Frankly I think this posting of github pages instead of the authorative source is inappropriate. I see it a lot on HN and find myself having to check that this is the official repo by going to the original website. If you know it you don't need to be told about it but if it's being posted to get it to people's attention then the canonical source is the right link even if it only has a link to github.
> Frankly I think this posting of github pages instead of the authorative source is inappropriate
The website is the authorative source of... what, exactly? Not the codebase, not the documentation - neither of those things are on there. The only piece of content is a stealth link - and I'm not alone in having missed it, and only found out about it via the comments. "Did it get HNed...?" I wonder if the link would've been more obvious on desktop, where I might've hovered over the icon with my cursor.
I see linking github pages as more along the lines of e.g. deep linking specific developer.apple.com pages. This has always been how the web was designed to work - link directly to the useful information. If you want the authoritative source readily available, link back to your authoritative website via your readme.md, so everything's nice and cross-referenced (as I note they have.)
> I see it a lot on HN and find myself having to check that this is the official repo by going to the original website.
Ahh - just because you're on an official-looking website doesn't mean you're on the official website. You could easily be looking at e.g. a community fork of a project, possibly renamed, instead. Linking to a website doesn't actually solve this problem. That github links encourage you to look at the wider ecosystem might even be a feature... ;)
I can see the rationale behind linking to the original website, even if the UI isn't immediately apparent. However, the linked page is served via HTTP, so, depending on your threat vector, there isn't an authoritative source.
