RE: Canonical content and AMP

My initial reaction is that AMP, while good intentioned and helpful for performance, is similar to the “mobile web” XHTML subset or the TV subsetting that we discussed. The difference is that AMP is opt-in by the developer community, so it’s not device- or browser-makers building it in…

From: Andrew Betts [mailto:andrew.betts@gmail.com]
Sent: Tuesday, April 19, 2016 1:31 AM
To: www-tag@w3.org
Subject: Canonical content and AMP

At the F2F in March I was concerned that the best progressive web apps were generally mobile-only, and I’m getting a similar feeling about AMP.

I’m wondering whether TAG have a view on the use of `m.` sites.  From an architectural perspective `m.` is a bad thing, surely.  But I can’t build my desktop webpages using AMP - because I would lose essential features like reader comments, so I'm stuck serving (at least) two copies of the same document.

HTML sandbox / Content Performance Policy seems like a better approach, but I’m wondering whether I’m alone in thinking that a) technologies like AMP can promote duplication of content in multiple formats in different locations on the web, and 2) that’s a bad thing.

Seems to me that we've spent the last few years gradually moving away from m., and we're now heading back towards that territory.  Is this a conscious recognition that One-Responsive-To-Rule-Them-All was a bad idea?