Wednesday, January 21, 2004

Why I don't blog politics

Well, for a lot of reasons. But my post about John Kerry a few weeks back illustrates one: I was very, very wrong. I'm not alone, of course, most people wrote off John Kerry. But still, why should I add one more ignorant voice to the clamor?

Tuesday, January 20, 2004

The Distance That Connects

An interesting thing to consider about the four hero types is the ease or difficulty with which each could be made part of a game, say an online computer game. There's no lack of muscle hero-type games, if we broadly consider this type of hero to be any type of head-on skill match. Indeed, this is the vast, vast majority of games. The self-control type, it seems to me, is much harder to simulate, even if we wanted to. That caveat isn't trivial; for cultural reasons, I think, this hero type is less attractive than others, as it lacks the tangible evidence of victory. For that very reason, though, it seems like an interesting challenge.

Trickster might be more possible, indeed, since the challenge moves from simulated strength and skill to the actual tricky (or not) mind of the player. In some sense, this is a purer form of interaction. Similarly, the prophet, at least on the mundane level of discovering secrets, is possible. However, the sort of customized, individual mysteries makes this sort of game scale very poorly. Infocom had some interesting mystery games back in the golden age of text adventures, and there's Law and Order computer games now, but it seems much harder to do in a shared persistent world.

I want to go back to the self-controlled hero. And real hero characters are mixes of the four types, so all heroes to some extent share this issue. The self-control must be self-control over something, or rather against something. The archetype here is Siddharta Gautama becoming enlightened as he is tempted or frightened (or not) by demons, or Jesus in the desert turning down Satan's offers of riches or temporal power. Could these tests, of refusal of appetite, or endurance of pain, be meaningful in an interactive game? Ironically, this is more powerful when witnessed at a remove reading a novel or watching a film. We ache with the character's self-imposed pain ("Must... hold... on...!") or sympathize with hunger. Yet without that remove, moving into first person, it suddenly loses dimension and thus loses power. I call this the "distance that connects," the paradox that by removing the conflict of the hero from the witness, it becomes more empathetic.

Monday, January 19, 2004

The Hero Types

Buried in a footnote to an appendix of Lewis Hyde's Trickster Makes This World is a fascinating point about the nature of heroes. He makes a claim that the trickster is not a hero, contrasting him with the two other types. I think he's partially right but only partially. Hyde's two types are the "muscle-bound" hero who imposes his will on the environment (for indeed, it is usually a man), and the hero who succeeds because of self-control. I think in the latter category we could also include those who are true to honor or other ideals, in any case strength emerges as a consequence of the denial of more earthly urges (a common theme in Hyde's book).

I do think that Hyde sells the trickster short, however; in my view, trickster is a hero as well, although true, he (again, usually tricksters are male) neither through imposing his will on others nor surely not by self-control. In literature, characters like Slippery Jim DiGriz or Miles Vorkosigan are tricksters; they achieve their ends through the assumption of identities, by stealth, by the confounding of their foes, and, in tune with Hyde's analysis, by slipping the unslippable trap. They may not be quite as ethereal as Coyote or Anansi, but then neither is the usual muscle hero a Hercules.

This doesn't cover the possible spectrum of heroes, though. A fourth type is the one who, by speaking truth, especially if by doing so reveals hidden truth, changes the nature of the world. This is a high-sounding phrase that includes such ordinary types as Hercule Poirot or the Continental Operative. Borrowing from Hyde (although I think perhaps he would disagree with my doing so), I call this hero type "the prophet." Prophet not because they predict the future, but because (as is the original meaning of prophet), they speak eternal truths. I think this archetype might explain the enduring fascination that people have with mystery novels. They are powerful reminders of the power that revealing hidden truth has.

Thursday, January 15, 2004

Invent the Future of Computing

Onward! Seeking New Thinking and New Paradigms is a sort of min-conference within the larger OOPSLA conference. While OOPSLA is focused on object-oriented programming, Onward! is broader, seeking new ideas for computer science and programming from any corner. It's a good track, and, I humbly think, an important thing to step back and look for bold ideas. Please consider submitting your ground-breaking work. Here's the complete Call for Participation, found also at

Onward! Call for Participation

OOPSLA 2004 welcomes papers describing new paradigms or metaphors in computing, new thinking about objects, new framings of computational problems or systems, and new technologies. Papers in the Onward! Track need not advance the state of the art, but should aim, instead, to alter or redefine the art by proposing a leap forwardóor sidewaysóregarding computing. Papers in the following areas are welcome, as are any papers representing radical thinking of interest to theoreticians and practitioners at OOPSLA:

* programming language theory, practice, and design
* architectures
* software development and methodologies
* applications
* environments
* education
* ethics
* biological or other non-traditional models of computing
* paradigms, metaphors, philosophy, and problem framings

An Onward! Track paper need not contain a fully worked out theory or implemented system, but must be well thought out, well-written, and compelling in its vision or uniqueness of thinking. Papers submitted to the Onward! Track will be reviewed by a separate program committee, and in some cases papers submitted to the regular program may be directed to the Onward! Track, if appropriate. Accepted papers will be presented in parallel with the OOPSLA regular technical program and published in an ACM publication. Onward! will have a Keynote speaker during a session open to the general public.

Important Dates

Firm deadline for receipt of submissions: March 19, 2004
Notification of acceptance or rejection: May 14, 2004
Deadline for camera-ready copy: July 19, 2004

For More Information

For additional information, clarification, or answers to questions, please contact the Onward! Chairman, Geoff A. Cohen, at

Program Committee

Geoff A. Cohen, Coherence Engine, USA, chair
William Cook, University of Texas at Austin, USA
Richard P. Gabriel, Sun Microsystems, USA
Cristina Lopes, University of California, Irvine, USA
Martin Rinard, Massachusetts Institute of Technology, USA

Wednesday, January 14, 2004

The New Electoral College

I was going to build up to this piece by piece, but actually I find that writing about electoral systems is, well, kind of boring. So, here's my proposed solution to the US Electoral College, without benefit of exhaustive analysis. Note that I'm not claiming that this is an original solution, although I'm not actually aware of a specific proposal that has all of these proposed elements.

1. All the "faithless elector" problems are removed by simply mandating by law in each state that the vote cast will reflect the majority vote. There's no legal basis for an elector to cast any other vote; the EC is just a ceremonial position (although I think, for other reasons, an important one).

2. Each Congressional District has one electoral vote, which is cast for the candidate who receives the majority of votes. If no such candidate exists, the district casts no vote.

3. Each state receives electoral college votes only for its number of representatives in the House, not for the Senators. That is, each state has two fewer electors than under the current system. The District of Columbia receives one elector, although if it were up to me, it would be retroceded to Maryland already.

4. If no candidate receives a majority of possible electoral votes (including districts that cast no vote due to no candidate having reached 50%), the candidate receiving the most votes wins.

Why is this a good idea? First, restricting the scope of electoral elections to congressional districts accomplishes two things. It makes elections more compelling for individuals, since it is more likely that their vote will count. Two, it fundamentally changes electoral strategy such that candidates will be encouraged to make incursions into what is now hostile territory. For example, Democratic candidates will visit Salt Lake City and Raleigh-Durham, trying to strip off a vote or two from otherwise solidly Republican states. Similarly with upstate New York and much of California. Overall, this is a tiny net positive for Republicans, since they would get more from solidly Democratic states (NY, Penn, and Cal) than the Dems would pick up from Republican states (a few in Florida, one or two in NC, a fair number in Texas). If this change had been in place in 2000, Bush would have defeated Gore by a slightly larger electoral margin.

Yet Texas brings us to a potentially fatal flaw in this plan, which is the gerrymandering of CDs. This approach would be unworkable with the current, extremely political districting. Only when congressional districts are assigned by nonpartisan professionals, as they are in Iowa, would this approach be a good idea.

The net positive for Republicans of the per-CD vote is offset by the removal of the two electors per state, since overall that benefits Republicans. More small states (Wyoming, the Dakotas, Montana, Idaho) vote reliably for Republicans. Also, and more to the point, the small state bonus is extremely undemocratic. While avoiding one-person, one-vote is okay for building consensus among disparate interest groups, a 13-fold difference between a Wyoming voter and a California voter is beyond the pale. If this change alone had been in place, Gore would have defeated Bush.

There's an argument that this would hurt small states, but this misses the whole point of electoral strategy. No one campaigns in Wyoming anyway, because it's solidly Republican. Candidates don't go where the votes are, they go where they can most efficiently reach undecided voters. Swing states, or rather swing districts, will still exist, and they will still receive the attention of the candidates.

Will this happen? Well, no, almost certainly not without another Constitutional Convention. For all that it's more fair, too many states would perceive it as a loss of influence. I can' t see Wyoming ratifying an amendment that removed two-thirds of its electoral votes. And intensely partisan states, such as Texas, where Republicans dominate the state government, would never make a change that gave votes to the other side.

But we'll need a Constitutional Convention to absorb Canada, so it's not too soon to start thinking of a new system.

Enough of this. Tomorrow, Onward!

Monday, January 12, 2004

Electoral College, Good and Evil

Okay, a little break for a post-New Year's cold.

Let's consider, for a moment, the effect of moving to a general election system. Wind your mental clocks back to November of 2000. Sure, as election night draws to a close, it looks like Gore is going to be the next president. But what's that, you say? Voting irregularities reported in over 10,000 precints in 31 states? Recount! Except instead of just being a recount in one state, in a few counties, it's every county, every state. Because to make up a half-million vote deficit, the Bush people will look under every rock in the country. Not just precincts where the vote count was close, because it doesn't matter. Imagine the Florida scenario in your home town. Worse, such a system, to put it politely, invites creative vote reporting. When a state is boolean win-or-not, there's little incentive to cheat, except if the vote is within such a small margin that cheating might both change the outcome and remain invisible. Such margins aren't impossible, but at least they're rare.

This is the unfortuante flip side of depressing voter turnout: a winner-takes-all system might make voters less likely to turn out, but it also makes voter fraud less attractive and limits the scope of post-election investigations and recounts. It's a deep philosophical policy question of whether this is a good tradeoff to make. On one hand, anything that depresses voter turnout is implicitly bad for democracy, right? And yet the logistical problems of running elections are real, and as we've seen, contribute in their own way to damaging faith in democracy.

Could a perfect voting system remove these concerns? Possibly, but such a system doesn't seem a likely reality anytime soon.

Tuesday, January 06, 2004

Why an Electoral College?

The Electoral College system exists as it is simply and purely because of historical contigency: it was the architecture necessary to reach compromise among the states and ratify the Constitution. But let us float above those messy details, and look at the wider picture. I've written in this space before about the difficulty of reaching consensus. Consider the Electoral College (EC) as a system for reaching consensus.

First, "consensus" doesn't mean that everyone agrees, and it doesn't even mean it's the decision of the majority. The post-2000 firestorm of criticism against the EC was based on the fact that Gore received more votes than Bush, in an absolute sense. The unspoken logical step was that the receiver of the most votes (although Gore did not, in fact, receive a majority of all votes cast) most deserves to be President. It was this very scenario that had people calling the EC a "ticking time bomb" or similar overblown metaphor. Aside from the fact that, despite some very angry Democrats, the nation was not plunged into another civil war, this misses the point. The EC functioned exactly as it was designed to do.

The EC basically performs two functions: it overweights the contributions of small states, and it makes it far more difficult for regional candidates to win. Both of these are important functions when attempting to reach consensus among a diverse set of actors that differ in power.

However, the EC also has some problems. It discourages voter turnout, it marginalizes the political minority in each state, and it marginalizes the issues of most states.

Moving to a new system, such as a general election, would solve some of these problems but create new ones, as well as losing the advantages of the EC system.

Monday, January 05, 2004

Election Year

Happy New Year, and welcome to 2004, a presidential election year in the U.S. Although I mostly intend to keep to my no-politics ban in this blog (very difficult, for those who know me), I think one area is fair game. The systems we use to choose our representatives and presidents is, in a real sense, a form of social software. Both as an interesting case study in ways for large groups (> 270 million people!) to make decisions, and as an impotant intance in its own right, the mechanisms of the U.S. electoral system deserve constant scrutiny.

I'll begin with the oft-criticized Electoral College, analyzing some common criticisms, some overlooked strengths, and humbly offer some adjustments.