I just finished a significant upgrade to baus.net. I regrettably let the site atrophy to the point where I wasn't sure how it is was running. Baus.net felt like a messy desk. While I could get other work done, it kept nagging me every time I looked at it. While I haven't made any aesthetic changes yet, I did complete the following tasks:
- Download pyblosxom blog code to my Mac and get running in WingIDE.
- Get Subversion plugin that hosts my entries running on Mac.
- Make some code updates to take advantage of the Python Subversion API changes.
- Get pyblosxom core running under WSGI on twisted.
- Research re-enabling comments (decided to make it another project).
- Move static content to S3.
- Move image hosting to Flickr.
- Centralize code and content into a singly rooted Subversion repo.
- Upgrade from statcounter to Google analytics.
- Create CentOS VM on VMWare Fusion to stage deployment.
- Configure nginx to proxy to twisted and feedburner for my feeds.
- Upgrade memcached and libevent.
- Centralize logging for all services.
- Document the installation process.
- Write startup script.
- Provision new Linux VM from Linode.
- Perform installation on Linode.
- Update DNS to point to new installation.
This project ending up costing tens of hours of my personal time, and there was a point that I was ready to scrap the whole thing and move to WordPress.
But I'm glad I finished the project. I use baus.net to try out technologies in a pseudo production environment, and having my own personal content in total disarray didn't sit well with me. I also realized that my previous experiments with using Subversion to store my blog content had a direct impact on a project we are working on which uses Subversion as the back-end of a content management system. I had confidence that it would work after running baus.net this way for years.
I take pride that in my free time I've created a system from end to end including system administration, Python development, and some basic CSS/HTML hacking, which includes a pretty novel use of Subversion. That might sound silly for something as small as baus.net, but I think there is something to be said about building an entire system no matter how small. As a project manager responsible not just administration or software, but whole systems, at some point I have to walk the walk to maintain credibility. If all I do is go from one meeting to the next projecting ROI, discussing synergies, and saying absolutely have I added any real value?