Hashtag Gee Tee Dee
As I’ve detailed many times before: I want to move off of Tumblr. I have a chronic fixation upon the act of starting to move. I have half-written, hand-crafted, artisanal blogging scripts that do work —more or less— but I haven’t followed through on a single thing. I’ve dabbled with Jekyll, Octopress, Pelican, WordPress, etc. I just don’t like any of them all that much. I am pretty convinced, at this point, that rolling my own will ultimately be where I wind up just so I can have every little thing exactly as I want it. I’ll just need to overcome the nigh insurmountable hurdle of my lack of education in software.
This is why I usually jump at the chance to write small scripts. Things that can move just enough data —connnnnnnntent— to make me feel like I have accomplished something, but not so large of a project that I pull the zip-cord on my procrastination parachute.
Joe Steel: Poorly Skilled Software Developer
I’ve mentioned, jokingly, that I have a BFA in Computer Animation — that’s really what the degree is in — but that’s not always a “funny-ha-ha” joke. It’s a way of explaining that I am ill-suited to software development of any scale. This is why I like to take on small tasks to actually give myself just enough to do to learn something.
These are the reasons why I’ve been experimenting with Casey Liss’ Node.js app, Camel. It’s a pretty succinct app (as long as you don’t research all the dependencies’ dependencies’ dependencies’) so it’s actually going to serve as something useful for me in an educational way, if nothing else.
For now, I was content to make a content converter script. It’s a little thing, and I knocked it out in a few hours. (Shut up, PROS, I know this is like 10 minutes of code for you.)
I had tried Pelican recently, because of its Tumblr import utility. Unfortunately, I found the utility to be half-baked. It parses through your perfectly adequate HTML from Tumblr and tries to turn it in to ReStructured Text, or Markdown. I modified the script so it would stop doing that and I captured all of my posts. Unfortunately, Pelican is a little rough. It made a total mess of the index pages when it tried to process stuff, and even the things it got right had the weird appearance of looking like I had posted them instead of making it more obvious that they are Tumblr reblogs.
With all this data, I could just pipe it in to other blogging platforms, right? No. It seems every blogging platform has their own, slightly unique system for storing files and file metadata. Camel’s metadata is one of the stranger ones. I made a conversion script that copies the files and sets up the expected folder hierarchy, and converts the metadata.
This isn’t the kind of thing I see anyone using, but I wanted to see what putting it up on GitHub would be like. I’ve used gists on GitHub, but I’ve never needed a project. I had a private repository with a friend on Bitbucket, for one of the blogging component experiments, but Bitbucket is kind of weird.
I even did all the command-line git stuff to put it up. That’s quite an accomplishment for me, because I come from a background where version control is just incrementing an integer in a filename.
Anyway, at this rate, I should have 10,000,000 pointless scripts and no blog.