Thursday, May 15, 2008

Microsoft and the OLPC deal

I have been following Microsoft's courting of the OLPC project on Slashdot. Ever since my involvement with the FaunOS project, I've slowly convinced myself that it will be "portables" that pave the way down to ever cheaper, commoditized computing. And Linux has always figured prominently in my speculative conceptions of likely, future computing landscapes. So now what to make of Microsoft selling $3 copies of Windows [on OLPC]?

I believe a simple rule of markets is that once a producer breaks the price of its product in one market, it will eventually have to break prices for the product in other markets as well. (I tried googling for some such succinct rule but, alas, I couldn't come up with the right query terms.) From a long term investor's point of view, this cannot be good news for Microsoft. It is a short term coup for Redmond: having OLPC run on Linux risked cultivating a new cadre of non-Mr. Softies. A $3 copy of Windows thwarts that threat: few children would opt for the Linux version of OLPC--for obvious reasons not worth enumerating. But this sets up a bad precedent. Will Microsoft fight Linux with $3 licenses wherever Linux enjoys a clear cost advantage? My guess is that wherever a user interface is involved, Microsoft will fight to keep market share. Today the battle is staged on what I'll sweeping characterize as ultra-portables, like OLPC and Eee PC. (Okay, maybe it's not a full scale battle, yet, but hear me out..) From Microsoft's perspective, the distinguishing characteristic of the ultra-portables is price: that they are portable is irrelevant. But hardware prices inexorably fall, and we can safely assume to see hardware the size of an ordinary laptop selling at today's Eee PC prices. Cheaper hardware means either a free OS like Linux, or significantly lower Windows licensing costs. This deal is an example of the latter, and is a reminder that Microsoft's fat margin's and pricing power are receding.

Now don't mistake this entry for a Linux fan's attempt at finding a silver lining in a deal that effectively ditches Linux. In fact, I am very worried. Imagine a world in which Microsoft still makes a decent living selling mostly $3 operating systems. Somehow they figured out how the business model could work--e.g. leverage their monopoly position less obviously, or make up the lost margin through greater volume. Whatever. Microsoft could even be a lot smaller company and still control the desktop.

I doubt that is how history will unfold. Linux will win the desktop, but it'll likely be a long and dirty battle in which Microsoft will make a lot of noise while quietly imploding as its pricing power wanes and margins collapse.

Friday, May 9, 2008

Idea: Wave power without moving parts

When I was an undergraduate in applied physics, I had this idea about using the conductive and dielectric properties of sea water to convert wave motion directly into electricity. I recently googled the topic and came up naught. (May be because it's a bad idea?)

Any way, the idea comes from 2 observations about capacitors.
  1. It takes work to slip out a dielectric slab sandwiched between charged capacitor plates.
  2. It takes work to increase the distance between charged capacitor plates.
Now if we can set up a large scale capacitor with the ground plate placed horizontally underwater, close to sea wave troughs and the "positive" plate well above wave crests, and a suitable timing mechanism to apply and draw voltage to and from the capacitor as the wave enters and leaves the contraption, then some of the wave power can be converted to electricity.

You don't actually need capacitor plates: a mesh of conductive wires should suffice. The power generated by such a contraption is proportional to the area of sea surface covered. But because of its simple design (all the smarts and expensive parts can be centralized), it should scale well.

Numbers? I'll try to add a back-of-envelop calculation, later..

Addendum: 3-20-2010

I'm beginning to think this might not work as decribed. Even if it did, from a practical standpoint, the wires would soon be covered with debris, barnacles or other insulating stuff. A better approach might be to create a giant, barely floating, sealed "waterbed" filled with a suitable dielectric liquid. Two wire meshes spanning the top and bottom walls of the bed from the inside would be used to create the effect of an array of capacitor plates. A component of the [gravity] wave hitting the side of this bed would continue to propagate through this bed, creating a moving bulge through it as it passes. As before, an adaptive control mechanism would time a voltage increase across the cascading, virtual capacitor plates ahead of the approaching bulge and draw power as the wave leaves the region.

Thursday, May 8, 2008


Years ago, before marriage and kids, I used to keep journals. I still write, on occasion, for my own pleasure, but in recent years this writing is mostly scattered across email conversations I've had with family and friends. The nice thing about the meat-space journal was that at least it was all accessible in one place (not to mention all the other obvious advantages of a real notebook). The downside of writing in a notebook, well.. is that only you read it.

I don't know if this journal will find an audience. I hope it will, but if it doesn't, the thought that my kids might one day read whatever crap I wrote here is enough to sustain me. That the digital footprint we leave behind in our lives is for posterity, I think, is an under-appreciated facet of online activity.

This will hardly qualify as a blog. Little that I write, I suspect, will be timely. Nor will it likely be focused on any few topics. This is more like an online notebook..