Eugene Eric Kim
PMB 207
63 Bovet Road
San Mateo, CA 94402

Date: Thu, 25 Jan 2001 13:24:09 -0800 (PST)

From:   Eugene Eric Kim


Subject:   OHS Meeting at SRI on 010122 Demonstrate BrowseUp

On Tue, 23 Jan 2001, Eric Armstrong wrote....

As I recall, "we" (as a group) disliked that notion, at the time, because we wanted people to use a "vanilla browser", without having to do anything special to the client.

I take it that our collective opinion is now more in favor of a smarter client?

How much effort is it to install that proxy, anyway? Or is it pretty transparent? (Or does a remote server actually serve as the "proxy", so that you have to visit that server and, from there, go to other points on the Web?

As Jack pointed out [in his letter on January 24, 2001] using the term "proxy server" in his original description was misleading, although there's no reason why the system couldn't be implemented that way. At the risk of repeating what others have already stated, here's my summary, along with some brief commentary.

BrowseUp allows you to create links from any HTML page to any other HTML page. It consists of:

The links are granular, and optionally bidirectional.

You can also annotate the links.

The link server stores all of this information, as well as thumbnails of the information being linked, and the "quality" of the links. Link quality is an attribute used for displaying available links, that is based on the frequency a link is viewed and presumably other information as well.

The link server is written in Java, and apparently communicates on top of HTTP. They were apparently designed to be widely distributed, so that these link servers could be found all over the Net. BrowseUp has some solution for problems such as link integrity and modified documents, although I think I may have missed that part of the discussion.

There was some discussion about social consequences of unrestricted linking that was interesting and important, but I won't rehash them here.

Some thoughts:

  1. The link server is, in my opinion, of greatest immediate interest to the project. Many of its requirements seem to overlap with ours, with some limitations and quirks. For example, the thumbnails are an interesting feature, but I wonder if the bandwidth and storage overhead is worth it.

  2. The plug-in is a win, but it's not necessarily the short-term direction that the core project should move in. The link server would enable us to move in a number of directions, from browser plug-ins to proxy servers to custom clients. There's a lot of room for evolution here. At the meeting, Lee said that he would like to see a Nelson-style transclusion view of links and documents (where documents are displayed side by side, with lines showing the interconnected links). I see no reason why such a view is not possible using BrowseUp's underlying system; they just haven't built it yet.

  3. I think collaboration is highly desirable, and a reasonable possibility. Alon seemed genuinely interested in seeing link databases pop up all over the world, and seemed open to the possibility of open sourcing parts or all of their system in order to enable this (and collaborate with us).

What are the next steps?

I know Doug is pursuing the collaboration angle, but there are certainly things that we can do in the meantime.

The best is to just try using the tool. I think there are many things we can learn about some of its fundamental linking features and about BrowseUp's own implementation of these features, which could lead to:

The plug-in is available from...

... and I'm sure Alon and company would welcome feedback.


-- +=== Eugene Eric Kim ===== ===== +=== Eugene Eric Kim ===== ===== ===+ | "Writer's block is a fancy term made up by whiners so they | +===== can have an excuse to drink alcohol." --Steve Martin ===========+

eGroups Sponsor

Community email addresses: Post message: Subscribe: Unsubscribe: List owner:

Shortcut URL to this page:


Copy to:


Post Script