RobbieAllen.com > Blog

One word to describe Saturday: Wow. I’m getting so much more out of this year compared to last. Last year I attended sessions I thought were most applicable to my day job - this year I’m attending sessions that are most interesting to me personally. Apparently there is a difference. I mingled and had interesting conversations with a variety of stars in the open source community. This will undoubtedly be a long post as I detail each session I attended. I can’t possibly write everything I saw or heard, but I’ll do my best to highlight the interesting bits.

Breakfast

At breakfast I found myself at the table of Andy Hunt, Dave Thomas, Martin Fowler, and Jack Herrington. Andy and Dave are the Pragmatic Programmers and started the Pragmatic Bookshelf (small publishing company). I pinged Andy prior to Foo Camp because I’m interested in the process of taking a final manuscript to print. (I have a reoccurring inclination to start my own publishing company some day.) It turns out that Andy lives near me in Raleigh.

All of us had our laptops open checking email and such. I had a ThinkPad and the other four had Macs. The conversation ventured into various applications people were using on their Macs and how it appeared that the majority of Foo Campers used Macs over PCs. Last year the mix was roughly 50% Mac and 50% ThinkPad/Dell/Toshiba/misc. This year it looks more like 80% Mac and 20% other. Apple is really getting a stranglehold on the high-end user. On an interesting side note, I heard tons of people complaining about their Macs this weekend. Everything from a pdf taking 5 minutes to open on Mike Hendrickson’s Mac to numerous complaints about the various Mac email clients.

The conversation eventually turned to writing books. Jack said something about how O’Reilly should put more of their books, such as Cookbooks, online. I told him that I pitched the idea of a Cookbook website to O’Reilly a year and a half ago, but there hasn’t been any traction yet. However, there have been some recent developments which seem promising. I told Tim recently that either O’Reilly creates a cookbook website soon or I’m going to give it a go on my own in some form. He reassured me that O’Reilly is committed to creating a cookbook website now (they’ve already started a Hacks site) and will bring me in on it.

Web 2.0 Meme Map - Tim O’Reilly

The first session of the day I attended was Tim O’Reilly’s Web 2.0 talk. Tim said he didn’t have much time to prepare due to the late session he gave last night on tech trends and data visualization. What he wanted to do was have an open discussion about what sets Web 2.0 apart from Web 1.0. He gave a handout with Meme Maps that described the Free Software and Open Source movements. He wants to fill out a Meme Map for Web 2.0. Some of apps that fall into the Web 2.0 category include Flickr, Google Maps, Gmail, Google Adsense, Wikipedia, and Greasemonkey. Some of the concepts (or bumper sticker phrases as Tim put it) include “The Long Tail”, “The perpetual beta”, “Users as co-developers”, and “Data as the Intel Inside”. Core competencies include “Services, not packaged software”, “architecture of participation”, “Remixable”, and “Software above the level of a single device”.

There was a lot of good discussion on this and Tim’s Map was pretty full by the time we were done. It would be cool if Tim posted the final map, which might help educate the masses on what he means by Web 2.0 (essentially the next gen web).

Some interesting folks that attended the session include Mitch Kapor (OSA Foundation and Lotus), Ward Cunningham (Wiki), Paul Rademacher (HousingMaps), and Dan Gillmor.

Why Software Sucks - Scott Berkun

I was looking forward to this session, but the turnout was pretty low, which resulted in limited diverse discussion. Scott started off by asking everyone to give him reasons why software sucks (which he wrote on a board). My contribution was that there are no blueprints for software and software is hard to get your head around which limits the learning you can accumulate over time (unlike more concrete things like hardware). Oh, and it is hard to visualize too. Larry Wall attended the session and I could feel the pains of Perl 6 coming out when he suggested “changing requirements” and “you can’t please everybody” as reasons, but I’m not sure those are specific to software. “Complexity” and “unmaintainability” were a couple other reasons. Andrew Singer and I both used “constructing buildings” as an anti-example. I personally don’t like “complexity” as a reason software sucks. Building a 100 story building is definitely complex, but it has been done successfully for a long time. The big difference in my mind is maturity. We’ve only been writing software for the last 25-30 years.

One interesting point that was made is that there are two sides of software, the user experience and the technology (I called them “users” and “systems”). Many software projects start having problems because they are focused on the systems and not on the users. Scott said that software process should start with the users and work its way down. Larry said that often users provide conflicting requirements or don’t know what they really want so in those cases you have to work from the technology side. Ultimately, most agreed that the process is a negotiation between users and technology: you should first start with requirements from users, determine if it is feasible with the technology, then go back to the users with what can and can’t be done, etc. until you ultimately reach a balance in the middle (although the more slanted you can be toward users the better.)

BitTorrent and the Future of Media - Bram Cohen

I got to this session late and the room was already overflowing. This topic wasn’t of major importance to me, but considering all the media attention BitTorrent has received over the last year, I thought it would be interesting to see what its creator had to say.

Bram is an interesting fellow - young and sharp, but would obviously stick out in a room of media execs with his all black attire and easy laugh. Bram made a few interesting comments. He thinks radio and TV will eventually go away. “TV is good for aliens”. TV and radio cannot target users very well, unlike what you can do on the net.

The discussion was dominated by Bram, a guy from the BBC, and a guy from the MLB. Apparently the BBC and MLB are doing quite a bit of media distribution in various forms. The guy from BBC (sorry, I don’t have his name), told the story of how the BBC Symphony released their rendition of a Beethoven piece freely to the public. A lot of classical labels jumped all over them for this because they didn’t want a government funded organization releasing something that could hurt private business. The discussion went in many different directions but touched on all the main points you might expect: DRM, decreasing distribution costs, how to make money giving content away for free, etc. Bram is working on a commercial model for P2P.

About half way through Jeff Bezos showed up, but didn’t have anything to say. I left before the session ended to catch up on some email.

Lunch

At lunch my table was occupied by Dan Kaminsky (DoxPara), Howard Rheingold, Doug Kaye (IT Conversations), and Brewster Kahle (Internet Archive). I overheard Brewster talking about a model in which they are trying to scan libraries all over the world. Instead of doing it solely by machine, they are trying to come up with a model where humans do the scanning.

Dan Kaminsky is a former Cisco guy and now noted hacker whom I spoke with last year at Foo Camp, so we caught up on the last 12 months. He left Avaya and was considering ISS before their recent spat of bad PR. We talked about the good ole days at Cisco when we were both too young to care much about the politics.

Then I jumped in a conversation with Doug Kaye about IT Conversations. I’m a big fan and Doug seemed thoroughly pleased at how well received his site is these days. I told him about similar sites (MIT World and Yale Leadership Forum) and how it would be really cool if there was an aggregate of all the cool talks that people give. It would be even cooler if the technology was easy enough so that anyone could contribute audio or video samples. I told him about all the speakers I listened to last year at MIT. He said they are looking at model in which users can contribute their own content and have a rating system to go along with it.

Tips for Writing Essays - Paul Graham

Paul Graham is the interesting chap that has written several popular essays and the very successful Hackers and Painters. Since I’ve written numerous books and articles, I’ve recently considered venturing into the boundless world of essays. This was a great opportunity to hear from one the best technology essayist of our generation.

Here are some of my notes:
- It takes him more than a week of full time work to write an essay
- He doesn’t make an outline
- He gets a good beginning by the end of the first day of writing
- By the 4th day he has a first draft
- Then he spends a lot of time cutting
- Two types of stuff to cut: good stuff and bad stuff
- It is easy to cut bad stuff, but hard to cut good stuff
- To make it easier, he has a file he copies the good stuff he cuts so he can refer to it later (but never does)
- First time you write something down it is often too assertive and inflammatory which is why editing vigorously is so important
- Lot of time he cuts out flames because “an essay is no good if no one reads it”
- The key to “good writing” is understanding what is a digression and what requires exposition to explain a point
- Tries to make his essays like sea cucumbers (slides right down). If he gets bored as he reads an essay he knows it still needs revision
- An essay is really good if people quote different things from it. If everyone quotes the same passage, then that may be the only useful passage in the whole thing.
- He intentionally tries to make his prose quotable

Tricks he does in the last stage of writing:
- Print the thing out and read it
- Sit down with a red pen and draw lines through things. It is easier to go back later and put a little ok by a marked out line than if you do it on the computer.
- Removes a bunch of “that’s”
- Reads it out loud to catch things you wouldn’t normally

- He thinks it is amazing how often you can rip out a whole paragraph and no other surgery is necessary
- Make it as short as possible. Short sentences.
- Read it a week later because you become too familiar with it when you’ve been working on it non-stop for several days.
- Use plenty of pronouns
- He writes to help figure stuff out

- He talked quite a bit about the essays that he hasn’t released and how he has to be careful not to be too controversial in his essays. He doesn’t want to get tagged an “insert here”-ist because some dumb person misinterprets a sound argument that leads to a logical conclusion.

Overall the session was one of the best of Foo Camp. It was informative and inspiring.

Beyond Books - Rael Dornfest (and others from O’Reilly)

Rael showed off a new Wiki that O’Reilly is creating to help with writing Hacks books. The intention is to help automate some of the writing/editing process by putting the content on a wiki. Good idea. Many improvements can be made in the publishing process today and I’m glad O’Reilly is finally starting to do something about it. O’Reilly is storing all the Hacks and Cookbooks in a database in a nice structured way, which makes it easy to reuse the content later (such as for a subscription based Cookbook website I spoke of earlier).

After the talk I approached Tim O’Reilly about the need for an author portal and more focus on improving the author experience. He recognized the need and agreed, but said it is a matter of resources now. They’ll get to it eventually. Sidenote: I just noticed the other day that even Wiley has an author portal. O’Reilly is far ahead in some regards but far behind in others.

Ice Cream Break

Next up was the Ice Cream break. They provided vanilla or chocolate with a variety of toppings. Good stuff. I was in line with Andy Hunt and he commented how last year the ice cream was hard as a rock which caused the line to back up while the servers tried to scoop it out. They had the same issue this year too! Two fairly young girls were struggling to scoop ice cream out of these huge containers. They should have sat it out a bit before serving. Now that I think about it, I should have volunteered to help out. That would have been great way to meet a large number of Foo Campers. I’ll have to remember this in case I go next year.

Andy and I sat down and we eventually talked a bit about his publishing company. We didn’t go too deep because his partner in crime (Dave) was getting ready to present the details behind their publishing process right after the break. So that’s the session I headed to next.

Agile Book Publishing - Dave Thomas

Dave Thomas’ talk on how he and Andy Hunt run the Programtic Bookshelf was fascinating. Dave really opened the kimono and gave us an inside look at their process. It was a shame Tim wasn’t in the audience to see how automated they’ve made the process. O’Reilly has a lot to learn from these guys.

Since Dave and Andy are programmers by trade and mainly publish programming books, they’ve tailored the writing and editing experience toward programmers. They don’t even use Word to write their chapters! They use homegrown markup called the Pragmatic Markup Language which is a derivative of DocBook. There was a lot presented and I only took a few notes. I hope Dave will make his presentation available.

- First shipment Oct 03
- 10 books in print (14 by year end)
- 56,000 total units shipped (11,000 via pdf)
- Sold 11,000 units directly
- Found that books haven’t been very price sensitive (hence the large number of direct sales)
- Use subversion for source control
- They always know at what stage a book is at because they build their books every night
- They have a well-defined workflow
- A book spends 1.5 months in production (compared to 3 months for O’Reilly)
- Malloy is the printer
- Beta books have helped sales

Identity 2.0 - Dick Hardt

Dick gave a presentation about the next generation Identity movement - Identity 2.0. The presentation style was perhaps more interesting than the content. He had a set of slides that he went through really fast. Most slides had only a single picture or word on it. After each word, he’d move to the next slide . I saw this style first at the Web 2.0 conference last year by Larry Lessig. It reminded me exactly of that. On Dick’s last slide I saw a reference to Larry as being the inspiration for the style. Yep. When Larry presented last year, I thought it was one of the most moving presentations I’d ever seen. Dick did a good job, but I found myself more distracted by the style than focusing on the content. When Larry did it, he didn’t look at the slides - so they just served to augment the experience, not define it. He could have given the presentation without the slides at all. Larry used some sort of pointer to advance the slides so he didn’t really seem distracted by them. With Dick’s presentation, he looked down at his computer and advanced the slides by hitting a key on his keyboard. It wasn’t quite the same. I think it is a neat way to present and I’m sure the crowd was entertained, but it is only effective with the right kind of content and when the presenter uses it seamlessly.

Regarding the content - I’m still skeptical whether we’ll see a widespread digital identity solution in the near future. There are so many competing interests that have to play nice that it seems unlikely.

Dinner

At Dinner I sat at a table with Sean Devine. Sean and his wife recently moved from Boston to Sebastopol. He had a place down the street from where I live in the back bay (when I’m in Boston). Sean was a nice guy and we chatted a bit about the current state of Safari and some of my ideas on creating more interactive content (again, the Cookbook website idea). He got it.

The most interesting side conversation I had during the whole trip came after dinner. I was walking to the schedule board to see what sessions were coming up after dinner and ran into Paul Graham. We struck up a conversation and I told him how much I enjoyed his talk earlier. We chatted a bit about Y Combinator. I told him about the New Enterprises class I took last Spring and how three of his essays captured a majority of what was covered in the class. Next thing I know Larry Wall is standing beside us and starts talking. The topic of conversation quickly turned to language design and Perl 6. Paul is a Lisp guy and Larry is (obviously) a Perl guy. Watching those two go back and forth was very entertaining.

The Role of the First Spy Satelite

This was a different kind of topic than the typical one you’ll see at Foo Camp, but it sounded interesting and Tim O’Reilly recommended it highly. It was first presented at this year’s Where 2.0 conference. It was all about the technology behind the first spy satellites and cameras. The speaker had a lot of old photos and antidotes that kept the talk interesting.

About half way through (8:45pm pacific) I decided I couldn’t take in anymore information, so I left to work on my notes from the day while they were fresh in my head. What a day!