This document is valid as of April 3, 2023.  If you want the most recent version, just ask, and I will update it.

There is a difference between the toots that a mastodon user posts, and what uncensored news imports.
I ignore boosts.  That would violate the other person's privacy. 

I ignore toots with visibility set to private or direct.

Some curators post off topic.  This can be recognized with a few keywords.  So each curator gets a list of keywords to ignore, to cause the toot to be ignored.

I ignore toots without links.  I am not interested in what the curator says, I am interested in the articles he recommends.

I ignore toots in languages other than English.  Eventually we need a version of this site in all of the major languages.  French is a likely next language. 

I ignore links to twitter, no point giving them traffic.  That just helps them.

There are 3 types of links in toots.  Mentions, Tags, and article links.  I ignore article links to mastodon posts.  I like nicely formatted web sites with articles.   I consider them more valuable.

I ignore links to the curator's own web site.  Better to import the website using RSS.  RSS articles are imported as listed, not recommended.  And it would be inappropriate for a curator to recommend his own articles.  If your site does not support RSS, then I can allow links to your site, but they will not be treated as recommendations.   

If the hashtag is #yellow or #red, the article has a yellow or red background.  This is for flagging misinformation, disinformation, or propaganda that a curator has tooted about. 

There are three types of article data structures.  RSS Articles are imported from about 163 trusted RSS feeds.  Links are articles which I have added semi-automatically.  I enter the url, it crawls the site, and extracts the title, description, and image.  Toot articles are links which the curators have recommended, and which are also automatically crawled. 

I treat toots, and the articles that they mention separately.  Multiple toots or tooters can mention the same article.  One toot can mention multiple articles.  When that happens, I do not know which part of the text refers to which article, so the toot text is not shown, just the downloaded article preview.

Processed toots are connected to articles.   So every toot links to the articles it mentions, and every article links to the toot which mentioned it  Displaying an article, uses the text from the first toot that mentioned it, and gives credit to all of the curators who mentioned it.

When the curator suggests a new article, the article is added to /mastodon-articles.   For people focused on a topic, I can add their articles to a different category say /healthy-food, or /ban-cars or /stay-grounded.  If you are kind enough to use the current hashtags, I can (soon) post the article to the correct category.  Here is the taxonomy of topics, hover over a topic to see the correct hashtag to use.

Mention hyperlinks are extracted from the toot, just the person's id is left in, it is not hyperlinked.

Hashtags are extracted from the toot.  If there is a single hashtag by itself in the text of the document, it is left in place, but the hyper link is removed.

And of course article links are extracted from the toot.

Each such article link is used to download the article title, description, and possible image.

Number of boosts and number of likes are recorded, they are also updated when the toot history is recrawled.

If the article fails to download, the toot is discarded.

Articles are shown, most recent first.  In every category, the relevant articles are shown, most recent first.  if an article was downloaded from an RSS feed, its publication time is when the article was published.  It if is downloaded from a toot, its publication time is when the toot was tooted.

It is a complicated algorithm.  And it is continuing to evolve, albeit more slowly.

If you login, I can give you access to a private page that shows your most recent 30 toots, before and after i process them, along with links to the articles. 
You can search your recommended articles using your user name @Snoro, or @takvera for example.

Future Plans

I do not yet automatically sort tooted links into categories using a /#slashtag.  But then again no one is yet providing slash tags.  

It is possible to do toot curation using a reply with a /#slashtag.  That would place the tooted link in the correct category based on the slashtag.

I want to be able to import Youtube RSS feeds and conver the links to make them into proper embedded youtube videos.  Right now I just link to them.

 Built using the  Forest Map Wiki