I'm working on a very small, no frills site management soft at the moment that includes a simple multi-author news publishing system. It exports a primary rss feed as well as author-specific feeds that syndicate the last N news items and the last N items for each author. These are static files that are republished whenever news items are added, edit, or deleted.Date: Fri, 15 Mar 2002 08:20:33 +0000 From: Julian Bond <julian_bond@voidstar.com> Subject: Being kind to clients As RSS consumers we have a duty to be kind to RSS providers by not collecting the RSS too often. Can I make a request to providers to be kind to clients by limiting the size and age of items that are sent to clients? I store 4 weeks of news from each of the providers I read. There are a few blogs that include items in their RSS that are older than that. So every 4 weeks I see the same old items come round again as new. It would seem to be reasonable to limit the age of items to about a week. Anyone who is reading the RSS reasonably regularly will then be sure of picking up the items. There are some RSS providers that only include items from the same (or previous) day. This is a bit too aggressive. If I miss a collection for24 hours, I can miss the items completely.
http://thundercat.monokromatik.com/thundercat_files/syndication/(ugly until I tweak the stylesheets a bit more and all of the "news" is junk that I've added to the db for testing purposes)
In the case of static files, can't RSS clients use the date value in the Last-Modified header to determine the freshness before grabbing the feed? Most, but perhaps not all, servers will provide Last-Modified headers for static files. If the RSS is produced on-the-fly by cms/site management tools, they can be written so as to supply Last-Modified headers (plus Cache-Control and Expires headers, which might go towards addressing the polling frequency issues raised by Bill Kearney). Many/most browsers seem to make use of this information, when it's available, to good effect.
That being said, many/most systems that I've seen that export RSS seem to be doing on-the-fly generation of RSS in response to requests and not supplying any useful header information. Getting them to go the static files route or at least whip their scripts into shape to produce these headers would probably be a hard sell.
-brian donovan