Jack Park
Street address
Palo Alto, CA Zip

Date: Tue, 23 Jan 2001 18:29:23 -0500

From:   Jack Park
Reply-To: unrev-II@egroups.com

To:     unrev-II@egroups.com

Subject:   OHS Meeting at SRI on January 22, 2001
Demonstrate BrowseUp

[Commenting on the demonstration of BrowseUp at SRI on January 22, 2001, to support the OHS/DKR project....]

Here is my take:

BrowseUp is, indeed, very innovative and well done. I hesitate to use the word 'innovative' because much of what BrowseUp does has been anticipated in papers by Douglas Engelbart, Ted Nelson, and Francis Heylighen, among others. Nevertheless, as an execution of the many ideas, BrowseUp is really slick.

I formulated a model of how it works. Right or wrong, here it is.

Imagine that your browser is set to work through a proxy server, a tiny local server that, itself, does the web connection for the browser. WBI, IBM's transcoding engine, is one such server.

Because your browser does not directly interact with the web, the proxy server has the opportunity to look at the URL you have requested and feed that URL to another web connection, which happens to be BrowseUp's link server. The link server can download what it knows about the selected URL while the selected URL, itself, is coming in on another http connection. Now, two bodies of html information are available in the proxy server. Before the server sends that to the requesting browser, it can perform whatever computation it likes.

BrowseUp appears to add a tiny bit of html to the page before being displayed. That added html forms an href link such that, should you click on it, you now go directly (through the proxy server, of course) to some URL inside the link server, where another window opens complete with all links others have established with the link you just clicked on. Got that?

There's more. Suppose the proxy server could open a tiny dialog of its own such that you can reach up into your browser image and grab something and drag it into the new dialog. That establishes a target. Now, go to some other web page and click on something and, presto, or words to that effect, the proxy server opens a nifty display of some linkages you are about to make. Both directions are linked, but you can 'uncheck' a box at either to break a link. Meanwhile, you can annotate the link(s), complete with search words and so forth. Got that?

So, now, you have imagined a really nifty kind of engine that gets awfully close to a transclusion engine as described by Ted Nelson. The only difference is that BrowseUp does not 'transclude' (meaning, actually insert the referenced material into the page being displayed). Rather, it gives you the equivalent of a menu to select those links you might want to browse.

Now, that's powerful, in my extremely humble opinion. So powerful, however, that I raise a couple of personal opinions (hip shots!) for further discussion. Note, these opinions actually apply to just about any NIC one might build.

I am talking to the so-called 'web of trust' concept advanced by Tim Berners-Lee in his Semantic Web initiative. We all need to trust each other to 'do the right thing' (whatever that is). And, BrowseUp opens pandora's proverbial box to all sorts of not-so-right things one could do. Imagine, for instance, someone linking your home page heading to, say, a really grotesque gif or jpg.

Here, I am thinking that it may be that establishing links ought to be a priviliged operation. Only those who are authenticated and have permission to do so should, perhaps, be allowed to do so. I am thinking that if everyone on earth had the ability to slam links onto whatever they want, there would be hell to pay.

But, I am not saying that BrowseUp, or even it's eventual clones, whatever, is without merit. On corporate intranets, you already (theoretically speaking) have a web of trust. On networked improvement communities (NICs), the opportunity, if not requirement, exists to authenticate those who participate. No, I'm not talking about private exclusive NICs; anybody can join, but they must be authentic, and tracable, because the links can be traced, through logs, to individuals, and that's probably the way things should be.

Moving away from the web of trust thing, consider legal implications. What are the laws regarding linking (especially, willy-nilly linking). It is my understanding that eBay got an injunction against a dotCom that was either linking or transcluding auction information at the dotCom's web site. I recall (maybe with imperfect memory) phrases like 'deep linking' (Google got 224,000 hits on that one). In fact, the second hit was this:


....which just happens to deal with the notion of deep linking.

Here is a quote from the wired.com article cited at the deepLinking url just cited: "Legal experts did comment, however, saying the legal landscape surrounding deep linking, or hyperlinking deep into another's Web page, is fraught with unpaved ways."

There you have it. Due Dilligence, here, would suggest that, before any NIC goes live, particularly one that permits linking around the web, some deep research ought to be done on issues such as those raised here.

Well, that's my 0.02 EURs for the day.



Jack Park