Automation’s great – but I trust humans more

Repeatedly I’ve found myself being let down by the latest and greatest RSS tools.

I’m not really sure how I can get around this problem, other than by being constantly vigilant – to the extent that it’s almost easier to forget RSS altogether and just monitor ‘by hand’.

This is how it should work

I’m currently working on the best way to use the various RSS tools out there to create a flexible yet powerful monitoring system. To my mind a good solution is:

  • Yahoo Pipes for processing. Yahoo Pipes is an RSS mash-up service, so you can do stuff like take feeds and bring them together, split them apart and filter them, all through a nice graphical interface. It is so flexible and powerful, and you can create really nice modular pipes that slot together like Lego. Want to search for blogs? Insert your ‘blog search engine’ pipe. News? Same. Microblogging? No problem. One of these days I’m going to put a quick tutorial about Pipes on this blog. But not for a while yet. I explain why below.
  • Feedburner for future-proofing. Feedburner is an RSS ‘add-on’ service, enabling you to add titles and descriptions to feeds but, for my money, its most important feature is the feed renaming service. So if you have a feed with the URL http://x.y.com/feeds/asdhJAH72jjaaSS99.xml, just plug that into Feedburner at one end, tell it you want it to be called ‘My lovely feed’ instead, and from then on it also has the URL http://x.y.com/mylovelyfeed. So whatever RSS feed is coming in, it retains the same address. This means that, if you have to use different feeds coming in, you don’t then have to scratch around looking for whatever you had those feeds plugged into. If you know they’re going into Feedburner then you just have to change it there. Everything else stays the same.
  • Google Reader for archive and analysis. Google Reader is an online RSS aggregator, so all you need is a Google account and you can use Google Reader’s very powerful features with no installs or upgrades needed. So, you’ve got a cool RSS feed created from Yahoo Pipes, going into Feedburner and retaining the same lovely name. You’re now ready to plug that lovely feed into any other RSS-enabled tool. The next thing you want to do is analyse it, so plug it into Google Reader and suddenly you can filter for ad-hoc queries, star or share items, go through archives, even produce web pages for clients and extra feeds.
  • Netvibes for display. Netvibes is an online RSS aggregator too, but while Google Reader’s good if you like lists, sometimes people like columns. So take your lovely RSS feed and this time use Netvibes to create a ‘front end’ for your feeds. So, you monitor the Google Reader stuff, while the client gets to see a really neat dashboard type display. You can add charts and all sorts of bells and whistles.

There just one problem: it doesn’t

I know this can work. I’ve seen it work. But there are frustrations along the way and recently I’m starting to wonder whether these services can be relied upon to work.

A couple of weeks ago I noticed that some critical Feedburner feeds had ‘died’. I was relying on them for data to come through for some important monitoring work. On further inspection I noticed one of the feeds had gone above the 512KB limit for Feedburner, but, while annoying, that didn’t explain the other problems I was having. Other people were commenting on Twitter about similar problems, and about the same time I noticed my Feedburner-enabled subscriptions had halved.

The solution was to bypass Feedburner altogether and just use the Yahoo Pipes addresses instead. But this was far from ideal. I want to use Feedburner for control over the address. I want to feel I can rely on it.

But the real culprit in all this is turning out to be Yahoo Pipes. I have invested considerable time and effort into getting to know it. I’ve got a system that builds queries from keywords, goes out to just about every RSS-enabled social media source I can find, grabs those feeds, filters for them, appends information to the titles and spits them out in virtually any configuration needed. I’ve tested it all, and I know it works.

But about two weeks ago I noticed Yahoo Pipes getting sluggish. It didn’t matter what I was using to access it – my PC at home, my laptop at work, IE, Firefox, whatever.

And this weekend, I can’t do anything with it. I need to add some tweaks to the system, but it either times out, or refuses to save my changes.

I mean, as I said earlier, I’d love to pass on some of what I’ve learned on this blog. But I cannot. Even when the system is working it’s just too slow. I find I’m wandering off to stroke the cat or do the crossword while Yahoo Pipes churns away.

So again, I have to ask: can I rely on it?

I can’t see the silver lining for the cloud

This is, of course, a criticism of cloud computing. While I absolutely love the idea of harnessing the power afforded me by Yahoo’s server farms to do weird and wonderful things with RSS, I hate, detest and loathe the notion that I’m totally dependent on them to be able to do so.

If, as has been happening for the past few months, I continue to creep into the cloud, I know that one day I’ll have really seriously important stuff in, say, a spreadsheet on Google Docs, that I cannot access when it’s critical that I can access it. Or I’ll get into trouble with a client because they’ll blame me for not making sure their RSS feeds are working properly.

Is the answer that I just don’t put all my trust in these services? Do I keep local versions of docs, just in case? In which case, what do I do for RSS monitoring? I mean, can I pay someone money to give me a better service? Is that the real issue here?

Perhaps the Luddites are right

So, to get back to my original point: at what point do I totally lose faith in these services?

I’ve spent enough time testing my systems to know that they work. The problem is, the services themselves don’t seem to work properly.

So do I monitor constantly and vigilantly to make sure everything is tickety-boo? Do I just hope that, come the day I’m dependent on Yahoo Pipes to work, and it doesn’t, I can quickly think of a workaround as I did the other day?

Or do I eventually decide that actually, it’s more reliable and in the long run more cost-effective simply to monitor individual blogs by visiting them on a daily basis? I mean, there’s something to be said for this. I would certainly get to know those bloggers more intimately. But this solution just doesn’t scale up. It’s not workable.

No. We need services like Feedburner and Yahoo Pipes to provide the service they say they will. I know they’re not bound by the kind of service level agreements that would be in place if we were actually paying them, but they surely have to operate within the bounds of, well, operability.

Because if they don’t, someone else will. I’ve already been checking out Microsoft’s Popfly mashup creator today to see if it can do what Yahoo Pipes should. And it’s already looking promising. We’ll see.

Are you one of the people messing up my feed stats?

This is a plaintive call to anyone and everyone who uses RSS to read this blog:

Why? Well, I’ll tell you…

When I started blogging, I didn’t really want anyone to know about it. Perverse, I know. So, I called myself The Friendly Ghost, in relation to being a copywriter (ghost writer? Geddit?), and called my Feedburner feed http://feeds.feedburner.com/FriendlyGhostCopywriter.

Then I decided to ‘come out’ as Brendan Cooper. I bought the new domain, had a redirect, and started up a new feed – http://feeds.feedburner.com/brendancooper_mainfeed – and simply told people to switch over. This made even more sense after I stopped being a copywriter.

But they didn’t. Several times.

Moreover, I noticed that the old feed was continuing to gain subscriptions even though I thought I’d removed all mentions of it from the blog. Then I realised, with a certain amount of annoyance, that a bug in the RSS widgets to the side of this blog meant that, although I was specifying the new feed address, and the RSS icon pointed to it, the text did not (you can still see this effect if you mouse over the ‘Brendan bungs’ text). Of course, WordPress support hasn’t answered my queries about this.

What was I to do? I was running two feeds, which meant my feed stats were a bit weird. I wanted to run everything on just one feed, but didn’t want to lose all the subscribers on the old one. So I deleted the new one, with a redirect to the old one. And I’ve checked – there are no mentions of http://feeds.feedburner.com/brendancooper_mainfeed anywhere on the interweb.

At first this was perfect. The deleted feed reset to zero, the other one increased. All was well.

But now I see that people are still subscribing to the deleted feed! How is this possible?

Furthermore I also recently realised that, if people use the funky Google Reader subscribe button or any other reader’s ‘subscribe as you surf’ feature, it subscribes them to this blog’s original – that is, non-Feedburned – feed, which could be either http://thefriendlyghost.wordpress.com/feed, or even http://brendancooper.com/feed. This means I have no idea how many subscribers I really have, because Feedburner won’t pick them up!

There is a Feedsmith widget that would solve all of these problems. But that’s for self-hosted WordPress blogs, not those run at WordPress.com like this one.

So my situation is that:

  • People still seem to be finding the deleted feed and subscribing to that. I don’t know how/why.
  • There is a completely unknown amount of people subscribed to this blog’s feed that Feedburner can’t tell me about.
  • I cannot run any utility to bring all these feeds together.

Hence the plaintive keening at the beginning of this post. I’ve learned a very valuable lesson: make sure you give your feed a name that isn’t related to anything transient or specific, such as your online identity (which may change) or your job (ditto).

What a palaver. Perhaps I should just go back to quill and parchment?