OK, after getting some helpful comments on my earlier post, mostly by a pMachines programmer (along with a somewhat more passive aggressive but at least marginally helpful post from somebody at Bloglines), I think I’m beginning to understand the problem. The short version is that there is something wrong with my blog feed but the Bloglines crawler and customer service reps both handle the situation very poorly.
Bear with me on this, because it taxes the limits of my technical competence….Apparently, there is an HTTP specification called “RFC-3229” that’s designed to minimize bandwidth when refreshing a page that’s been modified. It lets the software that’s fetching the modified page download only the bits that have changed (or the “delta”). This is generally a good thing, especially for RSS/Atom feeds, where each feed reader may hit your feed page as often as twice an hour. It helps feeds bear up under heavy traffic.
But there’s a catch.
And the catch is that RFC-3229 is not universally implemented yet. According to the pMachines rep, Apache support for the (presumably related) “226 header code” is “non-existant.”
Expression Engine (the software that powers my blog) does implement RFC-3229. There was a bug in versions 1.1 and 1.2 with this part of the code. (I upgraded to 1.1 at about the time that the Bloglines feed froze.) It has since been fixed, and I will upgrade my site to the patched code as soon as I can. (Unfortuately, it may be another week or two before I can clear the time.)
So, if there’s a problem with my feed, why is Bloglines the only aggregator that’s barfing? So far, Bloglines isn’t talking, but the pMachine rep suggests that the other aggregators either ignore RFC-3229 or have some kind of fail-safe code that kicks in if it’s broken on a particular feed. This makes sense.
If Bloglines is intentionally being hardcore about enforcing the use of this protocol, then they should prominently post a notice letting people know that their feeds will break if they don’t support and provide some sort of validator on their site. If, on the other hand, they are breaking feeds by simple oversight, then they ought to fix it, pronto. According to Scott Leslie, my feed is not the only one that is silently failing on Bloglines.
As a sidenote, let me reiterate that Blogines’ customer service has been heinous throughout all of this. They still have yet to reply to me via email for any of the many tech support messages I sent them. Instead, after I yelled at them in public to get their direct attention, I get a note posted as a comment on my post (clearly intended to let my readers know that it can’t be Bloglines’ fault ‘cuz my feed is “broken”) by a guy who does just enough investigation to show that my feed is timing out for them and doesn’t bother to ask the additional question of why my feed breaks for them and nobody else. Nor did he (so far) check back on the comments to see my replies or the replies of the pMachine rep. Whatever his goal was for posting on my blog, it apparently wasn’t dialog.
I’m still having trouble with my feeds and bloglines as well. I too am using Expression Engine – I hope you get it fixed and then hopefully I will be able to as well.
Unfortunately, right now I’m also having trouble with my hosting company and cannot exactly focus on my RSS feeds.
I use pMachine Hosting; they’ve been great.
I don’t understand how a service could be hardcore about enforcing the use of RFC 3229. All we do is include the HTTP header line ‘A-IM: feed’. That’s it. What happens with your feed is a timeout; we don’t get a response. Timeouts can happen for any number of reasons (many temporary) and do not in general indicate any one particular issue.
Many (most?) desktop aggregators don’t include this line. We include it as a courtesy, to try to help reduce bandwidth costs for content providers. And has been documented elsewhere, it does indeed reduce bandwidth usage. It’s unfortunate that it apparently causes an issue with your installation, and I’m sorry that it does so. But for the vast majority of feeds in our system, it at worst causes no harm and at best helps the content provider.
Mark, permit me to state the obvious: If Bloglines implemented “A-IM feed” as a “courtesy,” you have failed to make me feel like I have been treated courteously. I am spitting mad. (And from what I can tell, the Bloglines subscribers who are e-Literate readers aren’t too happy, either.)
Try to see it from my perspective. Like many bloggers, I am not a technologist. I don’t know a damned thing about HTTP headers and, up until yesterday, I had no idea that they could prevent an aggregator from being able to read a valid feed. That said, I do everything that I’m told I should do to make sure my site works properly:
– I use reputable blogging software
– I use a reputable hosting company
– I run my feeds through a validator any time they are changed
– I test my feeds against multiple desktop aggregators
– I subscribe to my own feeds so that I can continually see that they are still working
– I hired a developer to help me implement my last major upgrade
Yet, despite all this due diligence, Bloglines–and only Bloglines–can’t read my feed. I have close to a hundred readers who think I have gone silent for three months. I have no way of informing them what’s going on. And I have had no way of diagnosing the problem, never mind fixing it. Nobody at the Bloglines help desk responds to my multiple emails requesting help.
If you are going to implement features that are not standard in feed readers and that can cause Bloglines to be unable to read a feed, you should–at the very least–provide a FAQ someplace prominent explaining common causes for your crawler to be unable to read a feed, how to test for different causes, and what to do about them. Better yet, provide me with a utility that lets me test to see if turning the non-standard feature (in this case, “A-IM feed”) on and off has an effect on the problem. Best of all would be to fix your crawler so that a feed thattimes out will be rechecked without the A-IM feed request.
The bottom line is that, if a blog feed validates and shows up in typical desktop aggregators, the onus is on Bloglines to make it work with your service or, at the very least, make it easy for content providers to figure out why it doesn’t.
I don’t think that’s unreasonable to ask.
I thought Blinger and I were the only two EE users experiencing this problem. Nice to know we’re not alone.
I’m still waiting for a reply that I sent to Bloglines CS six months ago when I first experienced a problem (the problem later resolved on its own).
Bloglines just started sucking some more!
People can’t log in, those who can get logged in to someone else’s account, like I get to some Chinese guy’s account and could post in his blog.
Email subscriptions — the last remaining distinguishing feature of Bloglines — don’t receive emails (connection to mail server mx.bloglines.com times out), Bloglines’ support doesn’t answer the reports of breakage, and the Support Forum is 98% pet food spam and never looked at by anybody at Bloglines, apparently.