A set of slides from a Flickr talk at Web 2.0 Expo called ‘Capacity planning for web operations‘.

A good overview of Flickr’s setup. Nearly all of the techniques used by Flickr can be applied directly to Amazon’s cloud EC2/S3 services with faster scaling, at a lower initial cost.

Subtraction points out a large Flickr set of user interface elements, a slick repository of web interface element metaphors. I love pointing developers at good examples, and then watching their imagination explode.

Make online scoops Flickr’s new upload screen (and other enhancements). Sweet.

Flickr’s slideshow viewer now does full-”screen”, an improvement that I’ve been hoping for.

Swivel is like Flickr, but for data trends (via David Ascher).

A set of design patterns in pictures. Metaphors are cool (thanks pk).

Some slides on Flickr’s architecture. It’s all sensible stuff, with very little magic: Php4, Mysql (innodb), ImageMagick, Perl, Java (for nodes), RHEL, and so on.

A Flickr-like tool for words.

I was talking about web appilcations with a friend tonight, about how certain sites hit that sweet spot. What is it that makes their stuff better? What is it that we see that defines the difference?

So what makes them special?

• Google: authoritative answers
• Flickr: inspiration in photography
• Wikipedia: knowledge out the wazoo

The smaller sites are similar too, though with a narrower focus:

• 37Signals: simple life hacks
• JoelOnSofware: interesting software rants
• BoingBoing: compelling wild and weird stuff

It’s interesting that most of these guys are third generation web, and they all eclipsed their competitors by being damned good at their shtick. Being best is far more important than being first, and picking something we all want or need is absolute.

In product-land, the iPod is another clear winner. It does tunes. It does them well, end-to-end. No fuss. Like Google or Flickr, the iPod is clearly the best, with a limited set of features.

You’ll also notice that big success lacks the sickly-sweet corporate fermentation process. You won’t find meaningless mission statements or shallow product vision papers. What you find is clear, pragmatic thinking. Make it simpler. Make it better. Do one thing well. Damn the competition, because we’ll kick their ass.