Software is easy

Software is easy

I just got reading "Dreaming in Code" by Salon co-founder Scott Rosenberg. Silicon Valley spits out a constant stream of self-inspecting books on high-tech. This book is probably one of the lesser valuable books in that stream.

The book tells the tale of Mitch Kapor's "Chandler" project. Along the way, it dredges up every bit of wisdom and superstition in the software engineering, such as open vs. closed, language vs. language, build vs. re-use, UML vs. agility, Hungarian coding, CMM, Brook's Mythical Man Month, etc..

The intent of the book is to prove the assertion that "software is hard", especially innovative software. However, they choose as their subject matter a failed project. The Chandler people found software hard because they were so bad at it. Creating new and innovative software is easy: start with a clear vision of what your 1.0 product looks like, then have a leader who drives it completion. The Chandler project had neither. Indeed, getting to 1.0 is generally the easiest part, convincing people to use your baby product is the hard part. Likewise, getting to 1.0.1 and 1.0.2 are painful, and getting to 2.0 is real tough as you realize that you are spending all your time fixing bugs for current customers while new innovators are designing their own 1.0s that will obsolete yours.

My own BlackICE project is a good example. It was new and innovative, and changed how the industry viewed the problem. Getting to 1.0 and wowing people was the easy part, but continuing to support customers and fend off competitors was the difficult long-term grind.

Anyway, it's the current "it" book in Silicon Valley, so you should probably read it to know what people are talking about,

Intel+Hafnium

I always find stories interesting like this one from John Markoff. The story is about how while Intel was behind on power consumption at 90nm, roughly equal at 65nm, they are better than competitors at the 45nm generation. One quote from the story:

Many executives in the industry say that Intel is still recovering from a strategic wrong turn it made when the company pushed its chips to extremely high clock speeds — the ability of a processor to calculate more quickly. That obsession with speed at any cost left the company behind its competitors in shifting to low-power alternatives.

I find it interesting because Markoff gets the story wrong.

Intel first announced this lead back in 2003 at technical conferences. Intel's domination at 45nm has been know for at least 4 years. The reason Intel sucked at 90nm is because while IBM gave their power-reducing technology to AMD, they refused to give it to Intel exceopt under horrible terms (like giving up the design to their x86 processors). Intel refused those terms because it knew it had AMD/IBM beat at 45nm.

In other words, Intel made the choice to sacrifice the 90nm generation (Pentium 4) because it knew it had everyone beat at the 45nm generation.

Apple announced their switch to Intel processors at the 90nm level. This confused many people because Apple claimed they made the switch for power-consumption reasons, but at the time, AMD/IBM had a better power-consuming technology than Intel. This is because Intel shared their roadmap with Apple, and that with a Controe+45nm, they were going to have AMD/IBM solidly beat.

What this blog is about?

I used to be a computer geek, but I sold out to The Man. Now that I see things from the point of view of The Man, I have become disapointed with geeks. Geeks pride themselves on "fighting The Man", but rather than proudly fighting for the rights of everyone, they have devolved into mindless populism. Specifically:
1. they aren't fighting to free people from authority, they are fighting to become authority
2. they aren't fighting with logic, but mindlessly follow anything that fits their prejudices
3. they aren't fighting to change the world in the future, but trying to re-fight lost battles of the past