There are two different needs that you seem to be conflating here:
The first "need" is to comment and participate in online discussion: users have an opinion or question that they'd like to engage a larger audience or community with. This appears to be the need that would drive the hypothetical annotation services you're talking about.
The second need is to keep track of information that might be personally valuable to me in the future -- this is what things like Evernote were built to do. This need actually doesn't really relate to the first need well (despite similarities in UI or execution of the product), because for this use case, many people would prefer to keep that information private.
So, if we focus on that first need, we arrive at exactly the conclusion Dan mentioned: there are already lots of ways to participate on the web already, comments, etc. but even Twitter and Tumblr provide ways for me to provide feedback (sometimes on specific pieces of webpages).
And now you're saying, "But Ninakix, annotations would fill [a particular need (e.g., people's need to provide detailed feedback or analysis of text, whatever)] better than comments do! They should still work!"
So, the specific issue with this logic is that when it comes to approaching needs that are already being filled by some other product, one needs to design a solution that's orders of magnitude better than what currently exists. If I can hack my existing tools and existing networks to get done what I want, as people do when they use ">>" type indentations to indicate quoted text, my willingness to switch to a new product is very, very low.
If you want to think about how you might build an annotation service, here's how I'd evaluate ideas:
- What would this tool allow me to do that simply wasn't possible before, even by hacking?
- What need does this "super power" fill?
- Would this new super power be something that I'd be using 90% of the time I have this need?
Obviously the relatively random effects of path-dependency mean that anything can take off, but, the better you can fit those criteria, the more you increase your chance of success.