And so, even in the year 2011, if your weblog gets Fireballed, there’s a good chance it won’t be able to handle it. That seems crazy.
That is crazy. But here’s some anecdotal evidence that suggests going back to static publishing isn’t necessarily the answer. I got Fireballed last year and TypePad (the app that serves my blog) didn’t so much as break a sweat. I don’t know all the ins and outs of how TypePad handles load balancing, caching, and so forth but I know a lot of great engineers put a tremendous amount of work into building a platform that can handle that kind of traffic. (I should probably also mention that I’ve been working for the company that owns TypePad for almost seven years.) :)
As one of those TypePad engineers, I can verify that the site is entirely dynamic.
The (not-so-)secret sauce: caching, a lot of hard work on optimizations both in the data stores and in rendering, and an absolutely top-notch operations team that keeps systems tuned for performance, corrects failures before they take us offline, and lets us know where the problem spots are so we can fix them.
Simmons correctly points out that, with a static publishing system, it takes a lot more traffic to melt a web server delivering files off a disk than it does to kill one serving each request out of a database through a rendering engine. I'd like to note that there will still be a point at which that static file server falls over, and there eventually will not be anything to do save throwing some more hardware at the problem. This, I feel, is akin to arguing against using a modern multitasking graphical OS because DOS 5.1 boots so much faster on your quad-core Xeon box.
Publishing on the web in 2011 is easier than publishing on the web in 1999, because a whole host of companies have already solved the hard problems like scaling up to deal with a slashdotting and have the resources to throw hardware and engineers at these problems.
Simmons remarks (somewhat snidely, I feel) that he's "talking about slightly-geekier-bloggers, the type who install a blogging system on their own server or on their shared space at DreamHost or wherever", and holds those bloggers up as the kind he's more likely to read.
Really? In the age of Twitter, Facebook, and Tumblr, people are still comparing virtual dick sizes about how much work they put into blogging? As if how you get your words on the web is more important than the words themselves? Guess I should only befriend people who serve their own social networking profiles off their personal web servers.
A lot of my favorite bloggers are here on TypePad, or on WordPress or Tumblr. I don't really care.
Publishing should get easier. Being a blog snob is something that only hurts you, when you won't read about the next big thing because the guy who wrote it didn't serve it to you from his personal web server on a silver platter.
To recap: thousands of sites deal just fine with massive amounts of traffic to their completely-dynamic systems through the clever use of caching. Asking for a return to the static web is asking us to give up all of the incredibly amazing things we can do on dynamic sites to solve a problem that doesn't really exist. Why should we go back to 1999 when 2011 lets us do so much more without giving anything up?
UPDATE: Apparently people are still going on about this. Simmons, again:
Twice today I’ve tried to go to sites that couldn’t handle the traffic from Daring Fireball. This is preposterous.
But he still makes the title about static publishing, which is still missing the point.
Meanwhile, one Mike Watts has decided to reinvent the wheel with a new static-publishing system written in PHP:
templates are very simple, staticDimension will simply replace $PAGE_CONTENT with your html text and projects with the title of the page. Any navigation you need must be built into the templates or page content. [...] there are no comments - if you want to use comments, you could use facebook comments or disqus and simply add their code to your template.
So this staticDimension system does way less than what Movable Type 1.0 did ten years ago, and it’s supposed to be a better way to publish your blog? To hell with that.






£180 for a commercial license. I admire that dude’s chutzpah. :)
Posted by: Walt Grayson | 04/07/2011 at 04:21 PM
Isn't the cleverness required the issue, though? TypePad can hell of scale because there's an operations team (and zillions of programming hours) making that system work. If you're self-hosting, you'd rather be clever about your job than figuring out all the operational requirements and being on-call for your own site. Stripping down to the simplest thing that could possibly work seems like a good strategy for that.
Posted by: markpasc | 04/08/2011 at 11:44 AM