It's not really actionable for most comics makers, because there isn't a standard toolchain to point to, other than "learn web technologies, mash them into shape, pray it doesn't break". Splitting up the comic into various assets for panel layouts, lettering, etc all comes at a cost in terms of software dependencies.

The advantage of sticking to a traditional "plain image" with scrolling or pages is that the information is consistent, easy to parse, and robust to technical changes. The work of reformatting can be delegated to a sufficiently smart client app. We have OCR, we can definitely build a panel recognizer.