Joel Spolsky - A Unified Theory of Internet Startups

Ahh. I get it now. Then the Unified Theory of Life is that I want to eat and poop. Occasionally, I also want to socialize and have sex. People might go about it in different ways in different times and cultures, but it's really the same thing. Of course, architect-astronauting it really gives me insight into the common nature of seemingly disparate ways that people live.

Entertaining, but I really think it's as simple as: all computer programs read data, manipulate data, or output data. Internet startups create one kind of computer program. Text is one kind of data.

Darn, someone in the comments section of the post beat me to the best zinger of all. Please see Joel's 2001 post titled "Don't Let Architecture Astronauts Scare You."[1]

"When you go too far up, abstraction-wise, you run out of oxygen. Sometimes smart thinkers just don't know when to stop, and they create these absurd, all-encompassing, high-level pictures of the universe that are all good and fine, but don't actually mean anything at all."

[1] http://www.joelonsoftware.com/articles/fog0000000018.html

I love how the web is pretty much just a big printf statement.

Or:

"How it's still totally profitable to provide services that do nothing interesting at all with user data."

Or:

"How highly praised start-ups are often just big ugly interfaces."

And sometimes, as is the case with both Twitter and Facebook, they still don't get their easy shit right. The amount of terrible UI or just simply bugs in Twitter is a joke.

"Yeah but scale blablabla"? Usually has nothing to do with shitty UI and awful front-end coding.

I'm sorry I clicked through to read that.

Large parts of the Internet are simply a media clearinghouse, used to transmit and store information between parties with little value add beyond providing interfaces. Who controls the media and how it is structured, and how it addresses its audience, is usually the interesting part. Throw in some aggregators (which mine and organize the clearinghouses), some with search indices, and online shopping and I think we've this thing covered.

EMail (SMTP), Usenet (NNTP), Web {Forums|Blogs|CMS|etc} (HTTP). All made from the same stuff. But the interesting part of a book isn't usually that it's made of paper, nor is it interesting to compare two books on that basis.

Joel's abstaction is ironic considering it comes from the person who criticized architecture astronauts for excessive abstraction.

Joel's post is really about levels of abstraction (well that, and being snarky.) At a high enough level of abstraction, what he's saying is precisely true (iamwill points this out in a more aesthetically pleasing way[1])

Some people complain: "Images are not text!" But at a high enough level of abstraction, images are text, but interpreted in a special way by special software that someone else wrote.

[1]: http://news.ycombinator.com/item?id=2867214

There are so many examples where this description does not really fit... Think about google maps for instance.

If he's angling for a patent, I have prior art.

Well that is obviously a mockery and a good one at that.

Here is a simpler theory: "Internet startups always do something with the internet." Less funny but more unified(?)

The best startups are those that show their data in a way that feeds into some psychological need, or become immensely useful as more and more people use it.

Anyone can also create a text box where you can type in 140 word status messages, but noone would use it... but once you design it so that it looks like a feed that makes you look like a Wall Street stock trader WHIZ, everyone loves it.

Anyone can create a Facebook, but it's worthless until you got millions of users. Same with Quora. Anyone can build a Q&A, but noone would answer any questions if there ain't people to impress with your knowledge.