Unintended consequences of syndication

A while back, when this blog lived over there, I decided to include a recent links widget in the left column. So I injected some JavaScript into that column in order to read my JSON (JavaScript object notation) feed from del.icio.us and convert it to HTML. One unintended consequence of this arrangement was a change in how I used del.icio.us. Of course it’s always true that your stream of bookmarks is public — except for the relatively new option to bookmark privately. You may even find that people are pointing feedreaders at your stream of bookmarks and subscribing to them. But still, it doesn’t quite feel as though you’re publishing those bookmarks in a really explicit way.

If you do decide to explicitly publish your bookmarks in a sidebar widget on your blog, it may change the way you bookmark. It did for me, anyway. The balance shifted away from purely personal information management and toward the kind of editorial sensibility that governs the blog. It was around this time that private bookmarks became available in del.icio.us, and that’s been helpful. If I’m researching something and I just want to collect a list of resources labeled with some obscure tag meaningful only to me, there’s no need to flow that stuff onto my blog page. Conversely, if I want to draw attention to something in a public way, I can. It sounds great in principle, but in practice I think the friction involved in making that choice on a per-item basis made me less likely to bookmark either publicly or privately.

Here’s another unintended consequence that illustrates the surprising things that can happen in this web of information we are spinning. I realized the other day that my public del.icio.us bookmarks are appearing on every page of my mothballed InfoWorld blog. To stop that happening, I’d need to tweak its template — but I no longer have access to it!

Meanwhile, ironically, I haven’t yet figured out how (or whether) I can inject the del.icio.us JSON feed into the hosted WordPress blog I’m running here. For now, I’ve decided to embrace this constraint. Perhaps if my del.icio.us account feel less directly connected to what I am publishing now, I’ll use it more freely.

These scenarios are rather odd, quite interesting, and slightly scary. As building software systems with components morphs into building information systems with feeds, they’ll become increasingly common.

18 Comments

  1. I use a feedthru tag to pass particular delicious bookmarks through to feedburner, where it gets mixed on a daily basis with my feedburned blog feed; I make sure I always annotate these links, which I don;t always do for other bookmarks.

    The other day I adopted pageflakes for my personal page (http://ouseful.open.ac.uk) and reusing the feedthru delicious feed seemed a sensible thing to do there, too. My attitude towards things I tag with feedthru is very much that I am publishing/recommending those links.
    tony

  2. I don’t see why you couldn’t put your stream of links into your WordPress template, I have in mine. I use a personal tag for my craft blog that’s unique in my bookmarks. Any bookmark that I tag with that unique tag will get published to my sidebar.

  3. Hi Jon,
    I fully understand what you mean. Over the past few years I’ve experimented a lot with my various del.icio.us streams, in particular how I structure each bookmark to make it easier to grasp its scope and importance. At the moment I focus primarily on bookmarking blog posts and news items about RSS technology. I publish my bookmarks as an animated headline widget in the header of my blog CleverClogs. Until today I used FeedBurner’s Headline Animator for this; right now you can see the MuseStorm one in action instead.

    As you can see if you open my blog, I add information such as the author’s full name, the site name and the publication date to the end of the del.icio.us Description field of each bookmark. I also write a 255-char summary of the post into the Notes field. This summary may contain some comment or insight that I might have about the topic. This way I create some sort of mini-blog that works pretty nicely for me.

    One discovery that highly disturbed me today is that the del.icio.us tag “RSS” is heavily being spammed. Have you noticed this too?

  4. Did you consider 2 differente users? One for public stuff. Another for private use.
    I use a similar approach. Everything is public in my public user and the other way around in the second.

    I guess the standard bookmarlet mechanism –I don’t use it– to post sites can’t handle this kind of arrangement.

  5. “Oh, what tangled webs we weave, when at first we practice…”

    Well, what is it we are practicing here. We practice using these nifty syndication technologies we create, but IS there a practical use? It’s very cool for those in the technology trade, but which (single) button do I press to extend the reach of my “real world” message that will push it the rest of the way around the “real world?”

    “Button, button, whose got the button…”

    John
    by hammer and hand great works do stand
    by mind and thought best words are wrought

  6. “Would a ‘myblog’ tag work for you Jon?”

    Yep, that’s a good solution.

    “I don’t see why you couldn’t put your stream of links into your WordPress template”

    Actually you can. I was, of course, making it too complicated. I was looking to reuse the JavaScript code I used before, and it’s not clear to me how do that in the hosted WordPress. But of course there’s no need, you can just drag the del.icio.us sidebar widget onto the template.

    “One discovery that highly disturbed me today is that the del.icio.us tag “RSS” is heavily being spammed. Have you noticed this too?”

    No but it doesn’t surprise me.

    “Did you consider 2 different users? One for public stuff. Another for private use. I use a similar approach. Everything is public in my public user and the other way around in the second.”

    I’ve thought of it. The problem there is I’d often want to search across and interact with the combined set.

  7. “but which (single) button do I press to extend the reach of my “real world” message that will push it the rest of the way around the “real world?””

    There’s never just one. And I’m learning to look for them in places I never previously considered. For example, I am trying to influence my school district to accelerate its adoption of a student information system. To give people in my community an idea of what the benefits would be, I made a screencast:

    http://weblog.infoworld.com/udell/2006/11/30.html

    It’s on the Net so anybody can see it, right? Except that’s the wrong channel in this case. So last week I dubbed the thing onto a tape and took it down to our recently-launched community access cable TV station. And when I tuned to channel 8 last night, there it was.

    My “real world” message has a lot to do with the Net. But it is a real world message, and I’m learning, finally, that the Net — even today — is not necessarily the first or best place to deliver that message.

  8. Greetings from Brazil, Jon.

    Thanks for the very thought provoking post. I think that one important unintended consequence of syndication is especially when you couple it with del.icio.us-style tagging: once EVERY view of del.icio.us information is an VERY easily-acessible URL and also an XML feed (by view I mean any tag, user, or more importantly, any combination/intersection of tags), I feel that with time we tend to start mentally dealing with tags and tag combinations THEMSELVES as “locations”, just like today each website address is dealt by our brain as something analogous to a physical location. So, when you label some information with a specific tag or set of tags, your brain will think of it as storing that information on a specific ‘information bank’ that is different from another ‘information bank’ that houses items with different tags. Your first comment on this post by Johnny K is a good example of that.

    At the same time, as we start dealing with tagged information in many web services, like Del.icio.us, Flickr, and Blogging services that support post labeling/tagging, there will be more and more of a demand for an universal interface for dealing with our personal information that is based on tags and tag intersections, sort of an ‘tagging-client’. I can see a first step of this in Google Reader, that allow the influx of new information that the user subscribed to be ‘tagged’ based on the feed of origin. It does even allow you to add post-specific tags as you read them. But as an RSS reader, you can only receive third-party information there, not feed it yourself with information by typing it. Anyway, it’s a first step.

    Once when we start to demand for such universal ‘tagging-client’ AND also start to think of those tags and tag combinations as locations, we will reach a third consequence that involves syndication and tagging: those locations will become triggers for dealing and for syndicating data in different ways. So you can imagine someone using his future ‘tagging-client’ of choice and feeding it with a myriad of different info (some from RSS feeds and some typed and edited by himself). He could set up that any information that he puts into the ‘my blog’ location (by that I mean any information that he tags as ‘my blog’) gets automatically published into his blog account of choice. Also, any other tags that the entry bears that are not the ‘trigger-tag’ could be also published on the blog, provided that the blogging engine supports tagging/labeling. With that, provided that I have nothing to hide, I could even run an universal blog and have people to subscribe just parts of it by planning ahead on the tagging. By tagging something as ‘my blog, personal’ I could create a hook for my personal friends to subscribe. By tagging something as ‘my blog, professional’ I could create a hook for my co-workers or contractors to subscribe. But if I prefer to run two separate blogs, i could also make those two different locations to separetely act as ‘trigger-locations’ for two different blog accounts. So, if I tag something just ‘my blog’, this would not trigger any special action, unless if accompanied by any of the ‘personal’ or ‘professional’ tags.

    But, while it is important that this hypothetical ‘tagging-client’ has the ability to feed itself with RSS information, tag it as configured by the user and make it fully searchable, it would be important to provide an efficient way to separate self-generated content from feed material. Otherwise the user that subscribes to a lot of feeds would face a hard time on finding his own info.

    Also, it would be interesting to see as this would enrich the metadata about information that gets to be published and re-published on the blogosphere, because each blog post could end up ‘inheriting’ many generations of tagging: 1st) how the original blog owner tagged the entry; 2nd) the tags that the receiver’s RSS portion of the ‘tagging-client’ added to the entry based on how the user configured the feed when he subscribed to it. 3rd) fine-tuning of a specific entry tags by the user when/as he digests and returns the information back to the blogosphere.

    Returning to your specific case, if such universal ‘tagging-client’ existed, you could save your bookmarks privately anytime (let’s say you also tagged them all with the ‘bookmark’ tag for your own organizational sake, alongside with other case-to-case bookmark-specific tags). But if you decided to publish one of your existent bookmarks to your del.icio.us account, you could simply go to the ‘bookmark’ location, pick the bookmark you want to publish and add the ‘delicious’ trigger-tag to it. The client would then interface with del.icio.us to add the URL to your stream.

    This is pretty much everything that I mean by people starting to think as tags and tag combinations as ‘locations’. Other thing that also strikes me is that although we normally (and rightfully so) display RSS feeds in a river-of-news fashion (sorted by reverse-chronological order), XML is actually a format for unstructured data. With this in mind, we could use the concept of locations inside of this imaginary ‘tagging-client’ to add ‘location-specific datafields’ (attributes) to entries that fit into a particular location. So locations could also become a way of custom-building databases inside of the ‘tagging-client’, as once we have location-specific attributes, we could use these attributes to sort – and find – data.

    This is an idea that has been on the back of my head for some time, I even have a blog dedicated to it. thanks for giving me the opportunity to express it in some context.

  9. Paulo, thanks for posting that wonderfully thoughtful essay. If I were you I’d have just blogged it and linked to here :-)

    “I feel that with time we tend to start mentally dealing with tags and tag combinations THEMSELVES as “locations””

    Absolutely. This is what my elmcity.info project is trying to show people: that intersections in tagspace are, as you say, locations. It’s a tough slog at the moment, because of the abstraction involved, and I’m struggling to find ways to make it more concrete.

    “There will be more and more of a demand for an universal interface for dealing with our personal information that is based on tags and tag intersections, sort of an ‘tagging-client’.”

    And not just a client. There will be cooperating services too. For example, spelling and synonym services to help tie things together. Some will be general and some domain-specific. Some will be free and some will be paid for (and worth paying for).

  10. Jon, thanks for the very kind words. It started as something that would be a small comment and then soon got post-size :-o. I will very soon adapt it for posting on my blog too. I have some other essays there too, mostly describing how I think such a program would ideally function. I have some interest on sparkling some discussion on this as i’m a law student and information addict that have to deal with lots of information, and I see tagging and blogging as the keys to dealing with information overload. Better said, blogging worsens (or IS) the problem, and tagging would be a tool for taking blogging to the next level by somehow sort a huge outpouring of information. Can’t manage feeds on a hierarchical tree.
    Salutations!

  11. FYI, Jon/Paulo…looks like a relative url munge error is showing up in Paulo’s bloglink above (in comment #15…btw, comment numbers are nice! Makes things much easier to specify.) I’ll fix it by hand to track down his URL, but just wanted to point it out. Take care!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s