Yup, the Bloglines feed for e-Literate is still broken after three months. Yup, I’ve sent them multiple emails letting them know that the feed is still broken. Yup, I even let them know (in pretty blunt terms) that I was letting my readers know about the problem. (My experience with other web service providers is that most of them–even the big ones, like Furl and JotSpot–are incredibly responsive to users in general and bloggers in particular.) Nope, they haven’t solved the problem. Nope, they haven’t communicated anything to me about their progress other than a generic “We’re looking into it” reply after every complaint I’ve sent.
Nope, I’m not the only one who’s had problems with their service.
Silent failures of service are not OK. And silent failures that affect hundreds of your users (at least) for months at a stretch are really, really not OK. I don’t care if the service is free; it is broken badly enough that it is effectively unusable. Furthermore, this kind of shoddy performance is liable give a bad name to web services in general. It highlights just how vulnerable you make yourself if you become dependent on an application that lives on somebody else’s server and is totally out of your control.
The problem appears to be with your blogging/webserver software and If-Modified-Since HTTP header support. If we don’t send that as part of our request, we get the feeds as expected. If we subsequently send that header, using the value received in the previous request via the Last-Modified header, the request times out. If you look at your server logs, you can see this in action over the past couple of minutes (a little after 5pm pacific time on Wednesday, Jan 26). I cleared out the last-modified times that we had for your 3 feeds, and re-fetched them, without problems. I then tried to fetch one of them again using the Bloglines crawler, and it timed out.
Hope this helps.
Yes, it helps, and I’ll look into it. Why it took three months to get this answer, and why Bloglines can’t read my feed when other aggregators can, I still don’t understand.
Anyway, thanks for the info.
I actually did see this posting via my bloglines account…
Just an FYI
If you read the comment above yours as well as this follow-up post (which, by the way, does not currently show up on the Bloglines feed), you’ll see that the problem hasn’t really been solved yet, unfortunately.
Michael –
Chris Curtis notified me of your problems and so here I am checking out your feeds. Did you update your backend files recently? There was a problem with the headers sent but we had the problem fixed within a few weeks of EE 1.2’s release.
Mark –
I just ran a few tests using a quick and easy PHP page reader with the If-Modified-Since headers sent with different dates (a month ago, yesterday, the contents of the Last-Modified header, etc.) and there were no time outs for any of them, so if Michael has not updated his site I am confused why Bloglines is having any problems.
I updated my site, but to 1.1, not 1.2. And again, no other aggregator or feed reader that I can find is having trouble with updates on my feeds.
I did some further testing and it seems the combination of the If-Modifed-Since and the A-IM header will cause your feed to time out on request. I just tested this on my own site, which is running EE 1.2.1 and there is no timing out (and I believe Lisa-Jill said it was working in Bloglines). I would suggest you try updating your site to the most recent version of EE and see if that helps.
By the way, Mark, I am apparently not the only one experiencing feed update problems. You might want to check for problems with your crawler.
Paul, I’ll do the update as soon as I can clear the time, but I’m confused. Why is this timeout an issue for Bloglines and not for any other aggregator?
Depends on how Bloglines is reading and processing feeds. My first tests were simply using the If-Modified-Since header without problems. However, the A-IM header is part of the RFC-3229 request stuff and it would not surprise me if many of the other aggregators are not even using it -or- have some backup routine when there is a failure with it.
Paul, forgive my ignorance, but what is “the RFC-3229 request stuff” and why is it important? And Mark, is Bloglines strictly enforcing RFC-3229 for a reason, or is Bloglines’ lack of a backup routine simply an oversight?
Ooo! You want some really fun reading? Go [url=http://www.faqs.org/rfcs/rfc3229.html]here[/url].
If you want a slightly more relevant explanation of it, check out Bob Wylan’s article [url=http://bobwyman.pubsub.com/main/2004/09/using_rfc3229_w.html]here[/url].
Think of it as a way for feed readers to only retrieve the content that has changed since they last check. Unfortunately, Apache support for the 226 header code is non-existent.