« February 2005 | Main | April 2005 »
March 31, 2005
CPS at Purdue article
Ed's article about CPS at Purdue has been published.
http://www.campus-technology.com/news_article.asp?id=10787&typeid=156
This is the system we've been testing in CPT 355 this semester.
Posted by mikel at 01:50 PM | Comments (0) | TrackBack
March 29, 2005
Bayesian filter for ham
I agree with Joel that SpamBayes does an excellent job on filtering out what I consider spam. I don't know why I never thought of this logical extension.
Bayesian filtering is really just a specialization of a long established AI algorithm for sorting documents based on training, and we thought, gosh, why not generalize the algorithm so that in addition to sorting incoming email into "spam," "suspect," and "ham," it also sorted all the ham into piles, for example, "sales," "tech support," and "job applications"?
It seems so obvious. I wonder if SpamBayes will add that feature for additional sorting.
Posted by mikel at 09:33 PM | Comments (0) | TrackBack
March 24, 2005
Paul Graham on Startups
An interesting article by Paul Graham about startups
http://www.paulgraham.com/start.html
A few select quotes.
So as a rule you can recognize genuinely smart people by their ability to say things like "I don't know," "Maybe you're right," and "I don't understand x well enough."
It's no coincidence that startups start around universities, because that's where smart people meet. It's not what people learn in classes at MIT and Stanford that has made technology companies spring up around them. They could sing campfire songs in the classes so long as admissions worked the same.
And what I discovered was that business was no great mystery. It's not something like physics or medicine that requires extensive study. You just try to get people to pay you for stuff.
In a startup, your initial plans are almost certain to be wrong in some way, and your first priority should be to figure out where. The only way to do that is to try implementing them.
I'd really only disagree with the following quote and the section on getting investing.
I think it's wise to take money from investors. To be self-funding, you have to start as a consulting company, and it's hard to switch from that to a product company.
I wouldn't want to take money from VCs or Angels because they will eventually shape the way your company is run. I would agree with Paul you need to keep the number of founders small and taking on someone just for money isn't a good idea. Right now we're fighting the balance between consulting and products. But that's the good part about consulting contracts, they eventually end.
An apartment is also the right kind of place for developing software. Cube farms suck for that, as you've probably discovered if you've tried it. Ever notice how much easier it is to hack at home than at work? So why not make work more like home?When you're looking for space for a startup, don't feel that it has to look professional. Professional means doing good work, not elevators and glass walls. I'd advise most startups to avoid corporate space at first and just rent an apartment. You want to live at the office in a startup, so why not have a place designed to be lived in as your office?
Posted by mikel at 12:55 PM | Comments (0) | TrackBack
March 16, 2005
The long tail of software. Millions of Markets of Dozens.
Some choice quotes from Joe Krauss latest blog entry.
First, every business has multiple processes. Things like hiring, firing, selling, ordering, etc. Second, while some of these are pretty common in name from business to business (recruiting, for example), in practice, they are usually highly customized.
Said another way, there is a long tail of very custom process problems that software is supposed to help businesses solve.
The market doesnt like a vacuum and people do solve their software needs in the long tail. They do it using two basic tools: Microsoft Excel and email.
Whatever business your starting, think about how to serve millions of markets of dozens instead of dozens of markets of millions. Serving the head isnt a bad strategy. You can build a great business. But, figure out how to serve the tail of your market efficiently and youve got a blockbuster.
Not sure if he strategically chose that last word or not, but if should read "and you've got a Netflix"
Posted by mikel at 03:55 PM | Comments (0) | TrackBack
March 14, 2005
Free press releases
Next time we release a product, I'm going to look into prweb.com. According to the Chris Pirillo Show you can submit a press release for free. Looks pretty simple and it might drive some traffic our way.
Posted by mikel at 02:16 PM | Comments (0) | TrackBack
March 11, 2005
Install Pocket PC Cab files with NSIS
I've been using the Nullsoft Installer recently and its working out well for a desktop project. Here's an article on using NSIS to install CAB files on the Pocket PC. Useful for Swoosh and also a good reading assignment for CPT 355
http://www.pocketpcdn.com/articles/multicabinstall.html
Posted by mikel at 09:02 AM | Comments (0) | TrackBack
John Hollinger
I've recently started reading some of John Hollinger articles on the NBA that have a very stats oriented take. David just linked to the Hoiberg article and that got me looking around. Looks like John's site is currently without much info, so I'm just stashing this here for later.
Posted by mikel at 08:47 AM | Comments (0) | TrackBack
Rebound Rate
A possible new stat for Swoosh
Rebound Rate - While rebounds per 48 minutes is a very good way to evaluate rebounders, you can do even better by taking into account that some players have the opportunity to grab more rebounds than others. The most fair way to evaluate rebounders is by percentage of all missed shots when they are in the game that they rebound. (This is usually estimated by their team's and opponent's rebounds per minute). This is known as Rebound Rate or Rebound Percentage (Reb%).
http://www.nba.com/sonics/news/stats101.html
Posted by mikel at 08:41 AM | Comments (0) | TrackBack
Verizon DSL outage
My Verizon DSL went out yesterday afternoon and didn't come back on until sometime last night. Thankfully it's working this morning. I can't find any news on how wide spread the outage was or what caused it. All the tech support line had last night was that there was an outage in the 765-564 area.
On the other hand, yesterday I signed up for fixed point-to-point wireless at the new house in Camden with Fairnet. They'll be out to the house to do a site survey on March 22 so hopefully everything goes well with that. Should be faster than my DSL (1.5 Mb down and 512 Kb up) but the connection is obviously shared. Hopefully there aren't lots of people in my neighborhood using it. I was surprised to find out that DSL was available in the little town of Camden but it was more expensive when you factored in that we only need the landline phone for DSL.
Posted by mikel at 07:36 AM | Comments (0) | TrackBack
March 01, 2005
Web service exception handling
Some instructions on handling specific exceptions from a web service (MapPoint in this case)
Posted by mikel at 11:47 AM | Comments (0) | TrackBack