In part of my attempt now to get Koype showing more useful contexts when replying to content, I hit a wall when it came to extracting authorship information. There's some cues when people use the word "author" in their URLs pointing to authors or even using common standards for identifying authors like "p-author" or "rel=author" in their markup. It creates no visible friction for the user and it helps other forms of software more semantically show information to people if they're viewing your content from different sources. My gripe when working with this in Koype is that I didn't want to code in the 'way to extract' content for hundreds of sites. Although, I don't have to - the folks at TinyTinyRSS already ran into this with and have a solution via SiteConfig. The thing is, it's not an actual solution if publishers just post anything without immediate concern of how consumers will take it.
That made me realize that that's the whole point - to make it harder to use custom tools to "grok" the contents of the page and drive direct traffic to their sites from a Web browser. That's trash, in my opinion. I'd want to be able to read things in the medium that I find efficient. If it's my desktop reading it out loud to me, my phone showing me a snippet or me kicking back and swiping through news titles on my TV (still in the works) then so be it! I'd be more likely to subscribe if I had autonomy over how I could parse my content.
The most difficult bit for me has been extracting not only the name of the author but a usable photo and a link to the author's profile (be it one of the publication's production or their personal one). That'd allow me to have more rich displays of content and even do a more semantic referral system (end goal!) of other things to read. Once I can get that, viewing more personal content can occur versus the very monotonic presentation we have so far on the Web.
