So I just tried this with a bunch of medium-complex documents and it's wildly wrong. I suspect the authors have never seen an actually complicated Word document?
My examples are all internal/confidential, but if someone from the project wants an example I could probably do some search/replace redaction. It would be a lot of work though because there's photographs and such too, and indexes, and tables, and documents inserted by reference, cross references, conditional fields, bibliography fields, formula fields, etc etc.
I don't know why this was flagged, but you are right.
Google Docs [1] and OnlyOffice [2] also employ the canvas method to render office documents, and have found it reliable and consistent among different browsers.
