I’m not clear on the difference between atom and RSS. Atom seemed to be the better spec, but for my Astro blog I ended up sticking to the built in `rss` helper it ships with.
As a digital pedant I am very sympathetic to what prompted the creation of Atom. RSS2 for example under-specifies item "description" and "title," in particular how to put HTML in there, and using the most once-most-common technique (entity escaping HTML) makes it tricky to reliably do more basic things (encode/decode left angle brackets and ampersands, because now you don't know whether to do so singly or doubly).
But the undeniable victory of RSS shows the importance of being first and "easy" (even when "easy" means sweeping edge case problems under the rug). And of humans: Major publishers like the New York Times had adopted RSS and saw no need to switch to Atom because it was good enough. I'd argue the (also underspecified) CSV format is another example of this phenomenon.
(As for the entity escaping dilemma, people mostly just moved to using CDATA for their feed-embeded HTML, although I imagine people who write RSS readers still need to come up with semantics for figuring out if a title or description payload contains encoded html or not.)
15 comments
[ 3.4 ms ] story [ 37.6 ms ] threadMostly about Atom requiring IDs and timestamps, and having an overall cleaner and less ambiguous spec
<https://github.com/w3c/feedvalidator/commit/8ca2fd4e5a520158...>
Although I have found it occasionally useful for sites that have over-active bot-blocking on their feeds because Feedburner is often whitelisted.
But the undeniable victory of RSS shows the importance of being first and "easy" (even when "easy" means sweeping edge case problems under the rug). And of humans: Major publishers like the New York Times had adopted RSS and saw no need to switch to Atom because it was good enough. I'd argue the (also underspecified) CSV format is another example of this phenomenon.
(As for the entity escaping dilemma, people mostly just moved to using CDATA for their feed-embeded HTML, although I imagine people who write RSS readers still need to come up with semantics for figuring out if a title or description payload contains encoded html or not.)