Sunday, June 11, 2006
Home Sweet Home
Tammy and I are home, safe and sound. I'm betting I'll be online playing Halo2 via Xbox live so look for me--houseofwarwick is my gamertag--and let's hookup for a game.
Robert Scoble leaves Microsoft, joins PodTech startup
Robert Scoble, the world's most humble blogger, confirmed rumors that he's left Microsoft to join PodTech, a Silicon Valley startup. I've met Robert a couple of times and only know him through his weblog, but I'm happy he's made the jump. We'll all benefit from his "everyman" influence on the industry.
Congrats!
Friday, June 9, 2006
In Des Moines Again
Tammy and I are in Des Moines again, this time for a jewelry show that Tammy's doing for a new retail outlet. It's only for two nights, but we're already talking about where we're going for dinner. Tonight it looks like thin crust pizza. Yum!
Thursday, June 8, 2006
A happy Civic and happy owner
I just picked up the Civic from the local dealer after it went under the knife for some serious maintenance. I had a package done where they replaced the timing belt, water pump, all drive belts, flushed the radiator and replaced the coolant, changed the oil and filter and even did a valve adjustment. It's the most significant piece of regular service that you can do and I'd venture a guess that it's the most important.
The car feels brand new, pulling strong on a highway onramp, going much faster than it used to in the same amount of time.
I'm heading out to the highway now to put it through it's paces again.
Wednesday, June 7, 2006
My Experiences with Halo Multiplayer via Xbox Live
I've been playing Halo for about six months now and joining online games via Xbox Live for the last five months. The online component of the game makes all of the difference in the experience, allowing you create a friends list of gamers and track their status easily.
Halo online is very addictive. You are constantly ranked and compared to others, driving the perception that you're getting to a higher level "real soon now". Right now I'm ranked a 10 which is a lot harder to get than you think. The jump to higher ranks, say in the 20s will be hard. You have to learn to use the sniper weapons effectively and also prosecute kills quickly and in rapid succession.
For example, one of the game types is "free for all" where you are pitted against 7 other Halo players of similar rank. You're placed on a map at random and separate locations and a time limit is imposed. Everyone starts with the same array of weapons, but there a weapons on the map that you can pick up and use. Game play consists of finding a target and "killing" it, thereby scoring a point. When the first person reaches 50 kills, the game is over. The ranking system is complicated, but generally, if you beat (kill) people with a higher rank, your rank could increase. All of the games require strategy on par with chess, but without the luxury of large blocks of time and the narrow rules of the classic board game.
One of the maps is called Lockout and it remains a favorite of players of all levels. It's not very broad, but it is tall which gives you several levels to see other players and head in for a kill. What makes Lockout unique are viewing angles and enclosed spaces. The map is crammed full of corridor intersections which are relatively small and that size affects how some of your munitions behave. Standard grenades have a more contained blast and are likely to cause greater damage, allowing quick kills with any weapon. Since the map is stacked tall, there are plenty of gangways that lead from one corridor or room to another and they give you great viewing angles to odd places on the map. Lockout requires someone with quick intellect and excellent spatial analysis or what fighter jet pilots call "situational awareness".
I'll write more in successive pieces about the technical side of battle, the social networking of Halo, it's class structure and even the fanbase. Stay tuned.
Mike Arrington wonders aloud about Google
Mike Arrington writes about Google again and near the end asks the obvious question: what's Google up to?
The commenters on the post went in standard directions for the most part, but I'll shoot this out for others to read: Google is in the behavior business.
Huh?
Google creates applications for the average person to use, but they aren't groundbreaking or even different. Instead the applications are designed to capture a user's time, record their actions and then infer behaviors. With a large base of behavior knowledge, a company like Google could then sell the information to others, sanitized of course.
I think Google is about capture and control of the attention of the Internet. That will manifest itself in several different ways, the first of these are the web-based apps. Applications so boring and normal that people don't think twice about using them, interacting with them and finally staying with them.
Investors: Google is not a technology innovator. Google is a data warehouse storing behavior profiles about people and computers. Don't invest in Google if you think they are going to kill Microsoft or compete with Yahoo! because they won't.
Monday, June 5, 2006
I'm back with UserLand plus Radio UserLand News
This went out to the Frontier Kernel project mailing list this morning:
I'm baaackkkk!Exciting news. First, I'm once again serving as the product manager for Radio UserLand. Second, I am leading an effort to use the source code base from the Frontier Kernel Project as the basis for the next version of Radio UserLand's kernel.
As you all know, UserLand is a small company with few internal resources, but a very high regard for its developer community. In 2004, the Frontier kernel was released to open source on the hope that by drawing on resources outside the company, some of the lingering problems with the kernel would finally be resolved.
Over the past several months, as I watched the Frontier Kernel Project make progress, I started pitching UserLand on the idea of using the Frontier OSR code for the basis of the Radio kernel. The Radio.root would remain private, but the kernel would be available to anyone who wanted to compile from source. UserLand management agreed and that brings us to where we are today.
Here are some of the more obvious FAQ's:
Q. What does UserLand expect of us?
A. Nothing. UserLand will download the source like everyone else and contribute changes back to the project in accordance with the license and the spirit of the project. I will be the person responsible for compiling the kernel, testing and reporting the bugs and successes.
Q. Will UserLand be contributing to the kernel project?
A. Unlikely in the short term. But by recognizing and adopting the work of the project and using the kernel to create an improved product, UserLand hopes to add momentum to both and thereby attract more interest and people to the project.
Q. What happens if UserLand wants us to do something that we don't want to do?
A. Don't do it. UserLand isn't taking over the Frontier Kernel Project or trying to co-opt it. It is just another participant and user of the OSR. If a conflict with the kernel's project develops (unlikely), UserLand will have no choice but to submit or branch (with its separate contributions being added to the repository).
Q. Who's the UserLand point of contact?
A. For now, it's me ( steve@userland.com or steve.kirks@gmail.com ) but Lawrence Lee ( lawrence@userland.com ) is available too. I'm not an employee or owner of UserLand Software, but they are letting me lead this effort and contribute to both the Kernel Project and the future of Radio. I expect to get some compensation for my trouble, but that's mostly to justify time spent on the project to my wife.
Q. Where do you want the Radio UserLand kernel to "go"?
A. What do I expect? Here's a short list:
1. Upload some missing icon files for the Mac build.
2. Work with someone to get a functioning UB build of Radio running on an Intel Mac.
3. Work with someone to get a functioning build of Radio running on Windows using Visual Studio Express C++ 2005
Past that, it will be up to the community for Frontier coders and the Radio community.
To everyone who's contributed to the kernel: thank you for making this possible.
This is very exciting news to Radio UserLand webloggers. The core "kernel" of the software can now take advantage of the kernel group's hard work. In addition, UserLand has a chance to contribute some of it's work back to the kernel group. Win-win!
Wednesday, May 31, 2006
A fantastic birthday week
Yesterday was my birthday and in all of my 36 years, I can't remember the last time a celebration has stretched out across a week's time. On Thursday of last week, we left for Des Moines, IA to spend four days with my wife's sister, her husband and their two adorable children. I was able to do simple things like sleep late and drink good coffee, play with my niece at a pizza place (run by an animatronic mouse), hunt for bargain software, buy good spices and eat fabulous Mediterranean food.
On Monday, Tammy and I shopped for mattresses and finally settled on something good. It's amazing that few people don't spend the money on a good mattress considering how much time they spend asleep. Our old mattress was as aged as our relationship and the new one is about the same as the cost of a well-outfitted MacBook. I'd rather have a MacBook, but you can't sleep on it or share it with your wife now can you?
Tuesday was the actual birthday and the gang from the office took me to lunch and I found balloons and a card waiting for me when I got back. I got some great gifts that revolve around food, a subtle hint that I need to cook more.
Tonight, we're eating dinner with some good friends to round out the week. They won't tell me where I'm going, but considering my choices and their tastes, I know I'm in good hands.
Friday, May 26, 2006
New thinking about Radio UserLand's aggregator
Radio UserLand has shared it's XML aggregator with Frontier for years. The built in aggregator is simple: fetch the XML, compile it into a outline stored in the database, then use a format driver to read the table and extract standard information.
Format drivers are UserTalk scripts written for a specific syndication format that step through the compiled outline and look for things like site name, site URL, a description or details about an item like a GUID or permalink. The format driver scripts are big and contain repetitive subroutines because while it's fun to talk about how different things like RDF, RSS and Atom are, they aren't that different in ways that user's care about. In fact, the currently shipping format driver for RSS handles all "versions" of RSS and RDF *because they are so similar*.
I posted an example of a script that Patrick Ritchie wrote about a year ago with a fresh approach: element drivers. The idea behind element drivers was you only extract the information you want by calling the element's driver. Want to know the title of that website? As a script writer, you don't have to know the format, you just get the database address of the feed and ask for the site's title:
aggregator.compile.site.title (adrServicesTable)
Where "adrServicesTable" is the address of the table in the database where the syndication service information is stored.
This script (which doesn't exist yet but I am writing it) would return the site's title based on the format of the feed. In an RSS 2.0 feed, that's the value of the "title" element under the "channel" element. In an Atom 1.0 feed, that's "title" element under the main "feed" element. As an aside, the neat thing about Atom is that it tells you what type of content is in that element. For example, it can tell me that the text of the "title" element is escaped HTML and that allows me to tell my script to read the text that way. Cool, eh.
Here's the rub:
I think the aggregator format drivers need to go the way of the dodo and we need to use element drivers. It will help us reuse code and write things that are simpler. The issue is when someone needs to write drivers for a non-standard format. We can allow for that by giving a programmer the ability to completely bypass the compile process and use their own.
Nick Bradbury says pick one syndication format
Nick Bradbury: "So, if you currently offer multiple feed formats, may I suggest that you stop doing this? Just pick a format - any format. If RSS does what you need, stick with it and dump your Atom feed. If you need the extra features that the Atom format offers, dump your RSS feed. Either way you'll be fine, and your readers will be happier."
Amen to that.
