If we can take the same urgency, enthusiasm and clarity of vision you need to get elected to government, and apply those to governing, could we fix some problems? Yes we can.

Well, only sort of. Because it's nearly impossible for government to just hire the guy who led that construction, Harper Reed, and to allow for Harper to hire the people he needs in order to get that done.

Let's talk about why it's so hard for government to do that.

"To the victor goes the spoils!"

That was the old saying in American government up until the beginning of the last century. When elected president, the friends of the president got the cush government jobs and contracts that went alongside the victory -- a patronage system. Run loosely, this was a key cause of a lot of corruption, entrenchment, and even the assassination of James Garfield.

In 1883, we began to see move away from this patronage system to a "merit based" civil service with the passage of the Pendleton Act. From Wikipedia:

The act provided selection of government employees by competitive exams, rather than ties to politicians or political affiliation. It also made it illegal to fire or demote government employees for political reasons and prohibits soliciting campaign donations on Federal government property. To enforce the merit system and the judicial system, the law also created the United States Civil Service Commission. A crucial result was the shift of the parties to reliance on funding from business, since they could no longer depend on patronage hopefuls.

But today, especially in the world of engineering, those competitive exams seem to be competitive tests of will and patience, rather than merit-based engineering tests.

If someone like Harper Reed wanted to work for government as an employee, he'd have to go through an exhaustive series of background checks ranging from an extensive credit inquiry (why do you have a $200 outstanding debt from Verizon from 15 years ago in College) to international travel (do you have anyone who can vouch for all of your whereabouts when you backpacked through Thailand) to residency checks (please give us contact information for someone who can vouch for your good character in every place you've lived in the past 7 years). Applying to a federal position involves exhaustive background checks, but Harper wouldn't be asked for any code samples.

The spirit of the law is well intended. While the will of the people is important, it is probably important to the continuity of the business operations of government that the president does not get to fire everyone who disagrees with her or him. If you're an environmentalist: imagine a president firing the whole staff of the Environmental Protection Agency, and replacing them with the executive staff of BP America. If you're a war hawk, imagine the President firing the citizen staff of the pentagon and replacing them with volunteers from CodePink. But it also makes it so that people who are demonstrably under-talented or uncommitted to the job are difficult to terminate.

At the same time, it's made very talented people hard to hire. While many computer programmers may opt-out of higher education, not having a Bachelor's degree will effectively screen you out of the first round of hiring -- software will eliminate you before a human even gets to think about it.

If talented people are hard to hire, and untalented people are hard to fire, in a place as competitive as technology, how can government get anything done?

Do your bidding

A private business is allowed to hire and fire people as they (or they, in partnership with their employee union if that's the case) see fit. So in information technology, the federal government largely turned to the private sector. With the exception of edge-case programs like the Presidential Innovation Fellowship of which I was a part, real programmers (people who can write code in text editors and terminal windows) in government are hard to find.

So if the United States wanted to hire Harper for a federal project, such as the VA+DOD integration, it'd probably be easiest to hire Harper and his team to start HarperCo and contract with Government. But even if the president knew that Harper could do the job, and do it better than anybody else, he couldn't just hand over the contract to Harper. After all, that'd still be a significant patronage.

Instead, the VA and DoD would have to write a RFP for the exact requirements of the work and put that out to for public bid. Using a network of Contracting Officers (who are not technologists) and assisted by an appointed Technical Representative(who are rarely technologists), they draft up detailed specifications of what's needed. Sometimes, they seek outside community guidance in the form of public requests. However, should HarperCo start providing direct feedback on this RFP, HarperCo would likely be disqualified from bidding on this contract. So because government cannot hire people to figure out the solutions, and cannot ask qualified people what solutions to ask for, government tends to ask for a lot.

Which is why Recovery.gov ended up costing 18 Million Dollars. Take a look at the RFP they came up with. Data Cubing? XML Firewalls?

But let's say that government miraculously asked for exactly the right thing, and HarperCo responded to RFP with the lowest bid. Would Harper win? Not likely. Harper's firm is less than 2 years old (strike one), and owned and operated by a white male who is not a disabled veteran (strike two), and has no experience in federal contracting (strike three).

Instead, government would use the Alliant Governmentwide Acquisition Contract for a procurement like this. A Government Wide Acquisition Contract(GWAC) is a way for government, through the General Services Administration to pre-negotiate hourly rates, profit margins, and costs with outside contractors. Alliant is the largest of this kind of contract in information technology, with a ceiling of $50 Billion. There are about 60 companies (Raytheon, Accenture, IBM, etc) who have negotiated this kind of contract with GSA. It's not open to any new businesses.

So Harper would have to partner (or "team" in industry speak) up with one of these existing firms (we'll call our hypothetical one PrimeCo) to do the work. The prime contract would go to one of these 60 firms, and they, in turn, could subcontract out to Harper.

Harper, being a generally honest fellow inexperienced to government contracting, would probably estimate the number of hours it would take him to do the job to be similar to the campaign. Let's assume that HarperCo assumes it will take 12 developers, one designer, and two project managers one year of full time work. That's 30,000 man-hours. Let's say the billable rate for the project is an average of $200 per hour. That's a $6,000,000 cost. But government has mandated in its negotiation that PrimeCo is bound to a pre-negotiated profit margin of ~10%. There's no way that one of these companies can survive its largesse by doing projects like this and walking away with just $600k.

So what they can't get in hourly margin, they make up in volume. Before sending on his proposal to government, PrimeCo decides that Harper needs an additional 30 developers for support, and 20 new project managers. Remember, these are publicly held corporations. It's their legal, fiduciary responsibility to return as much profit to shareholders, not to save the tax dollars of citizens. Now we're getting somewhere with the profit we can generate. We've still only got the contract up to a measly $2.6 Million in profit, but you can see where we're going, and how it's done. Now that we've added all these people to the process we've created layers of bureaucracy and expense, and additional points of failure. Enough to chase off guys like Harper from even doing this.

You can see the policies and behaviors that end up causing billion dollar IT fiascoes, and more importantly, the good intentions that got us here.

Nobody Likes Incompetent Government

I can already hear a chorus of libertarians singing praise of this post. "See! Government can't do stuff effectively! That's why there should be limited government!"

Listen: unless you're an anarchist, you have to believe in some form of government. Whether you believe government should be very small, or that government should be an instrument of good, you obviously want the people running it to be the brightest and the best. These problems aren't a question of this very tired national argument of whether government should be big or small, but rather whether government should be competent or not. If nothing is done about this, even if Grover Norquist was able to shrink government down to the size where it could be "drowned in a bathtub," you can bet that the bathtub would cost ten billion dollars.

So we have to fix this -- the way that government hires and the way government buys are the central nervous system and circulatory system of government. They're vital to working well. We all want government to be a platform but the truth is, the platform's held together with duct-tape and billion dollar toothpicks.

So the question is, how?

Well, first it's by giving up on the nonsense of our current political debate. It's a lot like our obsession with our bodies being fat vs. thin rather than sick vs. healthy. It's a gross frame that doesn't do anybody any well, and as long as that's our frame we'll continue to make stupid decisions. As Stewart implies, it lets us off the hook -- where we'll knock on 5 million doors to elect a president, but won't lift a finger to help soldiers get access to their own medical records. We've got to start caring a lot less about electing, and a lot more about governing.

Once you do, then it becomes about two policies inside of government: Procurement and Human Resources. How government buys things, and how it hires. I'm helping to fix the former that through The Department of Better Technology -- by building tests that prove that governments can buy better technology by avoiding these huge primes and going directly to the HarperCos of the world. We'd love your help, either by contributing to the open source software, deploying it in your community, or by asking your local government to enroll in the pilot.

But obviously "helping my project" isn't the only way to help -- that'd be self serving. Government can be a platform but we've got to start advocating for sensible change in policy that opens government's doors to innovators and new businesses. While federal policy may change, it's easier to start with your local city or county. Start asking questions: how much did yourcity.gov cost? Who made it? How did they win the contract? Why?

Fixing this problem isn't a question of policy, it's a question of will and attention. Should the public start paying real attention to these problems, and start questioning these policies that make government and the private sector partner up so poorly, we may have an opportunity to make change. But as long as the word "procurement" causes eyes to grow full of glaze rather than action, we'll see no progress.

That said -- while its probably the most disgusting act of cowardace we've seen in awhile, it's not the only act of cowardice. From Columbine to Sandy Hook, from Mark Barton to Sueng-Hui Cho, its clear that guns are getting into the wrong hands more often than I'd like, and that something needs to change.

The bill that didn't pass yesterday certainly was not the solution I was looking for, and it certainly wasn't the step that most people were looking for. But it was a compromise, and compromises are designed for consensus, not partisan victory. More than anything, I think Americans were looking to see if the United States Senate could accomplish something -- anything -- of a step forward. Having a bill that 90% of Americans, 80% of Gun owners, and 70% of NRA members support seems like it ought to be accomplishable.

Think about the last time you heard about the number 90% in politics. Do you know how hard it is to get 90% of people to agree on anything? Very few members of Congress won with 90% of the vote. No president I've ever heard of won with 90% of the vote.

For the United States Senate to turn its back on 90% of the people's will -- that isn't about guns. It's about demonstrating that an institution that's vitally important is so hopelessly broken that it'd rather turn its back on the people it serves than confront a single special interest group.

See -- these things are related: you might hate that government doesn't let you buy fresh milk, and raids milk farms, and I might hate that government doesn't mandate ingredient lists on cigarettes and alcohol. You might want to tap the oil well in your back yard, and I might want to finance solar panels to put on my roof. But all of these things are related: the ability for a narrow set of people to manipulate our government towards their will instead of the people's. It's subtracting the common from our sense.

And that is far more dangerous to our livelihoods and our future security than any crazy guy with a gun. And that's why even if you are happy with the Senate's failure to pass a bill yesterday, you should still be pissed off. In order to fix the system, we have to excuse ourselves from our emotions on individual issues of the day, and look more towards fixing it when the system goes so dramatically against the will of the people. Until we understand that every time -- whether we disagree with the individual outcome or not -- that when the system does that, it is a failure of us as voters, we cannot change anything.

.

See the OMB Control Number junk at the bottom? You may also recognize it from your tax forms or all other government forms. That means that this "information collection" had to go through a minimum of a 45 day public comment period and probably a six month review process to be put on the internet. That's why you should hate the Paperwork Reduction Act.

How can you have a democracy, when asking a binary question like this takes a six month approval process?

Our best comedians could not make this up.

Now the obvious thing to do would be to do something poisonous to your product -- to somehow make it dangerous or deadly. Add lead to the toys. Put the spark plugs next to the gas tank. Put mercury in the sausage. But that's the stuff that makes for short term catastrophes that could end your business, not long term contempt that'll keep your business hated, but still keep you in business. You want to run this like a cable company, not ValueJet.

No, for long term contempt, you need stuff that nobody notices. Stuff that can stick around in your organization forever and not be corrected because it's long been forgotten. This is a problem that can't be solved with such sophomoric thinking as just accidentally running over some children with your trucking business. You need to bring in the experts at this: the "corporate policy" people.

What I'd do is create a policy that makes it really hard for my company's employees to ask questions of my company's customers. I'd make it a struggle to collect feedback. In order to collect any form of feedback, I'd make it so that you had to first ask for permission from an underfunded and understaffed component of the central office of my corporation.

Of course I'd also make it take at least six months to get this approval. That way, most of the people who wanted to ask my customers a question were immediately discouraged from doing so. And of course, the people that I'd put in this underfunded understaffed component of my central office -- I wouldn't make them professional question askers. They wouldn't be language experts or people obsessed with the "customer experience." Instead -- just to make sure that whatever questions to customers that did come out of my office were terrible -- I'd staff this office with economists and lawyers.

Then, just to be especially perverse, what I'd do is encourage my company to use social media. I'd create policies around it, pushing my company to go online on Facebook and Twitter and stuff, and to have "authentic conversations" with our customers. I'd tell them that it was totally cool to use social media to informally do whatever they wanted, except to use that information to inform product or service decisions.

This way, my employees will be completely cut off from their customers needs. And the only employees that actually make it to the customers are the people who know how to talk to the economists. That'll make it so whatever inputs and outputs of my business are so incomprehensible that they'll just create more frustration rather than solve problems. And making people go out in social media? That's just the icing that makes it so people think they're giving input to the company without that input actually making it anywhere useful. That'll make the customers nuts!

It's a machievellian scenario that, sadly, I didn't make up. This "corporate policy" is actually a law that makes your government act like this, and it's nefariously named the "Paperwork Reduction Act." It was the last bill signed into law by Jimmy Carter in 1980.

I don't know whether this outcome was Carter's intent. During my presidential innovation fellowship, I spent two out of the six months I had, simply trying to figure out the legal way for our project to ask a question. Not writing code, not talking to customers -- just filling out the paperwork and seeking approval to put this form on the Internet. And that's as a high-level appointee with air cover coming from the White House. Can you imagine what the people who dedicate their lives to this have to go through to talk to customers?

Did you know that when this president took office, it was illegal for the President to end a tweet with a question mark without a six month approval process from the economists across the street at the "Office of Information and Regulatory Affairs." No seriously -- they seriously had to give guidance to the rest of the federal agencies in 2009 that gave them permission to ask questions over the internet. It basically says: Sure, you can ask people questions, as long as you don't ask for structured feedback (feedback you can do anything with). Thus it became ok to end sentences on twitter with a question mark. I can't make this stuff up!

The Paperwork Reduction Act is a terrible law. It doesn't need to be revisited or revamped. It needs to be removed. In 1980, I'm sure collecting information cost government a lot of money. Forms had to be made. They needed to be proofread. It needed the mail or the telephone. It was expensive to key in the data collections. It was worth taking the time.

But today, it's a disaster. There's nothing I can think of that's more antithetical to democracy than prohibiting government from asking for feedback from its citizens besides, maybe, prohibiting them from actually voting. Though there's a case to be made that citizens providing feedback on actual policy is just as important than who they elect.

This law doesn't govern over "forms" it governs over "information collections." From tweets to RFPs to Facebook posts, to regulatory questions. Want to make a form that asks developers to report bugs in datasets? Forget it. Inside of your government it empowers the "culture of no." After just a few months of working on the inside, it kills your ambition to actually talk to customers, and instead encourages your government to operate blindly.

It makes insanity happen.

Why You should care

If you want to know why you don't feel like you're being listened to, or why government is flunking at social media it's not because of a lack of will. It's because of the Paperwork Reduction Act that's actively prohibiting your voice from making it to the right people in Washington. There's no other country on the planet with a law like this, And that's something Congress has to fix.

Whether you sit to the right or to the left, a democratic republic should be great at asking questions and getting answers from the people it's intended to serve. If it can't do that, it can't serve anybody.

To grossly oversimplify, imagine that there's two different kinds of programmers: cooks and chefs. Cooks understand recipes, they can follow them, and when they follow them, their food looks and tastes good. But only chefs really understand the food. They are the ones who make the recipes.

It's because chef understands how the food works together. They understand the fundamentals of flavor, and the interoperability of food. They understand the nuance of time and temperature, and how fat and protein work together. Asking a cook to change up a recipe is a recipe in trial and error. Asking a chef to do the same has a higher probability of success because the chef understands the patterns underlying the food.

For a programmer, putting Design Patterns on your desk at work means "I want to be a chef, not a cook." (This also explains why so many copies of this book have been sold, yet so few people have read the book)

As government increases its online capacities at the local and federal levels, it needs to make the transition from thinking like cooks into thinking like chefs. When you look at things abstractly, you start to see areas of opportunity to solve repetition.

Pattern 1: Vetting and Selection

For example -- RFP-EZ is the project I worked on during my fellowship. It's intended to make it easy to post RFPs online, to make it easy for small businesses to bid on them, and to make it easy for people inside of the government. When we (the fellows) were asked what the ideal software would be to manage the application process for the next round of innovation fellows, we started designing some rudimentary software for it, but it turns out we'd already done it: RFP-EZ, abstractly, is a method for posting a request online, gathering feedback from that request, and collaborating on deciding which piece of feedback was best. It could work just as well in a hiring decision as it can in a procurement.

So if you've applied to become a Presidential Innovation Fellow, you've used a version of RFP-EZ with enough garnishes that we call it HIRE-EZ.

I think there are a bunch of abstract patterns inside of the open government movement that we should be moving towards. The RFP-EZ/Hire-EZ one is an example of one I like to call "Registration and Vetting." We make this kind of application and do this kind of work for government a lot -- whether it's selecting a new fellow, running a procurement, or ensuring that you're qualified to drive or that your business is qualified for the benefits it receives.

But government isn't particularly good at this process. It can take months to register your small business as a small business with the small business administration. Nobody likes going to the DMV. Perhaps if it thought of the process a bit more abstractly, efforts could be shared, repeated, and more innovation could happen where it matters. A cook says "let's upgrade the software at the DMV." A chef says "let's figure out the right recipe to make a basic piece of software, policy and process work great, then let's share that and let people add the right garnishes to fit their specific need."

Pattern 2: Information Collection and Analysis

You would be shocked at the amount of time, effort, and money spent on government putting a form on the Internet and having the results of that form go into a spreadsheet for future analysis and policy making decisions. I would not be surprised if you could find the entire sequester budget cut solely in projects that simply do this. Millions of dollars in both time and budget are spent on convoluted systems that all do exactly the same thing.

This costs the rest of us zero dollars. Solving this problem inside of a government once will improve the citizen experience, and save the taxpayer billions of dollars.

Pattern 3: Redaction and Distribution

Remember the amazing costs of FOIA. It's a half-billion dollar problem that needs to get solved. Coming up with the right patterns (and software) for government to be able to quickly and properly release information is made for reusable components.

This goes beyond FOIA -- it needs to identify a dataset that should be open, to qualify it, to redact personally identifiable information from it, and to put it on the Internet. We need a technical process for releasing information and the open, interoperable software that goes along with our open government policies and directives.

Pattern 4: Service Requests

Open311 solves this at the physical municipal level. Services like SeeClickFix know this and are leveraging this design pattern into substantial businesses. But the pattern goes beyond potholes. How come there's no bugs.data.gov that allows developers to report back bugs in data from something that comes off of it? Any other open technical release is accompanied with an issue tracker. It's the same technical pattern.

Pattern 5: Performance Management

Government spends a lot of time measuring how it's working. Whether it's via internal performance goals, or through services like Recover.gov, or through legislated mandates like the Paperwork Reduction Act's Paperwork Burden Esimations.

I'm probably the least versed in this pattern than anything else I've mentioned. But I know government spends a lot of time measuring itself, its employees, its contractors, and its actions. It's an area rife with failure, and failure here means retaining bad talent, enforcing bad policy, and managing things inappropriately.

Abstract doesn't mean Enterprise

One might think that this is a great opportunity for "enterprise software" -- the case for more multi-million dollar IT projects that solve many problems poorly rather than a few, well, poorly.

These patterns are not nearly as much about technology as they are about process. It's likely, for instance, that it doesn't require a "data management system" from a ALLIANT QUALIFIED VENDOR to solve the redaction and distribution pattern. It probably just requires a dropbox folder installed on a government worker's desktop labelled "Open Data," and a technical and editorial process for what happens when a new file is added to that folder.

Talking about these types of patterns will be a boon to large contractors if you let it be. But in general, we're talking about figuring out playbook patterns to run, figuring out why they work, and reusing them. Usually they're small and simple, not large and complicated.

This is far from an exhaustive list -- but I'd imagine there are only a handful of repeatable patterns in the citizen/government space. And I think we ought to focus on identifying those patterns, and finding the right people to design the technology and process to implement them rather than shipping apps that fit niche needs for government.

Why you should care

The debate in Washington over funding and budget is nonsensical. It's as though America is a sputtering car on the highway, and we all agree we need to get the engine really revving again. The republicans say that the problem is that we've got too many people in the car. To make the car speed up, we should just throw a substantial portion of the passengers out. The democrats say that the best way to do things is to put more fuel in the tank and hit the gas.

The real problem isn't the passengers or the gas tank. It's that the car's engine only gets 10 miles to the gallon, and we pay 10x more for gasoline than everybody else. These design patterns help increase the miles per gallon, and reduce the cost of gas.

But more important than the data's release, the way that it's been released got me thinking that maybe we're overthinking how to provide online. To get to the data, you go to the consumer complaints database page, then click on the small "all data" link under the strange set of squares presumably to resemble a spreadsheet. This opens up a socrata instance, where you'll need to click up in the top right where it says "export" and then click where it says CSV underneath that.

That's a lot of software to put a 172 row CSV file on the Internet.

I think Carl Malamud has it right. With his projects, he's worried about suppliying the bulk data in as bulk data friendly of a way as possible. He's used mod_index (built in to apache), and a little bit of light design, and there's a user interface that's nearly universal. Malamud's challenge isn't software, it's organization and discovery. That's where the real challenge lies anyhow. Try and find IRS Statistics of Income data on bulk.resource.org. As soon as you see it, you know exactly how to get to it, and exactly what you're getting. You don't need to learn how to use this website -- you already know.

Please do not misunderstand me: I think Socrata is useful. I'm happy they're doing what they're doing. For average people, I think being able to take a look at the data in a table is great. Most people can't load a 50MB spreadsheet on their computer, and Socrata makes it easy to view a large dataset over the Internet.

But for the people who will actually do things with this data -- whether they be academics, journalists, businesses or programmers -- will likely find the Socrata experience limiting compared to what they can get out of SQL or Excel. Delivery of data over the Internet should be aimed at trying to make it as easy as possible for people to get your data into those tools. An ineffective and costly strategy involves trying to replicate those tools. You're creating learning curves and user burdens where they ought not be.

I'd love to see a bulk.data.gov aimed at replicating the bulk model for data distribution -- it's common and simple. Malamud's using the model for distributing the code of federal regulations and GovTrack uses it to distribute bills. I would have loved to see it before we got Data.gov -- as it seems lots of development has happened on a platform that fewer than we'd hoped are using (both on the data supplier and provider).

The bulk model Tauberer's GovTrack and Malamud's Resource.org use also easily allow for freshness and more openness. They provide rsync support -- a way to make sure data stays fresh and up to date without consuming a whole bunch of bandwidth. As the files add up on Data.gov, how do we only get what's new?

Finally, the bulk model allows for vendor independence. Again, while I have much love for Socrata and what it's designed to do, it's also the case that this kind of data distribution must be as platform independent as possible. While we talk about how open data can power the next Weather industry or the next GPS, I suspect none believe that it will be with Socrata as a middleman. By relying entirely on Socrata for data distribution, we place government in a concerning position where it's dependent upon a proprietary vendor for data distribution, and the economics of success depend on them staying in business.

So if you're out there, govvies, and you're reading this: Please, rethink your data distribution strategy. Do you need to put that csv file in Socrata? Or can you just upload it to your web server?

So how do great speakers like Larry Lessig prep and rehearse for a polished talk? I still don't know: when I asked him, he told me he locks himself in a hotel room for 2 days and I didn't get any more specifics than that. But after giving about 30 talks on the Information Diet since January, I can tell you how I prep and rehearse for a polished one.

Know the Scene

The first piece of prep you need to do should come weeks before you're scheduled to give your talk. Know what the room will look like, where you'll be, what the equipment will be like, and the structure of the event. Here's the bare minimum of the questions you should ask to anyone putting on the event:

1. How many people will be there?
2. Who are they, and what do you think they want to hear about?
3. Will there be a screen?
4. How big will that screen be?
5. Will I have a microphone? What kind of microphone will I have? (over ear, wired, podium)
6. What kind of slides can I present? Powerpoint, PDF, Keynote, Prezi?
7. Will this be run off of my laptop or your system, and do I need to bring a dongle?
8. Will I have a screen on stage to see the current slide?
9. Can I also have a screen on stage to see the next slide (i.e. presenter's mode. super handy)
10. Will this be recorded?

The last one's important because it means the talk you give is needs to not rely on external context. Your jokes need to be as funny to the viewer as they are to the audience member.

One thing I didn't understand for years is the dynamic going on between speaker and organizer. The organizer wants you to give a great talk, and will generally do whatever they can to make that happen. So don't be afraid to make a request. I always ask for a screen on stage to see the next slide, or the ability to use my own laptop (so that I can have that) and to date, I've never gotten turned down. So if you need something, ask for it.

Know your slides (at least one week prior)

There are many books out there about how to design great slides, and how to design a great presentation, but few about how you actually can become familiar with your slides. Check out Presentation Zen or any of Nancy Duarte's books for how you should design your presentations.

What I want to talk about, instead, is how to know the material you're presenting, and to get great at presenting it. Note that you should have completed all of this section at least a week before your presentation.

First, write the speech. Every presentation needs to tell a story to get across the point, and this will help you solidify your presentation's narrative. Go through each slide and write down what it is you're going to say on each one. Write it down. If it's more than three or four key points, and one of those points doesn't lead into the next slide, consider breaking the slide up into two, or rearranging the order.

Once you're done, do a rough draft presentation of your slides in front of a camera. It's OK if you're reading, but record yourself giving the presentation. Then watch yourself. Are there any places where you lose interest? That may be a place where you need to tweak the slide or your narrative.

Now, name each slide with a single unique keyword (unless its a duplicate slide, then using the same name is fine). Whenever you have a free moment, try and list those unique keywords from memory. This is a great thing to do in the shower, or as you fall asleep the night before, or relatively constantly up until the morning before your presentation.

Next, create a second presentation. This presentation should have exactly the same number of slides as your primary presentation. On each slide, type out the maximum 3-4 main points of its corresponding slide in your primary deck. This is your notes deck.

Do your presentation to a camera again. This time, instead of using your visual slides, do your presentation with the notes deck as a visual que of what to say and when to transition, but try and do it while looking as much and as directly into the camera as possible. Try and do this while you're standing up and not looking at your computer screen (just the camera). I do this by moving my notes deck onto my phone. This has the added benefit of it being accessible wherever you are for quick rehearsals.

Again watch the video you recorded. Figure out when you look down and need to read. Tighten up that messaging if you need to. Continue to do this until you aren't looking at your notes more than once per slide for transition and understanding what you're going to say next.

Do one more on-camera rehearsal for yourself with your presentation deck instead of your notes deck. If you need to have your notes, still, as training wheels, that's great. But the reason we didn't just put your notes into the "speaker notes" section of your favorite presentation software is because they're training wheels: the goal is to get you off of them so you can be present with your audience.

Test Audience (at least one week prior)

I tend to never find the time to do this, and find myself a much harsher critic on my public speaking than other people are. But it's still good practice to find somebody you trust to give you honest feedback on your presentation, and to test jokes and reactions upon. There's nothing like throwing a joke that you think is hilarious out there only to find that the audience doesn't find it funny to totally throw off the rest of your presentation.

The feedback that this person gives you isn't nearly as important as the data you collect by watching them react to your presentation. Either they're engaged with you or they're not. If you can't make your spouse not look at his or her watch, you're certainly not going to make a room full of people pay more attention to you than their twitter feed.

Take the feedback in anyway. Though your friends are usually untrustworthy. Most people view avoiding hurting your feelings to be a superior choice to being honest with you.

The Big Show

If you are travelling to give your talk, try not to travel on the same day you're speaking. Besides the unreliability of air travel, the TSA is great at causing you stress that can fry your brain and make your talk go poorly.

Try and get to the event in the afternoon the day before your talk. Ask the organizers if you can see where you're going to present, and if you can run through your slides. After that, lock yourself in your room. Go through your notes deck again, and give your notes presentations several times. Order room service. Give your primary presentation at least 5 times. Resist the temptation to change your presentation at this point. Your pre-show jitters are affecting your judgement.

Do not go out and drink. Do not have a heavy dinner. Avoid caffiene. Do not socialize unless it's to go get your speaker badge so you don't have to worry about it on the day-of. Go to bed early, trying to name all your slide keywords in a row.

On the morning of your event, a couple hours before your talk, go through your primary presentation one last time. Eat a warm breakfast. Get in the greenroom at least an hour before you're scheduled to go on stage. Do not clown around in the green room. Go through your notes. Do not check your email. Go through your notes. Do not check Facebook. Go through your notes.

20 minutes before showtime, shut down everything. Go to the bathroom. Relax. Take a deep breath, count to ten, and think about how you can be present with the audience. Clear your head and relax. 90% of the people presenting on that stage are not nearly as prepared as you are. Which means you're going to blow the audience away.

After your talk is over, stick around. Ask questions of other speakers. Be a participant, at least for a few hours. This helps other people know that you're accessible, and that's going to maximize the number of opportunities that arise as a result of your speaking in public.

Finally, don't forget to thank the organizers for their hard work. At large conferences, there's a lot that goes on that you don't see. A conference organizer's job is thankless, and a bit of appreciation to the people who are running your slides and wiring your mic go a very long way.

I'm sure you've heard it before, but I'll spell it out for you: Fewer things generate a higher return on investment for time spent than giving a great talk. Fewer things are more disrespectful than having that kind of opportunity, squandering it, and wasting a room full of people's valuable time because you didn't prep properly. So when you get an invitation to speak, be of service. The doors that will open to you because you gave a great talk are plentiful. I've used this technique for everything from my talks on the Information Diet, to Best Man toasts. It never fails. I hope it's useful for you, too.

To accomodate for this, Apple is relying on developers to make routing apps on top of the maps app to provide for this. For some, this means a minor inconvienence. What we'll likely have to do is, upon upgrading, install the Google Maps app, and it will likely provide the right public transit routing directions for us, provided that Apple allows Google to provide this service on the phone.

This may be a big blow to open data and civic interoperability. Your transit data today is powered by data called GTFS -- General Transit Feed Specification, and for the usr, it's likely the biggest open data success story since GPS and Weather. Pioneered by the Portland area's TriMet transit agency, and Google, GTFS was an optimal solution for open data: an open file specification that provided so much value to the end user that people in urban communities demanded it from their transit agency. By opening up data, a transit agency could let riders plot out destinations using mass transportation in their city in tools they were familiar with like (yes) Google Maps, and Bing.

The other great thing that GTFS did was kneecap the regional monopolies of vendors who built businesses out of keeping transit data "proprietary." Today, for instance, Nextbus claims that real time bus location data belongs to them, not WMATA -- and whether or not you agree with that business stance or not, it's fairly clear it's better for the public for that data to be as free and open as possible, rather than locked up inside of the vaults of some proprietary vendor that may or may not go out of business.

The problem, I fear, is that we'll fracture the transit data community back to where it was a few years ago -- where it was dependent on regional feifdoms of data and transit providers. It's not as though our transit authorities have a particularly good history of working together, or that cities think about interoperability. My wallet has a SmartTrip card in it for DC, a Charlie Card in it for Boston, and a Clipper pass for the Bay area, and though these cards all rely on the very same technology, they're not at all interoperable. Nothing, from city to city, ever is.

I'm afraid the user may lose out to the locked in, on the approved vendor list, not-competing-in-the-real-world-of-technology vendors will make for us sub-par apps that vary from place to place. From San Francisco and traveling to Atlanta and want to take MARTA? Download the MARTA app to complement your BART app to complement your NYC Subway app to complement your Trimet App.

I wish Apple would embrace GTFS and build some routing and scheduling software on its own. It's the kind of now critical kind of functionality that really belongs in the default. Alternatively, I'd love to see some kind of open source project emerge -- an OpenRoute type of project that really focuses on global interoperability based on GTFS, but I'm afraid the calculations of those kinds of routes on a national scale are computationally too expensive for anyone but a Google or Apple or Microsoft or what not to really build.

For now whole move reeks of Apple being more eager to screw Google than serve the user. We'll see what happens when iOS6 is out of beta.

After awhile it becomes irritating. You find yourself with a fully booked lunch/coffee/drinks calendar, dispensing advice that you're otherwise paid for, in exchange for a coke and, at best, a steak salad. Most don't mind the lunch, but left unchecked, the lunch economy can suck the soul right out of you: you spend day after day saying the same thing, dispensing the same advice, to people who then provide little or no value back. It's a cost of doing business for people who make and do things, but the return on investment for the lunch recipients in the lunch economy is remarkably low compared to the value they're creating for the lunch buyer.

The worst thing about the lunch economy is wasted time. The lunch buyer wants to ask a few questions from the lunch recipient, but has to spend time and effort to get to a place, suffer through polite small talk to get to what he or she wants out of the expert, and probably ends up not getting all the answers or value they were looking for in the first place. The recipient, feels like they must be polite, but after awhile feels taken advantage of: nobody wants to say "no I won't have lunch with you, pay me $200 for an hour of consulting instead."

The Lunch Economy exists because there's a bit of information that people want that's worth less time and knowledge than what a consulting agreement would cost, but more than the value of the person's time that's going to lunch. But it's a tax on everybody's time and patience.

So I'm experimenting with trying to end the lunch economy by releasing a minibook called How To Run Apps Contests for Government Agencies and Non-Profits. This is what most people want to take me out to lunch for -- to talk about how they can run an apps contest. Instead of going out to lunch, over and over again and dispensing the same advice, I've now got something to point to that costs everyone less time and money. If you want to pick my brain about running an apps contest, you can instead buy this book for what you'd pay for that lunch, and get way more work out of me than you'd get watching me chew.

Is it arrogant or a little obnoxious? Maybe. The goal is to make it better on both sides: to give the person looking for information what they need, and to make it so that of the 15,000 or so lunches we both have left can be spent not out of obligation, but out of genuine enjoyment. Time is our only non-renewable resource, and this answer saves us all a little bit more.

1. "Does anybody know the name of Kim Kardashian's ex-husband?"
2. "Does anybody know the child poverty rate of the county that we are in?"

To date, about 40% of the audiences that come hear me talk know the answer to the first, and not a single audience member has ever yelled out anything close to the correct answer of the second.

I ask this question to illustrate the premise of the Information Diet. It's not that knowing the name of Kim Kardashian's ex-husband is a bad thing by itself, but rather that that information is likely not very useful for your day to day life. You're not going to cause much change with it, nor would you ever want to. There's nothing actionable about knowing Kim Kardashian's ex-husband's name.

Knowing the child poverty rate (or obesity rate, or unemployment rate, etc) in your county, however, is useful information. It's something you can do something about. You can affect change with that knowledge -- either by becoming an activist, talking with your community, or better knowing how to set your priorities in your community. You're able to do something (anything) about it.

So if it's so much more useful to know all this stuff, why do people know more about the Kardashians than they do about the statistics that are vital to their community?

I agree that most of the reason why is because the Kardashian's are more "fun" to read about than poor children. But that can't be all. Another reason why is that knowing all these vital stats about a community is too hard. Trying and find out the child poverty rate, the obesity rate, the unemployment rate, who your state and federal reps are (and who they're financed by) can take all day.

The day the Information Diet came out, my wife Roz mentioned this to me: having a healthy food diet is at least possible because you can buy broccoli in the grocery store. With an information diet, all that's available in the grocery store is potato chips. If you want brocolli, you have to farm it yourself.

So that's why, over the past few weeks, I've been working on something that I hope will start to fix that, and yesterday I put in a Knight News Challenge application. The idea is that there's story to be told in your local community that isn't being told because it's too hard to put a lot of the data about your community into context.

We talk about data like it's some kind of fixed object but data alone doesn't matter that much. It's when it is put into context. Knowing your county's unemployment rate is 7.5% matters, sure. But knowing that it's a full percent lower than the national average and is decreasing more quickly? That's more interesting.

Another reasons why the Kardashian's consistently trounce poor children in your neighborhood is because of Google Trends. to meet traffic and revenue numbers, journalists turn to Google Trends to find out what people are interested in searching for and write to that. What I'm building will helpfully create a new way for journalists to figure out what to write about -- based not on what people are searching for but on what's actually happening in their community.

I find myself loving the app -- every time I'm in a new place, even in a different neighborhood here in Washington, I fire up the app to see what's going on. Because I tend to develop things out in the open, and have a hard time keeping secrets, I'll let you know you can find a link to the prototype in my application.

So if this kind of thing is interesting to you, or you think it could be useful for your daily life -- or even if you just care more about poor children than you do the Kardashians, go check out the full project description and click that little like button at the bottom.