call the specified method, and dispose of the object instance all within the scope of a single request, if your object has instance methods instead of static methods (Shared in Visual Basic). Therefore, your object must be stateless. That is, your object should acquire and release all required resources within the span of a single request.

This pretty much means that they expect an object to be a pass through to some other data source (XML, DB, etc.). I'm just capturing some session information right now. I find it easier to create/populate an object and databind directly then try to make an object that's specific to this problem. I think for the time being I'll go that route...

I don't think so. The theory/hypothesis is full of crap. We learned this in our finance classes in our MBA, but I don't buy it. There are several versions of it, but even the most weak of the arguments claims that technical analysis won't work. From what I've studied of technical analysis, it's trying to use numbers and history to predict aggregate investor psychology. A simple form would be that if a stock is at it's 52 week high, more investors think it's time to take profits than investors think it's time to get into the stock, so the price will probably go down.

Efficient theory claims the market is a machine that absorbs information and adjusts the prices within seconds to account for that information. Nothing else matters. I think the market is made up of people, and people have feelings, and stocks become either overhyped or forgotten. And when that stock hits it's 52 week high, I know I start to consider taking my money off the table...

I like Warren Buffett's feeling on the matter (which comes from Benjamin Graham). Take a look at this commentary also.

I've found Mad Money to be an interesting and entertaining show. I am watching more regularly now that my daughter has arrived and I need something to do while she sleeps in my arms, and with the summer here, this is not bad. Anyway, I don't always agree with Cramer (the host), but where I have agreed and taken action on his recommendations, it's worked out well. The latest trade is Valero Energy (VLO), bought before Q2 earnings announcement yesterday. Earnings came out as (Cramer) expected. Yesterday wasn't a good day for the market, but today both the market and Vallero seems to be heading in the right direction...

I recommend the show, but only if combined with your own research and analysis. Make sure you're not "drinking the Kool-aid".