You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The idea here is to provide a way of grouping article from different feeds or from the same feed on a criteria.
The base implementation criteria will be "article have the same link". It's the simplest form of criteria I could think of while being quite useful since duplicate articles from different feeds are a recurring things in my day to day feeds reading.
Cluster will be created automatically on the server side. It should happen at the article creation.
Each article will have its cluster. Several informations from the article will be de-normalized on the cluster such as title, feed.icon, date.
The first article of the cluster is to be that cluster main article. It shall be displayed first. If another article is added to the cluster, it shall remain secondary unless the feed it comes from is marked as a primary information source in which case articles from it will automatically become main articles in their cluster.
If two feed are primary then appearance order will prevail.
If an article is attached to an existing cluster, informations from that article will be considered as second hand and dismissed unless the feed it comes from is marked as primary source. In that case only, the cluster will be updated.
The cluster table will be referenced by each article which belongs to it, and, has category will only have a user_id as foreign key.
Depending on usage, various de-normalization may be possible :
main article feed icon
main article title
main article date
UX and UI handling
The UI has three column : menu, middle panel with the article list and right panel with article content.
Menu
Not much will change here on the outside. Filtering will be altered though.
Selecting category / feed should filter cluster displaying only those pointing towards articles that belong to those category / feed. That should be through relation table (many to many).
Middle Panel
Middle panel will from that point only display clusters. It should not be any different from before since icon, feed title and article title will be copied from the main article.
The Right Panel
Only the article view would have to change.
If the cluster only has one article, the view should remain the same.
If the cluster has more than one article, each article should be available through a tab inside the right panel.
It is important to think of large cluster containing more articles that what can be displayed in the width of the page. Decision will be made later in development about either horizontal scrolling, a drop down list, or a better options than the two precedent one.
The text was updated successfully, but these errors were encountered:
Cluster Feature
The idea here is to provide a way of grouping article from different feeds or from the same feed on a criteria.
The base implementation criteria will be "article have the same link". It's the simplest form of criteria I could think of while being quite useful since duplicate articles from different feeds are a recurring things in my day to day feeds reading.
Cluster will be created automatically on the server side. It should happen at the article creation.
Each article will have its cluster. Several informations from the article will be de-normalized on the cluster such as
title
,feed.icon
,date
.The first article of the cluster is to be that cluster main article. It shall be displayed first. If another article is added to the cluster, it shall remain secondary unless the feed it comes from is marked as a primary information source in which case articles from it will automatically become main articles in their cluster.
If two feed are primary then appearance order will prevail.
If an article is attached to an existing cluster, informations from that article will be considered as second hand and dismissed unless the feed it comes from is marked as primary source. In that case only, the cluster will be updated.
Current database schema
Proposed database schema
The cluster table will be referenced by each article which belongs to it, and, has category will only have a
user_id
as foreign key.Depending on usage, various de-normalization may be possible :
UX and UI handling
The UI has three column : menu, middle panel with the article list and right panel with article content.
Menu
Not much will change here on the outside. Filtering will be altered though.
Selecting category / feed should filter cluster displaying only those pointing towards articles that belong to those category / feed. That should be through relation table (many to many).
Middle Panel
Middle panel will from that point only display clusters. It should not be any different from before since icon, feed title and article title will be copied from the main article.
The Right Panel
Only the article view would have to change.
If the cluster only has one article, the view should remain the same.
If the cluster has more than one article, each article should be available through a tab inside the right panel.
It is important to think of large cluster containing more articles that what can be displayed in the width of the page. Decision will be made later in development about either horizontal scrolling, a drop down list, or a better options than the two precedent one.
The text was updated successfully, but these errors were encountered: