(Disclaimer: My employer is about to deplay its first Rails application, and we admins are somewhat apprehensive about the performance implications.)
Rails is an ogre, and ogres have layers
Previous post: Euro License
Next post: Sex and Drugs
I'm a sucker for girl groups.
(Disclaimer: My employer is about to deplay its first Rails application, and we admins are somewhat apprehensive about the performance implications.)
Previous post: Euro License
Next post: Sex and Drugs
{ 3 comments… read them below or add one }
Which parts of the rails application does not scale?
@Lars: I have no idea, I linked to the Pilgrim piece purely for its entertainment value. It’s a parody of DHH’s PR skills, which IMHO rank right up there with Matt Mullenweg’s.
Seriously, I think Kellan had the best take on the original interview: Rails scales development time, not performance. That’s the whole idea.
I imagine that selling used cars will not require a Mongrel cluster of twittery proportions, and that in our case Rails will work just fine.
PapaScott. Good point, however I think that’s not quite right either to say we’re at a development time/performance tradeoff.
One the side we are talking (not downloading Wordpress or Wikimedia or SugarCRM to solve our problems, but the build a website from scratch side), Rails is in the 90/10 pareto principle.
Rails does the 90% of your code that takes 10% of your time.
Of course that could be the 80/20 or 70/30 or whatever depending in the problem. At a certain point it’s going to be worth it to use Rails. We all know that.
But what point? A long lived highly scalable web 2.0 application that is in the top 100? I tend to believe that’s somewhere around 99/1 than 70/30.
Thankfully, Twitter is nowhere near the top 100.