> (I'll pass on the question of how important it is to replace the most powerful context aware autoescaping HTML string templating with something else.) It's p

> (I'll pass on the question of how important it is to replace the most powerful context aware autoescaping HTML string templating with something else.) It's probably me, but I'm not grokking this ending sentence. Is this praising go's "html/template" or is it meant as satire? I believe it isn't but genuinely can't say for sure. From a cursory look, html/template looks like doing the right thing, but come on, you have to tell it explicitly and procedurally which kind of escaping(s) you want for every individual var expansion. Compared to SGML which automatically detects where expansion of an entity reference occurs (within attributes, within CDATA or RCDATA, within normal parsed character data, and so on) while also considering the entity type and type-checking the final expanded result against the expected content model (not just for preventing <script> injection) this is a joke. Maybe SGML templating is what the author considers 'strict structural containment'? I mean, SGML isn't just any rando templating system, it's what HTML is based on, has been with us since 1986, and should be the point of reference.