I'm in full agreement with you about HTML parsing being a bit undeveloped at the moment.
If you have any ideas about how the product team could close this gap - drop them under the "Ideas" section - they do read these, and if the idea is well supported by the community it has a good chance of making it into the product.
I find parsing XML really tedious to parse. It takes me ages to get the right tag name. Eventually found two solutions that are both different to the more linear official solution give. My solution attached.