Simply Amazing!!

Any plans to open source Stacker?

Not for the time being. Stacker is specifically developed for this unique UI style that we're using for Next. We don't open source our UI elements.

Twitter open sourcing bootstrap just made the whole web better. Open source projects especially. Projects like https://github.com/dcramer/sentry benefitted a lot from bootstrap.

Making UI open source helps everyone.

You know who did open source their UI elements? Twitter, with Bootstrap. And you know how awesome that was for everyone? Pretty awesome!

While I agree that Bootstrap is awesome (I use it myself), it's worth pointing out that what Twitter open sourced was not the magic of their core product, but rather the templates that they use to create internal applications. If you look at twitter.com, you don't see any of the same elements that are included in Bootstrap.

You mean besides all their UI widgets, right?

http://twitter.github.com/bootstrap/javascript.html

It's probably not that hard to reverse engineer what they are doing with Stacker.

Hmm.. Rspec + autotest + Growl

1) This installs EGOTextMateFullScreen plugin (https://github.com/enormego/EGOTextMateFullScreen). I didn't write it and author didn't put a package out.

2) URL Shortening? I didn't have time to write the CSS to format my code on the page. The long links breaks the layout

3) How does it matter how its done as long as it works, especially when I spent more time responding to your comment than writing that script.

> How does it matter how its done as long as it works

Executing unknown scripts hiding behind obfuscated URLs is generally not considered a good idea.

Given that this is a very non-standard and decidedly odd way of distributing a TextMate plug-in, a large degree of skepticism is warranted.

I also was skeptical. A quick read of the contents of the shortened urls reveals nothing nefarious, however.

Did you read the contents in your standard web browser, like I did? What if it sent back a different set of commands if the user agent matched that of cURL?

Well, you're still free to curl the contents of the URL into a standard text file, and view it with the text editor of your choice (maybe even TextMate!).

Updated the blog post

What matters is that it's a terrible hack. Even ignoring the security concerns with "install this completely untrusted code from the internet": there's no way to cleanly recover/uninstall if something breaks; there's no way to tell what version you have (not even in theory, as this clones HEAD!).

Most importantly, there's no chain of authority here. In the Linux world, for example, your packages generally come from the distro and are signed. Down a layer, they might come from a third party repo (rpmfusion, say), which is still a large organization with high visibility and good auditing. Down farther still, there are tools like Launchpad or openSUSE's OBS which allow you to build installable pacakges of your own, but these are still distributed out of a managed infrastructure and your identity is reasonably tracked. Finally at the bottom are the people ("developers") who like to pull raw source code and compile it. These people are expected to be communicating as part of a project, so they can be warned about compatibility goofs or (goodness forbid) the occasional malware incident.

This "pull and install automatically" gives you the ease of use of the top level, but an even weaker promise of authority than even the bottom level. That's a bad thing.

Have you tried using Missing Drawer Plugin for TM? I updated my post with details.

Hi Tobi, I was wondering how you guys are doing multi-tenancy.

True that. Even LOLCODE can scale :) Twitter's problems are very specific to twitter - Mega throughput (~7000 Tweets per second) in realtime, sharded DBs requiring multiple connections etc.

Ruby will give you a significant time to market advantage especially if you are a startup.

My understanding is that the particular scaling problem of twitter is high fanout through subscriptions. Receiving 7k messages per second and storing them in a database is actually fairly straightforward.