I must admit that your approach is way clearer and understandable than mine.
Once done with current problem i'll try to read plain HTML using given method :)
One thing i don't get is statement "avoid parsers": why so?
I was convinced that parsers were ment to extract data from more complext structures

On the other hand i've managed to access chans using JSON interface (following https://github.com/vichan-devel/vichan-API/), but my question remains valid for HTML-only imageboards