gregword

I posted a comment over on TechCrunch that sums up a lot of what I’ve saying to people in person or on Twitter the past few weeks about FriendFeed. Several applications have, of late, risen to prominence that have taken the promise given to us by RSS and improved upon it. A few of those focus simply on taking the concept of content and flipping it around to being person-focused; instead of subscribing to a blog, I subscribe to a person. Examples include Facebook, Twitter and Tumblr.

Other applications have gone one step further. They don’t (intially) offer you the ability to create content, but rather to collect (aggregate) your own content and make it discoverable via your identity (rather than a blog/brand). FriendFeed is the best example. This is RSS subscription, but you subscribe to a person, not a feed.

We have, right now, embraced the social applications idea. Blogs became just another form of media, not the personal avatar on the web, as Richard MacManus called them many years ago. The future is systems like Facebook and FriendFeed (Twitter, despite its obscene limitations on content, is building a powerful social network that could destroy FriendFeed in an instant if they chose to move in that direction).

It’s a final recognition that the web is made of people, not content.

The confluence of two events: the deprecation of the Google Search SOAP API and the release of the del.icio.us JSON API, both point to the fact that the vision of the web as a platform is dying. Instead, the success of online advertising, and it’s corallary, the page view, is dominating. So Google limits their search API to only be used on the web, and allows for greater control over how their data is used (and, doesn’t expose data as before). del.icio.us passes over traditional standards for a format which provides functionality that helps drive traffic back to their property while at the same time doesn’t open/expose their platform too much. The paradigm is widgets, which function to promote the host site, rather than allowing data to roam free and find itself reconstituted in new ways.
On the one hand, it makes sense from a business perspective; why should companies freely open up their valuable data stores if they get no benefit from it (an issue we’ve been dealing with for years, since we made all searches on Blogdigger subscribe-able as RSS feeds, and in terms of monetization, it’s been very tricky)? On the other hand, locking up and restricting data to self-promotion vehicles hinders innovation.

In the page view/advertising world, it makes sense it has come to this. But is it a good thing?

Got back from PubCon on Friday. Overall, it was a really good conference. I’ve been focused so much on blogs and social media over the past few years, in some ways I forgot that there is a whole other part of the web that doesn’t care about buzz words or hip technology, but about dollars and cents. It’s been really interesting talking to people who run their own businesses online and hear some of the issues they deal with in making real money online.

The Sessions:

A typical pattern for me at conferences: I begin by going to lots of sessions, then taper off and spend most of my time in the hallways talking to people, and towards the end go back to the sessions. I was a bit disappointed by the sessions at PubCon, I really didn’t find that they were giving me any information I didn’t already know (this includes sessions on topics I know really little about, like things like PPC and SEO). A few of the sessions I attended and my thoughts:

Feeds, Blogs, News and Search: This session was pretty good. Niall Kennedy gave a good overview of some of the technical details behind syndication formats. Rick Klau gave a good overview of some of the applications that consume feeds beyond basic aggregators. Owen Byrne gave an overview of digg, including an interesting tidbit about how many servers they have (about 90, not all for production. I always find it interesting to find out how many servers services are using, as well as their architechture, ‘cuz I think about scalability quite a lot, and love getting hints at how much iron it takes to serve various levels of traffic. We’re doing about 500K searches a day on 3 servers). Chris Tolles gave an overview of Topix. I would have liked to have had more discussion on the basics of syndication, aggregators, etc. I also think it would have been helpful to discuss the role that blogs and feeds play in helping with rankings in the major search engines (like it or not, 90% of your traffic is going to come from Google, not from feed-consuming applications). There’s been a lot of rumors and misinformation about how Google handles feeds, and if they help in optimizing search rankings. Also, based on comments from the folks from the major search engines, a blog sounds like the ideal type of site for getting ranked. Some discussion of this would have been helpful to convey the value of having a blog or feed.

Podcasting and Net Radio 101: A good session. While the session provided some good information, I thought it focused too much on podcasting from a business perspective, overlooking podcasting as content. I’m going to post more about podcasting in the future based on some of my experiences over the last two months, but I will say that, from my experience, promoting a podcast is completely different from promoting a web site. There wasn’t much discussion of iTunes or other community building techniques that work very well for podcasts.

Interactive Site Reivews and SERP Quality Control Forum: This was really cool. The panel included folks from the major search engines (Matt Cutts from Google, Tim Mayer from Yahoo) as well as Danny Sullivan and other search experts. Basically, people from the audience submit their sites to the panel for review, and the panelists tear them a new one (not in a bad way…they are really helpful). It was really informative, and amazing to me how many of these sites were suffering from so many problems that a decent content management system such as Wordpress could take care of in a snap (it’s not just for blogs). Most of the panelists stressed the importance of unique content (this is a big deal for e-commerce sites that post just snippets from databases, and all the sites use the same database), and offered starting a blog as a good way to build unique content. Many other really good tips, some obvious, some based on inside information from Google or Yahoo. Very informative.

All in all, it was a good conference, and I learned a lot, and had a good time (oh, and the Yahoo party, that was fun). If you’re looking for more coverage, check out Search Engine Roundtable for a good summary of the sessions. As for a general overview of the conference, and why I went, Chris Tolles said it best.

Please, for the love of God, leave RSS alone.

Many folks weren’t around for the whole RSS vs. RDF vs. Atom melee that took place a few years back. Nothing new really happened during that time (well, not nothing, but less) apart from a whole bunch of squabbling on blogs and listservs about formats and such.

And then people just kind of dropped it, and we got all sorts of stuff done, like, um, podcasting and Web 2.0 (or whatever you want to call it).

Call it a coincidence, but I don’t think so. Please, no more new formats. RSS has it’s quirks, and has had them for a while. Figure them out and let’s keep making stuff to make our lives better. That’s my thoughts, as a developer and a user. Take it or leave it.

While in the shower this morning, I had an idea for a FeedFlare unit. After writing up a post about it, my laptop crashed, and I lost the post. Then I realized that my idea was already implemented by FeedBurner in the paid-for service.

To state the problem: the basic FeedBurner gives me information on what my circulation/subscribers are. It doesn’t give detailed information about what, when and how readers of this blog consume the content. I can get click-through information, but for various reasons, I don’t find this something I want to do.
It occured to me that I could better track who is reading my content and when using a FeedFlare unit. FeedFlares are dynamic images that are loaded when the feed item is loaded. A simple FeedFlare that passed back the link of the item being read would, assuming the rendering aggregator parses HTML (Which most do), allow me to record, outside of FeedBurner, when and what people were reading from my feed. Unless FeedBurner further opens up their API, I wouldn’t have access to information regarding where someone read my content, or who they might be. It be nice if this information were available, but I suppose Dick and Co. have to pay the bills, right?

After I thought of this, I realized that this feature is included as part of FeedBurner’s premium package. As such, I’m not sure I would implement this and make it widely available, but I might test it out for myself.

Come to think of it, there’s no reason why a simple Wordpress plugin couldn’t be created to provide this same functionality (the tracking aspect; FeedFlare makes it easy to display information in a feed in a user-friendly way). Embed a tracker image in all feed posts, when the image loads, the URL to the image calls back to your blog, stores relevant information from the calling site (IP address, user agent, timestamp). I’m sure someone has already done this, I’ll have to check around, if you know of anything, please to be letting me know.

This would be an extremely valuable source of attention data, and implementing it in Wordpress lets the content owner own the data.

[Hat tip to Scott J; blogging about something really does help one organize one’s thoughts. I should do it more often. ;)]

Pete Cashmore says CoComment is the best Web 2.0 app launched this year. I haven’t used CoComment yet, but from reading their site, my feeling is that while it may be a useful service, to me, comment tracking should be a feature of an aggregator application, not an app in and of itself. Of course, I’m still waiting for the aggregator that does this (Alex King of FeedLounge dropped some hints a while back, but I haven’t followed up). Granted, comment tracking, until recently, was difficult, but it’s much easier now. Aggregators developers should be looking at comment tracking as an important feature.

Other than that, CoComment does look nifty and useful; I just don’t want yet another account to log into. It is interesting to consider the data they may be able to collect and how that may be used to track active/popular content, and how they may correlate that information across various blogs. Especially since some subsets of the blogosphere tend not to link, but they do comment like crazy.

I’ll post more thoughts once I get a chance to use the app.

If everyone followed Postel’s Law, no one would have to worry about Postel’s Law.

I was thinking of this in the shower, after working on some updates to our crawler (something along these lines)

(While looking for a page defining Postel’s Law, I cam a across a post from Dave, saying the same thing, only three years ago. Ah well, there’s nothing new under the sun, as the saying goes.)

I am fairly technical. It’s not a bad thing, except for the fact that when I build things, I forget that regular people need to use them. I’ve set up checks and balances to try to mitigate the affects of my techni-ness, but it’s still a problem. For a long time, Blogdigger only had a link to the plain old orange XML button; a few weeks ago I was talking to some folks interested in Blogdigger who were very non-geek, and one of the said, “How great would it be if I could get Blogdigger search results in My Yahoo!” Of course I told him you could, all you need to do is right click on the orange XML icon, copy the URL, paste it into My Yahoo on the Add Content page…and, of course, I was met with “The Blank Stare,” which is usally when the topic of conversation turns to something that doesn’t involve computers. The next day I added easy subsribe buttons to Blogdigger for all the major online aggregators.

The point is, subscribing to RSS feeds is still a major issue for non-geeks, despite the fact that efforts have been made to make the process simpler. I did a quick test today to see if I could use the one-click subscribe features provided by FeedBurner to subscribe to my feed in all the major aggregators; I wasn’t expecting to have any problems, yet I was astounded at the quality of the results.

A bit of background: my feelings are that a site should offer one feed and one feed only. You don’t need to have an RDF, Atom and RSS 2.0 version, it’s just not necessary. One feed is enough. I also like to own my feed, meaning, I want my feed to be a URL that I control, like http://gregword.com/feed/; but, I want to use things like FeedBurner, or have the option to switch if I happen to find a better solution than FeedBurner. As such, I use a temporary redirect from http://gregword.com/feed/ to my FeedBurner feed. FeedBurner explains how to setup redirects, so it’s really easy to set up.

Despite the fact that things like temporary redirects are Web Tech 101, I was amazed that some aggregators didn’t get it. If you visit my feed, FeedBurner puts nice links to the major feed aggregators right there so you can easily subscribe to my blog. Here’s a rundown on how those services compared in a one-click subscribe test (try it yourself, and tell me if I’m doing something wrong here):

• MyYahoo - although they get a listing of my latest entries, an error occurs when I attempt to “Add to My Yahoo.” A second attempt was successful.
• Newsgator - worked perfectly
• Pluck - worked perfectly
• Rojo - worked perfectly
• Bloglines - worked perfectly
• Google - adding to my Google homepage worked only after a few page refreshes; I have been unable to get the Google Reader to subscribe to my feed.

So, that’s actually not as bad I would made it out to be initially, but considering that Yahoo and Google combined probably make up significantly large portion of the market, it’s a scary thought. This has to work flawlessly if RSS is going to get anywhere.

Update: MyYahoo is working fine now…it must have been a problem with initially loading my feed. Google still has problems.

I just finished setting up Feedburner for my feed; I hadn’t set up a feed in a while, and so had not checked out the new FeedFlare feature, which allows you to add in various utilities at the end of each post, like an easy, one-click Add to del.icio.us link, or a count of the number of comments on the post, or backlinks from Technorati.

The comments/backlinks in particular caught my attention; if Feedburner was writing this data right into the feed, each time the count incremented, a typical RSS aggregator would consider the post updated, and redisplay it to the subscriber. This would be considered a bad user experience. On a normal webpage, one would use a Javascript component to provide this type of behavor, but most, if not all, RSS aggregators strip out Javascript for security reasons.

The only way I’ve figured to do this is to use a dynamically generated image, and sure enough, that’s what Feedburner is doing. This is also how Google and Yahoo do RSS ads, so a different ad can be served each time a new subscriber views an item, without causing existing aggregators who happen across the same item to think the item has updated. By hardcoding a link to a dynamically generated image, the HTML code remains the same, so aggregators don’t think the item has been updated, while the subscriber sees the dynamically generated values each time the feed is presented. It’s a smart hack.