Bush on Meet the Press: A Rebuttal from Juan Cole

Fantastic analysis and rebuttal to President Bush's statements on "Meet the Press" this Sunday. Definitely worth reading, and forwarding to anyone interested in the US presence in Iraq.

The CIA: Method and Madness

Ok column by David Brooks in the NY Times today about the problems with the way the CIA analyzes threats against the US. In general, the points he makes about the use of "scientism" to predict these threats to the exclusion of all else is correct. He doesn't specify the lists of errors that the intelligence community has made, but the general point that they've missed threats by non-rational actors is perhaps the most important he makes in the entire column.

If there’s any problem I have with the column, it’s the idea that “scientism” has no use whatsoever. Implicit in that argument is a defense of the creation of a new intelligence unit within the Pentagon. They operated on the sort of intuition that Brooks advocates. But without the sort of processes and actual intelligence know-how that exists in the CIA and elsewhere, intelligence information of dubious credibility was put forward as fact.

While relying solely on “scientism” to predict and stop threats isn’t the best idea, neither is turning to the other extreme (going exclusively with your gut, or instinct).

Scott Ritter was Right All Along

I heard him this morning on C-SPAN with Brian Lamb. It's very interesting that he was perhaps the only one before the war who consistently stated that Iraq didn't pose an imminent threat to the US and that they didn't have weapons of mass destruction. Before the war, and prior to hearing him say it in another interview, it didn't even occur to me that biological and chemical weapons have a "shelf life". They can expire, like over-the-counter drugs. The man suffered all manner of abuse and harassment, allegations of being "un-American" and a traitor to his country.

Now that David Kay’s report has stated that they’ve found no weapons, and likely never will, Scott Ritter has been completely vindicated. Though their motives may not have been the best, one could say that the French were right as well.

Especially interesting in the Brian Lamb interview was Ritter’s response to a caller who suggested the idea of impeachment. He said the right things about having all the facts, and due process. In addition, he laid out a pretty strong case that Dick Cheney might be guilty of fraud, if he knew some statements he made prior the war to a VFW gathering regarding an Iraqi defector were false.

I especially appreciated the way Ritter placed blame at the feet of Congress for giving the president a free hand to go to war. I’ve always believed that Congress abdicated its responsibility in doing so.

The other valid point he made was that the United States needs to be a fully active member of the UN in order to prevent Iraq from descending fully into chaos. To him, bringing the UN into Iraq did not mean the US completely exiting Iraq.

Overall, a number of the points Ritter made about the war sounded like things Howard Dean has said on the election trail, most importantly that the US isn’t safer as a result of the war in Iraq. Interestingly enough, despite being a self-described “moderate conservative”, Ritter said he was going to devote all of his efforts to getting Bush defeated, despite voting for him in 2000. Hearing more from Ritter in the press about Iraq would be a good thing, especially as the election gets closer.

An Absence of Legitimacy

Excellent commentary by Fareed Zakaria on what's happening in Iraq. The point that he makes at the end about the transition plan being accelerated to coincide with the US elections does a good job of highlighting one of the key drivers of Bush administration policy--whether or not it helps them politically.

It’s quite fitting the organizations Bush and company heaped such scorn on earlier are precisely the ones they must ask to help them out of the mess.

NASA Cancels Trip to Supply Hubble, Sealing Early Doom

NASA Cancels Trip to Supply Hubble, Sealing Early Doom

It figures. One of NASA’s most effective projects ever will be dumped in the ocean sooner or later in favor of manned spaceflight.

To me, it’s just one more decision that says we don’t have the first clue about what real science is. I still remember some of the debate over “super-conducting super-collider” vs. the International Space Station. At that time, it seemed clear to me that the likelihood of useful scientific discovery occurring was a lot lower with the ISS than with super-collider. But because it was in Texas, the Dems were in power, and they wanted to stick it to the GOP, the wrong project got killed.

History may not repeat itself exactly, but it sure does rhyme.

The Troublesome, Vote-loving Ayatollah

 

The troublesome, vote-loving ayatollah

This story from The Economist covers a lot of ground on America's adventure in Iraq. They do a good job of touching on the political implications for Bush's re-election plans if things go poorly. On the positive side, it looks like a significant number of our troops will get to come home soon. As misguided as I think the policy was, the men & women of our armed forces have done a great job of bringing it this close to success. I think it would be both ironic and unfortunate if after all this effort to bring democracy to the Middle East, the result was an Islamic republic in Iraq.

 

Teach Yourself Programming in Ten Years

Teach Yourself Programming in Ten Years

Great essay by Peter Norvig that begins by talking about the “Learn topic X in Y Days” publishing phenomenon, and counters with how long it really takes.

His recipe for programming success, and his contention that it takes 10 years to get really good at it both make a lot of sense. Since getting my computer science degree from the University of Maryland (in 1996), the things I’ve learned about programming from doing it for a living have really added a lot to my understanding of what I was taught in school.

Working on projects with other programmers is what I enjoy most. I definitely feel like I learn more and accomplish more when I work with a team. Unfortunately, the position I’ve had the past couple of years means that I have to develop everything by myself. I’ve certainly learned a lot from not having other people to depend on, but I think teams develop the best software.

Working on projects after other programmers has been one of the most frustrating parts of my career. Too often, it’s poorly documented and written in an “ensure job security” sort of way. In other words, it’s not written in a way that easily allows someone else to understand it. Programmers only get away with that if they’re working by themselves. More often than not, if I’m confronted with that situation, I’ll rewrite the application instead of spending a lot of effort deciphering the existing code.

Talking to other programmers about programming is something I don’t do enough of. Finding time to read other good code is a challenge (trying to balance full-time work and a part-time MBA program), but I need to do that as well.

Fiber to the People

Fiber to the People

A Lawrence Lessig article on customer-owned networks. He uses examples from Boeing Corporation and the city government of Burlington, Vermont. Governments, corporations, and individuals (if they have sufficient funds) buying networks definitely has some benefits. No more worrying about Comcast being the only way to get onto the internet. The knowledge that a municipality could hire contractors to build such a network themselves could also work as an incentive to spur competition (if corporations are being too slow about providing access to certain areas). I hope more cities and counties do this.

USB FM Radio



I wasn't even aware such a thing existed until my friend Rob told me about it. The one I bought is a discontinued model from D-Link (the DSB-R100). You plug one of the antenna's connectors into a USB port, the other into the Line In port on your sound card.

I found out the hard way that it's very important to completely ignore the software that comes with it. After installing the driver and radio software on the PC, my machine crashed. Every other time I tried to start up the radio software, it would reboot my machine. So I followed the two pieces of advice I'd been given earlier:

1. Install the GemTek drivers instead
2. Install Radiator

The drivers and the software are both freely available. Now everything works like a charm. I've got the same FM radio presets on my PC that I've got in my car. The really interesting feature of the software is that it allows you to schedule the recording of broadcasts in WAV or MP3 format. You can record from other sources as well (directly from line in, microphone, CD audio, etc.) which I haven't tried (yet).

Stored Procedure Naming

Don't name your stored procedures with sp_

I found the following note:

Note: Do not preface your stored procedure names with sp_, because doing so reduces performance. When you call a stored procedure that starts with sp_, SQL Server always checks the master database first, even if the stored procedure is qualified with the database name:

at this URL: http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnbda/html/BOAGag.asp

Presenting Data and Information

Presenting Data and Information (a 1-day course by Edward Tufte)

I attended this presentation last Wednesday. If you ever do presentations for work, your MBA program, or anything else, attending this presentation is the best money you’ll ever spend. The fee includes a copy of all three of his books, and he refers to them often during the presentation. If there are two key points that can be summarized from the entire day, they would be:

  1. Paper is a much higher-resolution medium for communicating information than a computer screen.

  2. Only overhead projectors are worse for communicating information than PowerPoint.

Tufte does an excellent of substantiating each of these points. He illustrated the second point in hilarious fashion by showing us The Gettysburg Powerpoint Presentation, courtesy of Peter Norvig (and the Microsoft Powerpoint AutoContent Wizard).

I’ll include my notes from the course in a future post, but attending it in person is the best way to get the full effect.

Presenting Data and Information

Notes from Presenting Data and Information

A basic Google search will reveal all kinds of pages of information about Edward Tufte and his ideas–including notes other people have written about their experiences during his presentation. Looking at those will give you a broad picture.

My notes from the presentation are below:

  • 6000+ year history of trying to communicate 3 or more dimensions of info in 2 dimensions (paper, computer screens)

  • Data density and increased resolution are other issues with presenting data

Recommendation: put the data directly on the chart, don’t shove it into a legend

Rules of analytical design

  1. Show comparisons

  2. Show causality

  3. Show multivariable data

  4. Integrate words, numbers, & images

  5. Document everything and show documentation (sources). Don’t trust people who fail to provide documentation.

  6. What happens in a presentation depends on the quality, integrity, and relevance of the content.

  7. Try to make important comparisons adjacent in space instead of stacked in time.

  8. Use small multiples. They are comparative, multi-variate, and adjacent in space. Repetition, in time-series, increases credibility and integrity of presentation.

Example: Napoleon’s March (originally an anti-war poster, according to Tufte)

Dimensions of information:

  • size of army

  • location (3 dimensions)

  • direction

  • temperature

6 dimensions of data in only 2 dimensions (paper)

“Why should our work show the chaos of the bureaucracy that produced it?”

Principles of analystical design are the same as those of analytical thinking.

Good design is clear thinking made visible. Unfortunately, the converse is true.

Recommendations: Get a Mac. Get a big, high-res monitor. Even better, multiple screens.

Don’t de-quantify data. In page 31 of Envisioning Information, the chart would be better if the X’s were replaced by the number of occurrences of the crime.

Edward Tufte is working on a 4th book which should be out this year (he self-publishes).

You must show scales of measurement.

If you have data with high periodicity, consider carefully the vertical scale of your data. Smoother data can reveal trends that “spikey” data can’t.

Side note: when sewage gets into drinking water, cholera can result. This comes from a discussion of process of stopping a cholera epidemic in London, and the way data was used to determine the source.

Make small, but effective differences. This brings out more info.

Defend your content against a production process that might mess it up.

Borrow strong design (don’t reinvent the wheel).

Displaying Financial Data

  1. It’s generally an assessment of change.

  2. Show averages and variation. Detail increases credibility. Clutter and confusion are design failures. Micro-macro design is a good idea.

  3. When comparing money & time over more than a year, you must adjust for inflation so comparisons will be correct. Annual reports typically don’t do this, so their data in tables & graphs is wrong. Data needs to be adjusted for seasonal effects.

  4. Don’t trust a display with no footnotes. Footnotes help credibility.

  5. Financial data should explain as well as describe. Annotation is a good tool for accomplishing this. Every table & graphic in a presentation should be annotated.

  6. Find and follow great conventions of displaying financial data. Borrow strength. Don’t be constrained by time or geography in your search for examples to borrow.

Consider what thinking task the analytical display is supposed to help.

Regression toward the mean is a statistical fact. So you must separate that from other possibilities when trying to use data to make public policy.

Recency bias is the tendency of people to give more weight to the most recent data.

Spark lines are a Tufte invention for displaying financial (or similar) data. They display better on paper than on a computer screen.

Design should not reproduce the hierarchy. It shouldn’t reproduce bureaucracy. Don’t reproduce binary searches.

Design should be as flat as possible. Should depend on scanning of reader’s eye, not scrolling. Sites like GoogleNews and Yahoo and other portals are successful because they cram lots of info into the first screen.

Paper is: permanent, high resolution, portable

“Users should never see an operating system.” (Edward Tufte) On the shop floor, and in military environments, user interfaces are devoted entirely to content.

Screen should be 90% content in web design. You don’t need more than a nav bar to navigate a website.

Side Note: Edward Tufte loves the iBook G4. The MacOS is ok, though he doesn’t like the dock because it reveals the OS.

Microsoft bought a company that developed an app called Presenter in 1984. The app became PowerPoint. A software corporation is its underlying metaphor (structured, nested, hierarchical, short lines). It is presenter-oriented, not audience-oriented.

Powerpoint’s problem is that it’s low-res. the bullet points are a big problem too.

Don’t abbreviate the truth to fit on a slide. Find a better presentation method.

An 11x17" piece of paper can hold the same amount of information as 250 PowerPoint slides. Feel free to use PowerPoint to display really expensive graphics (that would be cost-prohibitive).

Bell Centennial is both highly readable and highly condensed. Great for use in serious printing situations.

Solid rocket boosters are built in Utah, assembled in Florida. The post-manufacture assembly is what makes O-rings necessary.

Lessons to Learn from Challenger Disaster Presentation

  1. Show me causality

  2. Show me all relevant evidence

  3. Ask yourself: what do I really need to see?

These 3 things improve the chances of making the right decision. Tables should have numbers and words in them.

Supporting material can be found here.

Techniques for Better Presentations

  1. Show up early. Good for giving handouts and getting to know your audience individually.

  2. Problem, relevance, solution. Your presentation must have these elements.

  3. Never apologize in your intro.

  4. Particular, general, particular. This pattern can be used to present slide information (instead of just reading the slide).

  5. Think about your audience by what they read.

  6. The Gettysburg Address (what not to do)

  7. The audience is precious. Treat them accordingly, don’t patronize. Consider them colleagues. Be civil and diplomatic, but don’t self-censor.

  8. Humor is good. But don’t be generic.

  9. Don’t use masculine pronouns universally. Feel free to use plurals.

  10. Don’t hide behind the podium. Believe in what you’re presenting, and be demonstrative.

  11. Finish early.

Outside the Meeting

  1. Practice, practice, practice

  2. Practice in front of a friend.

  3. Practice in front of a video camera (so you can pick up any habitual gestures or movements, stutters, use of the word “um”, etc.)

  4. Get better presentation content.