i used to write PDP11 assembler too. it was one of my first assembly languages and the first real mainframe i had access to as a kid. i think i remember exactly zero. only that i am not a fan of punched cards. 
the Styled Text processor is pretty crude (at least from my perspective) i think it suffers from three primary issues:
-
there are many aspects of Rich Text Format (the input to the system) that have no obvious translation into HTML.
This means the translation will always be a “best effort” algorithm. It necessarily has some cases where it knows it will make a mistake because there is no “correct” translation. -
It has, for 15 years, allowed HTML to be inserted inline with styled text. This is a useful feature that many use — but it adds even more opportunity for the user to provide input that has no valid output.
-
It has been largely unchanged for 15 years. This means there are 10s of thousands of documents (more???) in the wild that depend on the quirks, both good and bad ones, remaining unchanged.
Fear of breaking lots of existing projects surely stymies any substantial behavioral changes.
i guess what i’m getting at is that with those as hard limits there will always be plenty of times when the expected output of the engine will be INvalid HTML. Garbage-in, Garbage-out was the adage from the old computing days righ?! LOL
I think the trick to affecting change is providing a bug report that shows that your input is NOT garbage. If the input is good and the output is bad — then that is surely a bug that must be addressed and cannot be brushed aside as merely legacy behavior. Surely the output for valid input takes precedence over any legacy quirks.
That said, for all of the above reasons (and a few more) I’ve decided that Stacks will be moving to it’s own text system soon, hopefully in Stacks v4.5 in about 12-18 months.
We’ll keep backward compatibility with RW’s styled text system but newly added stacks after the newer system is in place will have more rigorous bounds on both input and output to ensure always valid behavior that’s more consistent with Mardown.
isaiah