How Agility Allows Us to Build One Of the World's Biggest Rails Apps | Hacker News Reader

How Agility Allows Us to Build One Of the World's Biggest Rails Apps

infoq.com

21 pointsmickeyben6 comments

Open articleView on HN

Here's this talk as a PDF:

http://github.com/nealford/presentations/blob/master/Rails%2...

... from this collection of Neal Ford's presentation slides on github:

http://github.com/nealford/presentations

32k lines of code is one of the world's biggest Rails apps?

We clock in over double that, and still beat that number without counting auto-generated code (Ragel and Racc output), without counting views and templates, and without counting tests.

(We are, for the record, not proud of this number. "How did agility allow us to scale to such heights? Well, whenever we run into a problem, we fling more f'ing code at it, like a monkey with a handful of poop.")

There is no way either of these two apps --- ours or theirs --- can really be among the largest Rails codebases.

For comparison, rake stats reports BCC is about 6k these days -- not counting the Java client application.

When they reported 30kloc in the app the Enterprise developer in me immediately started thinking of "Where is the rest of the iceberg hiding?"

On the one hand, you know what's really happening here. Most of the internal enterprise Java apps we look at crush these Rails apps by LOC; intuitively, even though Java inflates the LOC metric, it's no contest even by any reasonable scale factor you'd assign. What's happening is that Rails nibbles at the edges of the market and gets relatively small problems to work on. Apps like ours are big --- for Rails apps. Not a hugely meaningful metric.

On the other hand: how many developers does Thoughtworks have on this project again?

I wish this wasn't an hour long, or had a text Table of Contents. Can anyone please give a breakdown of when are interesting spots in the video to watch?

the InfoQ article is missing the "(Enterprise)" qualifier from the original presentation's title.

Reply on news.ycombinator.com