I think I’m changing some of the baseline big refresh thoughts I’ve had. While I’ve previously been all in on node.js I’m starting to lean towards python. I’m not sure why I’m thinking that way, so it’s still in the air. Lately I’ve been doing a lot of looking at both options, for a home-rolled blog that is, and something just keeps on making me lean away. Not that I’ve actually done any work of course. It might be that I’ve always disliked anything that goes full monolithic. This isn’t to say that JavaScript is any more prevalent than python. I think it’s more that I don’t want to have anything be that prevalent in my own world. The majority of code I do at work is still JavaScript, but I’ll be branching into a bit more python in the next few weeks, and some Perl as well.
I’m kinda dreading the Perl as the syntax doesn’t currently work for me, and the system I need to do it for is a production server that has custom scripts written ten years ago, the comments are the documentation, and any upgrade takes six out of state engineers to do because of those scripts. Once I dig into it though, I assume it’ll be awesome. Mainly, those scripts give us more headaches than a box of worms for what like they seem to do at times, and I’d really like to clean that whole deal up.
Python would be for new and maintained PageDNA templates work wise. While I won’t be the one doing the heavy lifting on this, I’ll be training the person who will be. The comedy value, and my skill, is I’ll be lessoning what I train. The way they’re using python is fairly clean, though again, there are a lot of the older templates that have some pretty janky anachronisms. It’s the curse of having vendors write custom stuff over the years in that things that work don’t get updated, for good reason, but that lack of refreshes makes future updates more difficult. People bandy about the word refactoring, and that might be what I’m talking about, but going back in and cleaning or streamlining old work I think is an invaluable process.
I’ve instituted a change log process at work for PReS Connect, because reasons. As I get older my systems plasticity seems to be sliding away. Rather than having everything in my head, I want notes where they’re easily reviewed. Code has to be commented because I never liked the whole “good code is self-explanatory” trope. If you don’t need to document the why and what, did you need to do it to begin with? That sentence made sense when I wrote it a month ago, seemed gibberish five days ago, and read naturally five seconds ago. It’s a bad sentence.
I’ve been reading my book on Perl for the last few days. So far everything makes sense. We’re talking to the point that I’m tempted to look at throwing Perl on a few other systems for ease of log processing. I had come up with a surer ghetto process out of PReS Connect to get decent error emails, but it’s only good about half the time. This of course is a current feature request in PReS Connect that’s totally in the works. Seems stupid it wasn’t built in from the start, but whatever. I’m not sure how much of a prepress tech I am anymore. I mostly do coding and documenting. It paid decent though.
So, Python? The Stack Overflow annual survey came out, and it seemed like I should think Python. The only reason I keep doing second looks at node.js, is it means I can keep the front and backed all in one system. Spinning up a server is a dozen lines, and… I have to use JavaScript all the time anyway? Python send a little less janky, I have to keep everything separate, and it’s not a language I know as much of. At the end of the day though, I don’t like pre-built modules. I want to see the secret sauce. I want to see the nuts and bolts. There only real bump is I dislike the bigness that is Apache. It’s huge. It has features that only add attack vectors for me. I can mitigate this though. If I switch over to a custom server, I can turn off a lot of that stuff. I don’t need PHP. I don’t, need?, a database. I don’t need what they’re selling. I could look for a smaller box to serve my code, or I could just block all the box parts I don’t want. At the end of the day, Apache mostly just had to be kept updated, with the updates being verified. With that as the server, I can then focus on the language. I could mix and match, or go only Python. It’s up to me.
I think this post has rambled long enough over paragraphs and days.