About every six months I decide I want to build a RepRap 3D printer using parts printed on my own 3D printer, and as time goes on the landscape gets worse and worse. Several of the key companies selling kits are folding up. The RepRap wiki points to articles written 3-5 years ago and still in an unfinished state.
Basically the MakerBot I would use to print the RepRap, along with several other very good relatively inexpensive manufactured or semi-manufactured options, destroyed the market for oddball printers with smooth rod skeleton frames and exposed electronics. And I suppose that’s a good thing, but also (in a nostalgic-hipster way) kind of a pity.
Then I go to DIY CNC routers and do the ritualistic “wow, where would I put that, and my oh my that’s expensive, and I bet that wouldn’t even work that well” and fold it up until the next six month window.
I also have some cycle, wavelength/period unconfirmed at this point, of thinking about trashing my Efendi blog tool and going to a static site generator using Markdown files or something like that. But the funny thing is that I then start researching web frontends to a static site generator, since my lockdown computer here at work can’t SSH into anything and barks at me with soft blocks if I even dial up dropbox.com in a browser. So I basically need a static site generator fed by a dynamic site manager. Also known as a caching CMS, wakka wakka.
I’ve been thinking about writing/modding/finding a tool that Hoovers markdown files off Dropbox (soft blocks be damned) into a static site generator and onto my web server or GitHub pages or something like that. One of my more creative and ambitious options would be a post-by-email system, which would work great with a static site generator but would be a nightmare to figure out how to edit a post. Reply to the e-mail? That would be cute until it came time to implement it and handle rich text/MIME e-mail formats.
Efendi is devastatingly slow, even at the loads I run (which are basically limit-approaching-zero unless I get crawled by a Russian search engine). A lot of this is the cool feed merging stuff I built in to update from Google Code (defunct!), Picasa Web Albums (ruined by Google Plus!), and Cluster (near-idle!). I never bothered to optimize because (1) that’s not fun, and (2) I don’t need to optimize, but I do wish I could just deal in text files and serve flat template-driven HTML without a dynamic page build on every request.
Interestingly, the other direction I feel pulled with the blog is to get some kind of Twitter connectivity set up so I could do microblog entries on Twitter and cross-publish to the site. This is difficult to do, of course, and that ick feeling that all programmers (and probably most craftsman/artisans) know starts to set in, where the feeling goes from “let’s build this!” to “let’s avoid three months of untangling API hell and unintended consequences and forget about building this!” Plus I really don’t feel like Twitter works well for me to record anything meaningful, so it’s pretty much a red herring in the whole thing.