FrontPage SiteMap RecentChanges RecentNearChanges HowTo Blog RSS

LowTechSharedWebBrowser

A “Shared Web Browser” is a web browser where everyone using it sees the same thing.

For example, AlexSchroeder may point it at a particular page on CommunityWiki, and then everyone else who he was talking with at the time would see the same page at the same time.

Then someone else could take a turn, and point out another web page.

People could leave it pointed at a page that they thought was generally interesting, for other people to check out.

Discussion

I think we could implement this right away, with nothing more than just Frames and JavaScript.

If we have a special server, we might be able to have something that’s super-responsive. I’ve seen JavaScript messages passed in real-time using some super “Hold the HTTP connection open” technique- IntComm:ModPubSub, if I recall correctly.

Say we have a single wiki page somewhere, with a single link on it. (Perhaps the “DatePages” 2004-06-22 2004-06-24 2004-06-26 2004-06-27 2004-06-28 2004-06-29 2004-06-30 2004-07-01 2004-07-02 would be appropriate, or perhaps the Community Wiki Blog http://communitywiki.taoriver.net/ ). When it’s my turn, I delete the old link and add

“I/we are currently looking at WikiFeatures:FailSafeWikiDavidCary.

Then everyone knows I’m looking at that page. Later someone else could delete that line, and point out a different web page.

  • As a wiki page: people have to manually click the “back” button, then click on the new latest link.
  • With frames and JavaScript: page updated automatically, zero-clicks.

What’s your goal here ? Does the first (simplest) option meet that goal ? Is this just an improved SharedAwarenessSystem, or is there something I’m missing ?

DavidCary

This is a “just a form” of SharedAwarenessSystem, for use in live situations.

But SmallChangesYieldGreatChanges.

Now, right now, we have live sessions in the WikiChannel (InternetRelayChat.)

(You should join us sometime!)

In IRC, rather than messing around with wiki, it’s just easier to say a URL, and have people click on it.

This idea is about automating the lookup step, so that everybody is automatically on the same page, so to speak (and literally!)

Here’s one small benefit:

  • Participants don’t have to click on anything, pages just appear.

Here’s two large benefits:

  • For the first time, everyone knows that everyone is looking at the same thing, at the same time.
  • “Social Surfing,” for lack of a better phrase. People will just pick up the shared browser, and go with it, or leave the page on something that they thought was interesting.

I know about Social Surfing, because we test trialed WebTV at my dorm when I was in college, and we had great fun, just ocassionally looking over to see what other people were looking at, and thinking about. It’s almost like the old scenario: “What if people had little thought bubbles above their heads?”

And it’s- you just cannot do that in an IRC channel alone.

Nobody starts just dumping links into the channel, walking from site to site, and expects people to watch as they go around. People would call that behavior crazy, and likely ask that person to leave. I don’t know- it’s just sort of unthinkable to do that.

But if you have shared web browser, people will do just that.

Now, the LowTechSharedWebBrowser described in the DocumentMode here- it has some problems. For instance, if you click on a link, unless you have some sort of JavaScript intercept system of link clicks, (I forget if that is possible,) then whoever makes the click is going to have to copy the URL into the form that feeds the group communication.

But, you know. It’s half-way there.

There may be some software for shared web browsing out there, already. I’m going to go investigate.

Well, there’s a Shared Web Browser in Lyceum, which is not distributed or sold or released on the web.

For the Low-Tech Shared Web Browser:

I’ve realized that- if you pre-parse links, you can modify ever A HREF=”…” link to redirect through the main web browser. That way, people can just click on links like normal, and they automatically work through the shared system.

So- when the main browser fetches and displays a web page, it acts like a proxy. It receives the page itself, transforms it to make all links go through itself, and then posts the transformed document.

This will not work unless there is fast simultaneous voice communication (skype and teamspeak I tried out with thomaskalka working simultaneous on a webpage, both work excellent). Voice is only used for fast information, not for chatting. “Better use verb a insted of verb b.” “Ok. Let me edit now.” Or on a common browser: “When everybody is ready with reading I’ll go to page B”. Just a sentence every 20 seconds - not more. Voice is secondary. We tried it out. It improves working speed x 100. Believe me - it’s good for you. We did it half an hour twice - Thomas used spoken language as one instument of two - in fact as the less important one - from beginnng on, very cool. Mega-innovative. Common browsing: a great idea - but no way without a simultaneous and disciplined voice communication, according to me.

Just saw Jybe. ;)

By using VncServer? it is possible to have a SharedDesktop?, which allows to use any program. Wich other benefits do you see demanding seperate “shared apps”, if there are missing special Multiuser-Features (as many text-pointers, like MoonEdit) ?

I’ve used VNC a few times, and pointed to it several times on this wiki.

I don’t believe it’s the real solution, as much as I hoped it would be. I’m still trying to pin down the reason it doesn’t work for me; I think it has something to do with the comfort of using my own desktop, the ability to understand what things are for what, and- I mean, just practically, if you need to answer an instant message from someone who doesn’t know that you’re in the middle of something- you know: It just pops up, and everyone else must necessarily be like: “Wha? Wha? What’s this?” You know: Now they have twice as much instant messages to deal with- their own, and the messages of the person who they’re VNCing too. Maybe I should call this “too much information,” in both the privacy sense, and in the InformationOverload sense.

I think what we really need is a GroupUserInterface that everyone in the group can participate in manipulating. The UserInterface itself is a ConsensusReality?.

This way, when people say, “as you can see to the right,” everyone knows what you’re talking about. Just stuff like that. Every time you have to think about the difference between their perception of the UI, and your own, it’s a distraction, and it breaks communication.

Honestly, I’m still trying to put my finger on why it is that it doesn’t work for me. I hoped it would. But it doesn’t.

I can be swayed; If I see people using VNC effectively to communicate rapidly, and if that becomes something of a norm, that’d be great. I’d just adapt and start singing.

One thing that you could do, that I’ve been thinking about quite a bit, is make a space for virtual VNC servers. ;) That is, there’s nobody with this desktop, really. It’s a fantasy desktop, with its own UserInterface and stuff. It would be equipped with a standard set of communications tools, and that’s it. Everyone who wanted to share a conversation would go to the fantasy desktop, made explicitely for communicating, and share this space.

I would have zero objections to that.

Developing this other software is still important, though, because we need ActivityAwareness in general. You don’t get that by participating in a fantasy desktop that forgets who you are.

Here’s another thing. How do you have a chat in VNC? You have to keep writing your name, over and over, into the text box. There’s just all these little things.

I was thinking about something like adding a “sharing layer” to any app. There seem to be some solutions (for example http://w2.alkit.se/avnc/) based on VNC. I’ve tried to set up a public vnc-server, but did not find a solution how to secure it. Last idea was to run it inside usermode linux and to start a fresh virtual linux box for a session.

That’s a great idea! That wouldn’t interfere in any of the ways I described; it’d be amazing.

I’m going to be thinking about this for a while.

Maybe it is enough, for someone with a good Linux machine and a good user policy, to start a vnc server on a second X server (passing :1 as argument to X) started with a restricted user. For each session, you just have to remove and restore a clean $HOME dir from a tar.gz backup.

Another possibility would be to use http://w2.alkit.se/avnc/ to share an Xnest window again started when logged as a restricted user.

Does anybody know a tool to check the “restrictedness” of a user ? Up to now I am not fammiliar in setting up publicly available user-spaces.

I’m returning to this toppic and collect my thoughts on PublicSharedWorkstation.

EditNearLinks: MeatballWiki PatrickAnderson UserInterface JavaScript DocumentMode

Languages: