That's absolutely fair, SamSam – and thanks for taking the time to write (glad I found your comments, they may come in useful down the line!).
There have been several attempts in the last decade or so to catalogue science demos, some large efforts and some small. All have failed, in some cases taking quite a lot of money with them. Few reached content scale; none really solved the metadata problem; none reached critical mass for audience or contributors.
Most recently the UK has built a huge library of all the STEM education resources they could lay their hands on, and while the result is impressive, it's not yet solved the 'people actually want to use this' problem. I'm not convinced it will, which worries me because some of my own work is trapped within it.
ScienceDemo is itself an experiment to see what a minimum-effort site might look like, and to find out whether that solves any of the problems. So, yes, it's a bare-bones WordPress site, running the (new) default theme, with all the positives and negatives that entails. You're absolutely right that as the site grows the taxonomy will go from 'barely adequate' to 'utterly hopeless,' but that's only a problem if people actually use the site.
That's a problem I want to have. And if people do find it useful – particularly, but not exclusively, teachers – then I have to hope that I'll find a way to pay for some proper design and data model work.
There's a horrible problem in science teaching around the use of practical work. So far, ScienceDemo.org is but a modest contribution. It could end up being a huge asset. We'll see.