JB says: Here’s an idea; call it “Web Excursions” (Webcursions?)
You write a blog post, and in that post you have contextual links– hyperlinks –but they’re special, because rather than just linking offsite, these links take you down a pathway, on an excursion which eventually circles back to your blog post, or article or story or whatever.
For example, somehow you link off to Wikipedia, and in that Wikipedia article is some kind of mechanism for you to take the next step down the pathway that the author has chosen, sort of choose-your-own-adventurey style. Maybe the pathway links are a special color or have a special icon next to them
Steve says: Hrm, kind of like the old fan ring networks but different.
JB says: Right. The next step in the path is clearly identified, so you can follow the steps and don’t get lost in branches. Seems kind of semantic-webby.
I was just reading this blog post and it generated that idea. If you scroll down to the section headed “so how exactly did i get here?” where he talks about making a CD rom that contained “excursions”.
We should be able to build “excursions” into any kind of Web content. We can already link offsite, what’s missing is the mechanism to bring the user back around to our context. We can only do that if we build it into our web site ourselves, but that eliminates the possibility of (easily) referring to other people’s work to supplement your own. Maybe it shouldn’t be easy, but I kind of think this kind of reference is what the Web’s ultimate purpose really is.
Steve says: Right. Did you ever watch James Burke on PBS? He had a couple short series
Where he’d start with one thing, go off on a whole series of contextual type links and then bring it back around to the starting point.
JB says: Right. Exactly.
[OK Internets, I have pushed the idea into your ethertubes, now spontaneously generate a startup that will build this for me! Just like you did for the Networked Archive idea. Here is some more chat, in a more practical vein, about how one might actually pull this off.]
JB says: I think you could pull of this Webcursions thing with an Apache plugin. So in the originating page, the links that start a Webcursion point back to your server.
Steve says: ah, yes
JB says: The Apache server grabs the destination page, parses it, and replaces any links to the *next* pathway step with the special Webcursion code, and so forth, so the user never leaves your server.
I mean, this wouldn’t do for a real site but rather for a Proof of concept
Steve says: ah, yeah
JB says: Since controlling other sites like that is kind of evil.
Steve says: spoofing and stealing content… yeah
JB says: but just to show “here’s how it’d work if we all agreed to do it”.
Hmm. Or this is better: other Apache servers could recognize requests for pathway steps. That’s a better proof of concept. You get to leave one site, go to another, but it recognizes a special request and formats appropriately.
Steve says: Yeah, that would do it. Maybe a combination of a web browser plugin and an apache module. The web browser plugin would put something extra in the HTTP header which the apache module would recognize
JB says: Well, it could be a new protocol request couldn’t it? One that’s almost http but with semantic additions for the pathway built into the headers of the request.
Steve says: Yeah, thought about that too, but http is so entrenched
JB says: Well, yeah, i think this would be a whole grassroots sort of exercise.
Steve says: And then you’re talking W3C RFCs and such
JB says: And i don’t think this’ll ever actually get built, so I’m just thinking about what would a good way to do it that didn’t require everyone to have a browser plugin. ‘Cause shit, even Flash isn’t in *every* browser yet
Steve says: right
JB says: But there’s only two web servers to speak of. Apache and IIS. I bet the semantic web people have already done all of this thinking, but i also bet a new protocol is the trick
Steve says: maybe a different apache virtual host
JB says: yeah, kind of like how SSL is done
Steve says: So, rather than linking to http://www.hogswallowing.com/page.html you’d link to webscursion.hogswallowing.com/page.html
JB says: or instead of http:// you do sttp://
Steve says: A new protocol would certainly do the trick. That’d probably require a web browser plugin, though too.
JB says: really? why?
Steve says: For the browser to know where to direct sttp:// — what port
JB says: hmm, true
Steve says: or, make port 8222 the one used for this and run two instances of apache, the one at 8222 modified to do the webscursion, and link to http://www.hogswallowing.com:8222/page.html (or whatever port)
JB says: yeah, either your destination supports webscursions or not. Any way you link to it, it has to support them, so it’s finding the way to accomplish the link without burdening the reader/user that is key.
Steve says: right
JB says: It would read the page, find the next link in the path, and format appropriately. You would have to figure out “what do you do if links disappear”. On receiving a 404 does it go to the next step on the path? Does it mention anything? Put some kind of coding in the request headers to indicate the health of the path and let the browser or scripting interpret it, and include reference to the “Originator” of the path to allow authors to monitor the health of their paths and make adjustments when necessary.
Steve says: gets complicated
JB says: yeah, but it started complicated
Steve says: that is true