User:Johan:Rss
From Carls wiki
Below are summarised the various features of Johan's RSS reader.
Contents |
Wishlist
- Tagging
- ...of posts.
- ...of feeds.
- Mark posts for further review
- Trackbacks
- Blog this (It should be: 'put in blog queue' instead of 'blog now' (can be complemented with put in mailqueue) (this should be implemented as specialtags (like the inbox in gmail (this is getting lispy))))
- Subscribe to post comments.
- Webservice integration:
- del.icio.us</dt>
- Aggregate tagged pages (only from del.icio.us to the reader)</dd>
- Flickr</dt>
- Basically just sync the feeds (I dont know if the tags are in the feed url).</dd>
- Email (gmail, want those tags)
- Views
- View posts chronologically (rss is not email)
- View posts by feed.
- Tags are views (as in gmail).
- Search
- Annotate posts and feeds (wiki-markup).
- Some easy interpost linking mechanism
- iCal
- Implemented in Ruby on Rails (RoR).
- Open Source
Roadmap
- Version 0.1
- A basic web interface to read rss-feeds. No annotating, no comments, no syndication.
- Version 0.2
- Add syndication, polish interface.
- Version 0.3
- Don't know yet.
Issues
- Comments, Resolved
-
There's comment feeds and the software user's comments. Maybe the user's comments should be called annotations instead.
- Parsing RSS and Atom in a uniform way
- This will get tricky, very tricky. The easiest way to do this is to use Mark Pilgrim's Universal Feed Parser. It's written in Python though, so to use that I need some glue.
- Subclassing?
- There's lots of information that can be handled by subclassing, a
commentfor apostis a variant ofpost, atrackbackis also a variant ofpost. What's the superclass here, is it post or something else?
Relevant information from the web
- RoR documentation
- Feed info
- Mark Pilgrim's Universal Feed Parser
- Misc (blog posts and stuff)
Architecture
Database schema
Tables
Names are according to the ActiveRecord standard
| posts | feeds | tags | annotations | trackbacks | posts_tags | feeds_tags | tags_tags |
|---|---|---|---|---|---|---|---|
| post_id | feed_id | tag_id | annotation_id | trackback_id | post_id | feed_id | tag_id |
| title | name | name | post_id | post_id | tag_id | tag_id | tag_id |
| url | url | ||||||
| author | description | ||||||
| summary | etag | ||||||
| content | modified | ||||||
| issued | |||||||
| modified |
I don't know if the tag-tag-coupling is necessary in this case (it's useful when tagging photos). If this relation is implemented it should be a DAG (cycles define synonyms, and that's just different names for the same tag, which could be implemented (and if it's implemented it should be as different names for the same tag, not by allowing cycles (that's just lazy and the model becomes rather weird and hard to grasp)), but I don't think it's a good feature).
Praise
Collection of favourable quotes for the software.
"Wonderful! Is it free? I'll take two!" — Carl 06:36, 25 May 2005 (CEST)
