index.html: part of the document, or just for metadata

[frivoal]

Should index.html be the first entry in the document, rendered normally in addition to having lots of magic metadata, or should an e0 reader pull out all the metadata, figure out the first item in the spine other than index.html, and render that instead?

My general logic for e0 is "What would a regular browser do? Let's do that, but smarter." Based on that, I think an e0 reader should display index.html normally.

This makes it possible for a regular html file to be a valid e0 document all by itself, and combined with externalizing metadata #5 enables authors to create arbitrary experiences, structuring the document anyway they like: index.html as the cover, index.html as the whole document, index.html as the front-matter...

[therealglazou]

The thing is, I don't want the spec to say what a RS should do. The Navigation Document is a html document to be browsable, with visible contents, for that purpose. Some RS will render index.html, some will parse it and use the contents to generate some UI.

[frivoal]

I fully expect (smart enough) UAs to parse the content and generate some UI for it. The question I am asking is whether that is instead or in addition to showing the index.html directly.

Of course, in the end UAs can do whatever they want to, and specialized UAs that don't show anything to anyone and only extract metadata will exist as well.

But we should be clear to UAs whose primary purpose is reading that index.html is meant to be shown to the reader, that there may be unique content in there that would otherwise be missed, and that it is not just a verbose data-dump.

If we don't and some UAs just skip it because they assume the first bit of user-readable content is the next item in the spine, then authors won't be able to put anything there other than metadata lest readers don't see it, and then other UAs would need to skip it because it is a generally ugly and boring part.