A “Connection Point” is a small micro-wiki. You would attach a connection point to blogs, blog comments, emails in email archives- pretty much anywhere where there was information.

A story it illustrate: One day, LionKimbro was struggling with the IP_ADD_MULTICAST option in a setsockopt call. Windows kept giving him back error 10042, which means: “I don’t know this option.” “That’s rediculous,” thought Lion, and he did a google search. Lion was able to find several e-mails from people discussing this problem, long ago, but no solutions. Finally, after a few hours of working and searching, he discovered, deep within google, a link to a web page that had the solution.

What we’d like to be able to do is point those old e-mails, with high prominance in google, to the page that actually has the solution.

To do that, we need to be able to annotate the e-mail. The “ConnectionPoint” solution to the problem is to have little micro-wiki attached to each e-mail in the archive. PeopleFromTheFuture? will come and integrate the KnowledgeArtifacts? with the understanding of things that comes with time.


(We are presently discussing the merits of detached annotation systems and ConnectionPoint systems.)

Every archived email in a mailing lists archive, every blog post, every response, needs an associated “ConnectionPoint.”

That is, they need a little micro-wiki box next to them.

Whether it’s moderated, or whether it’s open and community managed, I don’t care.

Those little boxes just need to be there.

That way, the people who are searching and using those things, they can connect them all together with the network of related materials.

I heard that there was some sort of distributed web-annotation network that used wiki syntax. I wonder where I heard about it… On Meatball?

I have a problem with the annotation systems; They are so many, and they are so fragmented. I think we need a consensus reality. Maybe not, I dunno. Do you know what I mean?

I know what you mean. We need critical mass to get them going. That’s a real problem. We don’t have an incremental upgrade path from here to there…

Two strategies come immediately to mind:

  • Network existing annotation systems. (Do they allow reworking what came before? I don’t know much about them.)
  • Start sticking little mini-wiki ConnectionPoint onto stuff.

It’d be an interesting experiment to keep a blog or mailing list with these little boxes on the side.

It’d be kinda neat to see what people thought about what was in there, a year down the line.

Alex, maybe the thing you saw was the wikalong extension for firefox?

It’s a sidebar that opens along with your page a corresponding page on a wiki, so that you can leave comment on the viewed page.

In fact a comment on PlanDuSite? existe (not from me and maybe only when using the french interface)…

I’m the dev for wikalong . I’ld be happy to share any ideas about it, or work with you to extend it. (also, trying to get the hang of this wiki syntax) My next step was to do the same thing I’m doing with Firefox, with Thunderbird’s message-id’s. Thereby having a similar affect to your desire to annotate emails. -JohnCappiello john - wikalong

Perhaps we should just call it MicroWiki?.

John: I sincerely hope it works. My personal belief is that it would be easier (and still, enormously difficult) to get people who write web software to implement ConnectionPoints. Specificly: blog software, and e-mail archive software. (I’m not sure what to do with bulletin boards.)

My skepticism of annotation systems is this: The things I care to annotate are generally very obscure. For instance, I had a problem calling setsockopt with option IP_JOIN_MULTICAST, I kept getting back error 10042. The top google hits were all people confounded by the problem. Waaay back deep in google, I found a solution.

Suppose I use an annotation system: Who is also going to use it? The problem is that what I am looking at is extremely obscure. But even more obscure is the requirement that someone have this problem and be using the particular annotation system that I’ve chosen.

There is also a problem of “policing:” the annotation system is the sole government for annotations. (This, however, can be avoided, by somehow routing annotations for particular URLs to particular servers.)

Perhaps the way to grow the annotation system is to use it within a particular community, and establish habitual use there? Improve the device in your community. Then find sympathetic communities, and, by strength of personal connection, get them to use it too. Lather, rinse, repeat. With time, the annotation system may become dominant, and everyone in the annotation community using it. With further time, everyone in the world may be using it. There are likely alliances and protocols and standards formed and adopted in the course of this path.

Overall, though, I think it would be easier to make software with ConnectionPoints implemented. This is because it requires no special “browser” tricks; it is a shared reality.

I don’t mean to take any courage out of persuit for the annotation approach; I don’t really know which way will work. I just mean to explain why I am taking an alternate route. But I think both approaches should be explored.


  • Perhaps we should call ConnectionPoint “MicroWiki?.”
  • Skepticism re: AnnotationSystems?.
    • obscurity - useful annotations seem to be obscure, getting system going seems hard
    • governance - the single system is used to govern over the entire web’s annotations
  • Skepticism aside, the strategy would seem to be to grown communities of users in their particular domains.

I think the two are both complementary and could both be useful. The annotation approach will probably be most useful because it is purely “client-side” in that it doesn’t require any effort (or authorization) by the original content provider. However, ConnectionPoint could be useful for web site developers who want to add interactive features to their non-interactive site with minimal effort.

Re: stuff on MeatBall, FridemarPache and I were working on AnnotationWiki awhile ago – we didn’t have much time, so we found a way to control CritLink annotations from a wiki.

However, in the long term, I think a browser-integrated strategy like Wikalong or Annotea is the way to go. Right now I think Wikalong is the way to go.

And btw, JohnCappiello, great job! Wikalong is really cool!

We don’t have a WebAnnotation page of our own yet, so I’m just attaching this here:

  • Make a firefox plugin that can automatically link words or phrases. There may already be one, even. Then use WikiMetaData to form the database.

This is different than traditional WebAnnotation: Instead of marking pages, we’re marking information. It could also hypothetically refer to all information sources, not just web pages.

ex: You are reading an article, and it mentions "Ian Pearson." You are hooked into a Futurist’s collection. They want you to know that he's not well known for good predictions.

It’s not a specific page that’s being tagged, it’s the name “Ian Pearson.” Or “British Teleeommunications.”

PurpleNumbers was supposed to fix this… EugeneEricKim seems to treat purple numbers as just one possible implementation of GranularAddressability?. See his blog entry on Purple Numbers: Optimized for Synthesis. I talked to Eugene at Wikimania 2005 about purple numbers, and he was interested in that automatic generation of local named anchors based on words starting paragraphs (Oddmuse:Paragraph Link Extension). That was a proof-of-concept thing that never got used.

Nonono; I think you misunderstand. I’m not talking about addressing parts of a page as a link target.

I’m talking about: “Whenever I see a phrase on a page anywhere on the Internet, I want to be able to tap it and annotate it, perhaps with popup text, perhaps with a link to a page, perhaps with a link to a page in a PageDatabase.”

Following from the example above: Everywhere where the phrase Ian Pearson is found, we want it to link to Futures:BritishTelecommunications?.

(Warning: This technology can lead to tribal warfare.)

You’re reading a newspaper online, and his name is in there. Bam, you get a little notice: “Danila thinks this guy is full of (it), because he had wacky predictions for 2005 in 2001. (follow link to more Futures community thoughts on Ian Pearson.)”

Doesn’t a lot of the benefit come just from writing a blog entry that describes the problem and solution? And that entry could link to the existing pages you’d think of annotating. Not quite wiki-esque BackLinks, and Referrer-spam devalues links to some extent, but there’s still plenty of value there.

(If I now Google for “calling setsockopt option IP_JOIN_MULTICAST error 10042” (not as a phrase), this page comes up as the only match.)

Maybe Google should add a “pages linking to this” link for each item in its return list. That might make links even more useful…

Or does publishing a page seem like too much work?


Define external redirect: KnowledgeArtifacts GranularAddressability AnnotationSystems BritishTelecommunications MicroWiki PlanDuSite PeopleFromTheFuture

EditNearLinks: EugeneEricKim CritLink AnnotationWiki PageDatabase MeatBall ConnectionPoints