Friday, January 07, 2005

Why Does H-P Software Suck?

Hewlett-Packard makes great hardware, but the company's commercial software is often of very poor quality. Pixelated images in installers and icons, untested conditions, memory leaks, the problems are obvious to even end users. Why should this be so?

I had an interesting problem with my home printer yesterday, an H-P Officejet d145. I was attempting to fax a document to Australia from the US. The number naturally started with the US international access code 011, followed by the country code for Australia, 61, and the city code for Brisbane, 7. I had tested the number from my telephone handset to ensure that I heard both an Australian ring tone and a fax on the other end, but dialing from the H-P Officejet consistently resulted in reaching a US number in the 617 area code. The only thing I can figure is that an H-P programmer decided without thinking to strip leading zeros from a dialed number. That would result in a connection to a US number in the 617 area code. Hmph.

H-P is an international corporation with engineers, programmers and presumably QA people scattered around the world. Surely someone would have tested this condition?

My solution was to attach a telephone handset to the printer and dial from there. Once the fax on the other end was reached, the printer could highjack the line by manually starting a fax transmission. "If it is stupid but it works, it is not stupid". It may, however, be annoying.

Saturday, January 01, 2005

Change in Virtual Real Estate

Happy New Year!

I have new digs and didn't even have to move boxes. With the recent changes at Tucana, I have started a consulting company for the first time in ten years.

You may now reach me at Software Memetics (http://www.softwarememetics.com/). My new email address is david@NOSPAM.softwarememetics.com (please remove "NOSPAM." before sending).

I look forward to staying in touch in the new year.

Group Think

My recent experiences with certain people who shall remain nameless have caused me to think about the "group think" phenonenon. It turns out that a very simple agent model yields the appropriate emergent behavior. My model is a drastic simplification of Thomas Schelling's investigations of social systems.

The model follows the behavior of agents which may have one of two possible "opinions". The two possible opinions are represented by a color, red or green. Turtles may change their opinion on each turn based on the dominant opinion of their neighbors, thus showing the effects of a simple form of the "group think" phenomenon.

The model shows how the opinions of individuals rapidly form large-scale patterns. The model itself, with documentation, is here.