FriendFeed Changelog: see what code we are writing(blog.friendfeed.com)
blog.friendfeed.com
FriendFeed Changelog: see what code we are writing
http://blog.friendfeed.com/2008/02/friendfeed-changelog-see-what-code-we.html
14 comments
Here's the actual changelog, btw: http://changelog.friendfeed.com/
I've considered doing this for a while, too. I went to far as to make a post-commit hook script for subversion to post to twitter, but it was friends-only so only my collaborator and I could see it.
Even better would be to post the diffs :-P
Would people be interested in seeing the svn changelog on my current project?
Even better would be to post the diffs :-P
Would people be interested in seeing the svn changelog on my current project?
Very cool!
PS: paul, what do you guys use on the backend? I see you guys use nginx (great choice). What else?
PS: paul, what do you guys use on the backend? I see you guys use nginx (great choice). What else?
"Paul recently suggested that we could save ourselves a lot of time writing blog posts if we just sent out our changelists from Mercurial"
Wow. I mean, wow!! This is a great way to get technical people invested in the development of your app!
Wow. I mean, wow!! This is a great way to get technical people invested in the development of your app!
choice exerpts:
changeset: 50bb94a18270
user: bret
date: 2008-02-26 13:23:50
description: Undo shitty code.
---
Sat, Feb 23: Let's redo our entire UI two days before launch
I'd love to set an ethographer loose on this, and see what insights they have about how programmers and programming works. Might make a nice PhD thesis for someone :-)
There have already been a bunch of academics studying the version control histories of major open source projects, for example as a means to gain insight into why open source development "works" (i.e. is able to produce very high-quality software). One of the cool things about studying an OSS project is that everything is public, and the vast majority of the communication happens via an archived, digital medium: you can correlate VCS history with mailing list activity with bug tracker activity with website traffic logs, for example.
An example paper related to this idea is http://opensource.mit.edu/papers/lakhanivonhippelusersupport... ("How Open Source Software Works: "Free" User-to-User Assistance", studying Apache)
An example paper related to this idea is http://opensource.mit.edu/papers/lakhanivonhippelusersupport... ("How Open Source Software Works: "Free" User-to-User Assistance", studying Apache)
When I think about doing this myself I immediately start coming up with flimsy excuses for why I can't. I believe a lot of doing a startup well is daring to put yourself out there to be judged. This is right in line with that idea (to an extreme). Brave and commendable.
damn, that's a lot of commits. I only usually make 3-5 a day, but then again I only work on my work stuff 8 hours a day, too.
I find it depends on the type of work I am doing.
Creating a new module - a daily commit. Bug fixing - a zillion commits. GUI work - a zillion commits.
Since it was right before prelaunch, and presumably a lot of GUI/bug fixes, seems right.
Creating a new module - a daily commit. Bug fixing - a zillion commits. GUI work - a zillion commits.
Since it was right before prelaunch, and presumably a lot of GUI/bug fixes, seems right.
modern cold war