Weblog Without Honor or Humanity

Joel on Groove thinks the same:

“Why do you keep picking on Groove, Joel?” Three reasons:

• They have an interesting architecture that provides important platform functionality which I could really use in my own products,
• They make it impossible (or at least unrealistic) for ISVs to build on their platform, thus dooming themselves to oblivion, either out of greed or because they think Groove is an application, not a platform,
• and Groove inventor Ray Ozzie has a weblog, so he can answer me if he thinks I’m off the bat.

How to make money on the Internet (and anywhere else). Art of money making is a short, old and very delightful book about making money. It tells simple truths and it’s amazing how insightful those simple truths can be. Read it, it won’t take you long.

Hard-earned Apache knowledge: .htaccess files (and RewriteRules it contains) won’t work if AllowOverride is set to None for a given directory. The brute force solution (and probably strongly not recommended) is to set it very permissive for all directories, like this:

<Directory />
    Options FollowSymLinks
    AllowOverride All
</Directory>

This goes to httpd.conf. Of course hours were wasted while looking for elusive problems.

Bootstrapping BitTorrent. BitTorrent is a great example of good technology born outside of a big corporation. BitTorrent solves a problem that is undeniably present. Do you remember what happened when Ken Starr’s report on Clinton was released? It was only available from one place and many people wanted to read it. The end result: highway-like congestion. Everyone was fighting with everyone else for the right to download the document and because of that no-one could. The pipe wasn’t fat enough. The problem is that those who provide downloadable goods pay a hefty price for their success with bandwidth costs. Those who download get lousy service. Caching proxies were an attempt to solve that problem but they are not deployed frequently enough to make a difference. If you can pay, Akamai can solve the problem for you by providing many distributed, fat pipes. BitTorrent solves the problem differently: using swarmed downloads. What that means is that those who download will share the burden of distributing the content with the original source of information. Result for the users: faster downloads. Results for providers of downloadable files: reduced bandwidth costs.

So there’s a real problem and a real solution. Interestingly, the solution didn’t originate within Apple or Microsoft. Mostly because it’s not really their problem. They can afford to pay for their bandwidth. Their biggest problem is how to suck even more money from the costumers, user convenience is secondary. In a way it’s a shame because this technology would have the greatest impact when integrated with the OS. Otherwise gaining widespread usage will be tough since such tools have acute chicken-egg-problems. Both those who provide and those who download must start using BitTorrent for the system to work. But neither has incentive to do it given that the other party is not ready. One way to bootstrap usage of BitTorrent would be through enlighten self-interest of some large-scale provider of downloadable files. There are quite a few of those e.g. SourceForge, ibiblio, PalmGear, many places that offer shareware/freeware software. They pay huge bandwidth bills. Why not make BitTorrent usage mandatory (or offer it as an option, promising better download speeds)?

My prediction is that sooner or later similar technology will be widely used. It might not be BitTorrent but it will be something. It just makes a lot of sense.

Hard-earned PHP wisdom: urldecode() doesn’t decode ‘ and “. At least not in version 4.0.6. I had to write my own function:

function myUrlDecode( $str )
{
  $res = urldecode( $str );
  $res = str_replace( “‘”, “‘”, $res );
  $res = str_replace( ‘”‘, ‘”‘, $res );
  return $res;
}

Bookmarks out of control. My bookmarks are totally out of control. I bookmark a few sites a day, just so that I can get to them later (in an unlikely event that I’ll actually need to do it). But I just add them to the very long list without trying to arrange in categories (both IE and Mozilla don’t have good tools for that) and that doesn’t scale. There must be a better way.

The future is here, it’s just not evenly distributed (or a few thoughts on selling software). Some ideas are so obviously good that you might be sure they’ll take over. One such idea was selling bits on the Internet. There was a time where people were selling software via catalogs, had to ship physical goods (diskettes; when was the last time you used a diskette?) and accepting payments was a pain. Then came the internet and a few started offering downloads instead of diskettes and accepting payments on-line. The future was there, the writing was on the wall. This way of doing business was so much better that you could reasonably expect that everyone will switch to it. And so they did. I think that today we have another thing brewing up in the domain of software marketing: animated demos. The first animated demo for a software product I saw was a demo for CityDesk. It’s done in Flash (a good use for Flash, would you expect that?). Another one is a demo for ClickTracks. The idea just clicked with me. This is a powerful method of showing your software to potential customers. People are very reluctant to part with their money. The idea behind marketing is that you can convince them to buy your product using finely crafted ads. Well, now you can show people how good your software is. It’s not a replacement for advertisement, it’s not a replacement for manuals but it’s a great complement of those. It can give you an edge over competition. If I was selling software I would start making those demos and putting them on my web page. Now. 

Parenthetical remark: truth to be told this is not a new idea. I remember in very old days Lotus Cam (I think that was the name) allowed people to capture a particular session with software and create a stand-alone demo that would replay that session. I never seen them used much. Today we’ve reached a breaking point for two reasons:

• better tools to create demos; I’m pretty sure I could create such a demo myself if I tried really hard
• better ways of showing demos to people. If it’s Flash most people are already able to see it. If it isn’t, installing a plugin is not that hard those days

Where the Gods fight…. A e-mail discussion between John Gilmore and Jon “maddog” Hall about .ORG bids. A few low points where they try to prove to each other whose is bigger (now I understand why penis enlargement isn’t popular with geeks; who needs a pill if you can grow 3 inches with just a well-crafted put-down) but otherwise an excellent read. I found out that I can buy domain for $12/year at Joker.com (I’m one of those chumps overpaying $35/year using register.com and easydns.com; which just proves that markets (consumers) are far from well-informed and efficient (the economic theory goes that in well-informed, efficient markets all prices will go down to barely make a profit).

A new hotness. Here’s an example of something I strongly believe in: you can always do better. Furthermore, a way to make money on software is to produce better, unique apps. You would think that web server log analysis should be a solved problem by now. The market for good, useful tools is huge and the problem is old. In the past I used log analysis software and I wasn’t impressed. Two leading free tools (Analog and Webalizer) are doing industrial-strength but very uninspired job. They are old and busted. They are Ford Ts. They don’t go beyond producing basic stats and occasional graphs. (note: I haven’t used commercial tools, maybe they’re much better). I felt that it could be done better, that we need a Ferrari of log analysis tools, a new hotness. Unfortunately I didn’t know how to build this new hotness but when I saw it I knew this is it. I’m not going to bore you with the details, just see their 60 seconds demo. I was blown away by simplicity and usefulness. It really takes analyzing logs to the next level.

Interesting: the author of Analog works there. Is this really a 3 person company?

One painful thing: they use “patent pending technology”. Ouch. No matter how smart they are and how great their ideas are, this software doesn’t do anything a smart developer couldn’t figure out. It’s just extremely well executed good idea. Good but not earth-shattering. Parsing log files isn’t a rocket science. I hate patent system for encouraging this land grab mentality in people.

Information business as a relationship. Bruce Sterling at his best. His analogy of information business and relationships is great.