Myles Carrick

Tracking WordPress with Git

I use Git and want to track upstream WordPress.

[UPDATE] I was using a utility here called svn2git. That was crazy. Now I’m just using git-svn and it rocks.

Here’s what I do to track Wordpress SVN. It couldn’t be easier.

git svn clone -s http://core.svn.wordpress.org wordpress_svn

The -s flag tells it you’ve got a normal-looking SVN repo. I clone it into a git repo (folder) called wordpress_svn

It takes FOREVER… replaying svn history one commit at a time.

When you’re done… have a look at the branches

cd wordpress_svn
git branch -r

To refresh it (or if it dies halfway through) just

git svn rebase

Couldn’t be easier. I have mine at https://github.com/newington/wordpress - master branch and usually at least one recent topic branch should be current.

Using Chef to build cloud servers for WordPress and Drupal Commons

So I decided this afternoon that with a wife 39.5 weeks’ pregnant I’d get into devops fun experimenting with Chef.

My goals are simple:

• give me a clear, unambiguous, repeatable way of deploying servers of various types
• help me deploy a bunch of PHP stuff I need right now:
  • WordPress with Network stuff
  • Drupal Commons

I toyed with the idea of getting someone to do it, but sometimes it’s just easier to do it yourself (read: faster and far less complicated).

Why Chef?

The reasons are fairly simple, and all about me:

1. Chef’s all Ruby. I know and love Ruby.
2. I can use Opscode’s hosted platform for the servers. The last thing I want to do is run another server or bunch of servers to control my servers - I’m trying to be quick and lightweight here.

Where am I going to deploy?

I’m not entirely sure where this is going to end up. I have some unused Linodes that have to either start getting used or die. Other possibilities are Ninefold and of course AWS/EC2.

Next post: the details…

Genuine insight from Pixar on managing effective teams

Ryan Singer from 37signals mentioned this vid from Stanford GSB

Pixar founder Ed Catmull addresses the question: “Why do ‘successful’ companies fail?”

What’s more important? Great people or great ideas?

Ed goes through some important things that have made Pixar awesome, like:

• Undergo constant review - review it all every day - push through the embarrassment; make it so that ‘when you’re done, you’re done’
• Keep it safe to tell the truth
• Communication and org structures shouldn’t necessarily match - everyone should be able to talk to anyone about anything; managers shouldn’t be precious about always being kept in the loop.
• Don’t let success mask problems
• Restructure the postmortem / retrospective - keep it fresh; allow it to be sufficiently in depth.

He also addresses the strange feature re company themes / mottos - they tend to be truisms - everyone knows they’re true; all competitors have pretty much the same one.

Sometimes having the pithy statement means we can say it and not change behaviour

There are, Ed surmises, two types of ‘crisis’:

• you don’t like what you see & have to make changes. People will have to change. It’s hard.
• you finish and it’s rubbish (game over)

The first sort is a LOT better - it’s self-imposed. It’s harder. But it might actually work.

Bucketloads of ideas here for how I manage teams, methinks.

Yes, lecture capture sucks… now what?

Fellow contrarian Mark Smithers writes on the phenomenon of widespread, automated lecture capture in education›:

the large scale implementation of lecture capture is probably one of the costliest and strategically misguided educational technologies that an institution can adopt

At UNE we’re currently in the throes of deciding our plan of attack for all things rich media. Naturally this includes lecture capture, but the peculiarities of our environment have thankfully provided the context for a re-think. In particular:

• 80% of our students study all or mostly online
• An ever-increasing number of our teaching units have no face-to-face cohort - i.e. no actual lectures to record
• The tools for academics to record great, focused learning interactions are inexpensive, powerful, and much easier to use
• We don’t have the legacy of an existing lecture capture practice
• As Mark has highlighted, a gazillion studies have shown (and our own experience often attests) that an hour-long, lecture-style video is rarely the best way to engage, stretch, develop and enhance learning.

So what are we going to do?

We’re currently working at ensuring that our focus is on (a) the tools our students tell us they want to use… and (b) those our great teaching staff are keen to get their hands on.

Thankfully, those two are one and the same:

1. We’re rolling out great synchronous meeting / classroom platform. Based on Adobe Connect and tightly integrated with our Moodle site. It’ll give us a great building block for fantastic online discussions - for tutorial-style discussions, supervision, collaboration, and more.

2. We’ve put in place the absolute best media storage and management platform. Kaltura forms the foundation for this - we’ve started with terabytes of storage and are sponsoring the open source development of the Moodle 2.0 plugin. It’ll be possible to record or upload video straight into the browser… and Kaltura will automatically take care of the formats, resizing and distribution.

3. We’re investigating the best way to give all staff (and potentially students) the easiest way to capture audio and video from the desktop and in lecture-style situations.

Excited? I am!