I half-agree with your other respondent, olavk. If you're deeply entrenched in imperative-style coding and are just missing familiarity with - what's that other

I half-agree with your other respondent, olavk. If you're deeply entrenched in imperative-style coding and are just missing familiarity with - what's that other word? Prescriptive programming? - then by all means, you want to get down with it and get some practice. Set yourself some "toy" tasks and learn by doing until you're familiar with the concepts and effects. But meh, maybe after 6 years you've been way down that route. Something I've found helpful is to interact with a good HTML/CSS debugger such as the one in Chrome. You can highlight a chunk of your output page and see how the CSS comes together at various levels to create the final look of it. IIRC Chrome even lets you change those settings on the fly in the browser to experiment. On the other hand, you have my complete sympathy insofar as I find that model very cumbersome in a few places. It feels to me like something designed by a committee, and one very intent on reducing CPU load. There are things I can do with the table model, especially stuff like proportional columns or regions whose dimension is a constant +/- a product, that are like pulling teeth with the CSS layout model. There's a reason there are huge collections of "tricks" advice on how to do these things, which should IMO be obvious to a "programmer" type if they actually were. On the other hand, maybe you just need to let go. TeX especially is highly opinionated about how it slaps text on a page, and allegedly those are the opinions of experts in the field. So if your output is not looking like your mental model of the expected output, maybe your mistake is being so specific in your expectations? Both CSS and TeX allow you to micromanage your output more or less down to the pixel level, but very often I'm quite glad to let the layout engine do its job while I do the rest of mine.