njr 3.0

Several of the feeds I was subscribed to were moribund and/or broken. There are some significant differences in how the various readers handle them. Here are a couple of representative examples:

FatBits, John Siracusa’s old Ars Technica staff blog

The URL for this feed was http://arstechnica.com/staff/fatbits.rssx. This returns a 301 (Moved Permanently) redirecting to /staff/, which is not a feed. This is clearly breakage on Ars’s part.

• The Old Reader scraped the redirected-to page and pulled a feed URL out of it, http://feeds.arstechnica.com/arstechnica/index/, but kept the old title, so the main Ars feed appears as “FatBits: John Siracusa’s Journal”.
• Feedly just says “No unread articles”; I’m unaware of a way to view errors from it, or even the feed URL without pulling it out of the page URL.
• Feed Wrangler also scraped the redirected-to page, but changed the title to “Ars Technica”.
• NewsBlur somehow figured out that the correct URL should be http://feeds.arstechnica.com/arstechnica/staff/fatbits/ and used that. It changed the title to “Ars Technica » John Siracusa” based on the feed, as well.

Clearly NewsBlur’s behavior is the best here; if you try to subscribe to a broken feed, it also provides a search interface with some great suggestions. Feed Wrangler is acceptable; The Old Reader less so, and Feedly not at all.

Canned Platypus, now Platypus Reloaded, Jeff Darcy’s blog

The URL for this feed was http://pl.atyp.us/wordpress/?feed=atom. At various times this was a broken redirect to http://pl.atyp.us/wordpress/index.php/feed/atom/, which currently returns 404 Not Found.

• The Old Reader followed the broken redirect, but does not show the 404 anywhere, even when I click “Refresh”. Articles are shown, presumably from when the URL worked.
• Feedly probably followed a redirect at another time, so it has http://pl.atyp.us/wordpress/feed/atom/ instead, which also returns a 404 — this information is not visible anywhere I can see, and there are no articles in the feed visible.
• Feed Wrangler was able to discover a valid feed address, http://pl.atyp.us/wordpress/index.php/feed/, and even displays a combination of the old and new feeds with appropriate site titles.
• NewsBlur displays an ! icon that, when clicked on, pops up a resolution interface with useful options. Clicking “Fetch Feed From Website”, as recommended, causes the feed to correctly change to http://pl.atyp.us/wordpress/index.php/feed/, though I repeatedly saw hangs when clicking “Retry Fetching and Parsing”.

Feed Wrangler did the right thing automatically; in addition, its combination of old and new feeds is exemplary. NewsBlur has a pretty good solution, though it requires manual intervention, exhibited some flakiness and necessitated the folder be expanded to see the ! icon. The Old Reader and Feedly’s ambivalence to errors was not appreciated.

4:10 PM | Computing, Web