A recurrent discussion is happening in the French Web community about comments and blogs and how to continue to publish on your own blogs and still aggregates the discussion into a thread in your own blog. So basically how the discussions can be decentralized and distributed with a notion of multiple copies.
Let's use #sudwebcmt for the hashtag. It will help avoid putting plenty of people in cc in the discussions.
Before putting forward any protocols or technical solutions, I want to be sure we are having a common understanding of the problem space and what we want to achieve at SudWeb 2013.
- Gestion des commentaires
- Discussions
- Commentaires et communautés
- Protection for WordPress Pingback Vulnerability
- Un commentaire de trop
- Gloss is a bookmarklet for reading online. Any sentence can be highlighted with a click.
You are welcome to edit this document. There are revision history, so we do not lose anything, we just evolve it.
There are issues to solve on what we want.
- Do we want a thread of the discussions (aka following the inter-links in between blog posts)?
- Do we want a copy of these articles locally in each original blog posts?
- If yes where the aggregation starts and stops?
- Do we want a system where a comment can be made outside of any blogging systems?
- Do we want to have this comment being distributed on multiple blogs and in this case where and how?
There is a clear distinction between push and pull in social interactions.
- Push systems (notifications, traditional comments, emails) are usually not working in an open distributed network, because they generate spam. The reason is that the flow of information in a push system is dependent on the power to generate that flow at a low cost. To avoid the spam, receivers of the pushed information have to put in place strategies a priori (login, authorization, captcha) or a posteriori (bayesan filtering, heuristic, manual removal).
- Pull systems (RSS, click on a link, etc) are putting the burden on the information receivers. Their main issue is that it takes time and requires regular polling (bandwidth waste). Another issue is that it makes difficult to discover related discussions. We do not know what is out there. The benefit is that it completely avoid the spam. You get only what you want to get, but you do not get everything you might want to get.
Maybe a place to explore is the notion of "community of trust". We might want to give up entirely on the completely open network for discussions and put in place a system based on network of trust. Some communities have explored this such as Advogato. We might want to explore ideas.
See also the proposal