The Paradox of Infrastructure

The weird thing about coding a project management application is that you have to rely on a support structure that you wish had been written already.

One of the things I neglected to include in my initiation/requirements form a few days ago was a source browser or code review system. I think that's probably a part of the system, but I didn't want to commit to it right away. I have some interesting thoughts about integration, anyway. But even though it's an eventual feature of the system, developing the system requires it exists right away.

Clearly, for PHP developers, there needs to be an alternative to Trac for source code browsing. For one thing, if you're coding in PHP, it simply makes sense to have your project management and code review application also be written in PHP. This would hopefully lead to more opportunity for customization. It also makes it much more familiar to deploy than something like Redmine (Ruby) or Trac (Python) since you simply unzip the archive and go. I'm not sure where Ruby and Python developers ever got the idea that installing their stuff was easy -- there is a reason the "P" in "LAMP" usually stands for "PHP" after all.

Nonetheless, until the end product exists, some decision needs to be made about the tools that will be used to manage the project of building that end product. I've been cataloging some of this in this blog, but ideally these documents would better live on their own, in a way that other developers could add to or alter them.

It would also be useful to have hosted source code control and file hosting until the project itself is usable. Perhaps it would be best to proceed with a hosted infrastructure for the base bits, and then migrate everything to the in-progress system when it becomes at least marginally usable. In this way, we wouldn't migrate from software to software, but server to server, which would leave a cleaner result in the end.

Also, I think it would be useful to use the simplest solution possible. Rather than have a ton of features that would be difficult or impossible to transfer into the new system, it would just have some basic wiki and upload functionality alongside the source control. Hmm, this is sounding a lot like Google Code.

Google Code seems like a good barebones choice to start out. In terms of incubation, it's as good a place as any to have code to start as a launching point. Github looks like an interesting alternative, and I like some of their reviw features, but I dislike paying for features that I'm getting for free in an SCM tool (svn) that I'm already familiar with, namely what Github refers to as "Private Collaborators". When Stonepath is complete, the pay-per-user model will go away for anyone willing to host their own copy.

So look forward to Stonepath on Google Code sometime soon, to eventually be followed by its own project host.


0 Responses to The Paradox of Infrastructure

  1. There are currently no comments.

Leave a Reply


Development Headlines

What is Red Alt?

Red Alt is an outlet for Owen Winkler's web tools and provides a consolidated and organized archive of other online web development tools and resources.