I’m sure that long time readers noted that this blog’s been rather slow for the last few months, slower than the time before my six month hiatus. Well, the good news, or possibly bad news depending on your opinion of my writing, is that Weird Things is not going on another hiatus even though once again the schedule gods refuse to have pity on this blogger. This crunch is partly of my own design because a fair chunk of my blogging time now goes to polishing my Hivemind prototype for open-sourcing. This means that I have to rewrite its code to better comply with certain architectural standards, polish the utilities I designed to make it easier to modify for future developers since I’m going to release what amounts to an extensible starter kit, and test each approach for performance and maintainability. Even for this starter kit there will be around 250 unit tests to ensure that a modification to any of the translators, or repositories, or container factories won’t break critical functionality, so as you can probably see, there’s a good deal of work to do before it’s actually ready for prime time.
Why spend the time to do this on top of my day job and some other important life events? Partly because it’s a project on which I spent years of study and research, the vast majority of which was used to come up with the working model of Hivemind’s architecture and learn the different technologies which would be involved, trying them out in individual projects, and assembling these individual components to further refine the design. If an existing toolkit within a framework I used didn’t suffice, I wrote my own, while in the meantime making sure it was consistent with not only the end goal but the relevant science. One of the obvious benefits of experiments like this was a better insight how the tools I use in my day job work and how they can be modified to do what I want them to do. The other is that they also highlight problems and raise questions I didn’t originally consider but really need to answer. Now, after all this time and effort, I want to see it in action and make sure it works in its intended context. One can spend inordinate amounts of time building theoretical AI tools and pontificate on their applications with no end in sight, or just build them and see how well they’ll perform in the real world.
On an unrelated, but important note, I’ve also been thinking about redesigning the blog itself since it seems to have been locked into a format that doesn’t really allow me to be as flexible with posts as I like. Whether I’ll try to build my own blogging platform from scratch and add all sorts of convenient little features to it or just spring for another WordPress theme to customize, I’m not sure but if you have any ideas on the subject, please leave a comment or drop me a line with your thoughts. On the one hand, building my own platform will take plenty of time but it will let me customize everything exactly as I see fit so once it’s up and running, it will give me a lot of flexibility for future posts and for layout options. On the other, using a customized template will save both time and effort, but it won’t necessarily give me the same range of options unless I happen to find a template that I know can be extensively modified without a lot of testing or compromising its functionality. Regardless, I’ll post an update when I come up with a decision and again, I encourage you to weigh in and let me know if you have any particular thoughts on the subject. Since the choice hasn’t been made yet, I’m on the lookout for ideas and options I haven’t yet considered before setting my foot down and getting to work.
[ illustration by Marian Cerman ]