[feature] suggested links collection

[jibe-b]

The daily life of a 21st century individual is flooded with content one accesses but also with suggested content.

WebMemex could store suggested links (for example links appearing in a social media feed) and analyse them, keeping track of the fact that they might not be in the momery of the user.

If I am efficient enough, I may implement this feature in my fork. Don't hesitate to ping me to make me hurry up a bit :wink:

[blackforestboi]

Hey @jibe-b,

Thanks for dropping your idea here. A small concern: I could imagine storing the content of all pages a person experiences in their feed might cause a lot of unused pages to be indexed (and thus in the mid-longterm cause performance issues. Also it could cause some copyright issues to parse that data out. We might though parse out the urls of comments of articles you read on Facebook and then show it to you in the sidebar later as related content?

Also related to that, we actually intend to write a parser specifically for facebook that allows to keep track of all the pages, comments etc you like/share, so that this content can be found again faster, because it is ranked up. What do you think about that?

Enjoy your week!

[jibe-b]

Thanks for your answer, @oliversauter :)

A small concern

true, this feature would open the door for many concerns ;)

a parser specifically for facebook that allows to keep track of all the pages, comments etc you like/share, so that this content can be found again faster, because it is ranked up

I approve. I guess it is something one may expect.

Keep me in the loop for this, I may help.

storing the content of all pages a person experiences

At least, it could be an opt-in feature, which get deactivated automatically when it consumes too much memory. Or that only keeps the index of the last X days.

Still need to design this feature better, I'll try to propose a better version in the next days ;)

[jibe-b]

Another option, in order for someone to have what he doesn't read indexed would be to allow a user to:

• bookmark a link (right click, bookmark link on Firefox)
• have webmemex resolving the link if necessary (if bit.ly or l.facebook.com etc)
• index the content of the bookmarked page

[poltak]

@jibe-b if I'm understanding your idea right, this should be a feature already planned for the new version of WorldBrain :) on bookmark event, it will perform an XHR in the background to get the DOM pointed at via the bookmarked URL and do most of the same stuff on that which happens when you manually visit a page.

[jibe-b]

That's great!

That may need to take care to handle redirections well, so that when using bit.ly, tfe good URL is kept in the index.

[jibe-b]

This may be an example on how to deal with bookmarks with the Webextension API

https://github.com/rwanyoike/bookmark-refresh