| Author: | Benja <b.fallenstein@gmx.de> |
|---|
Several things, in my opinion.
The Web is a comprehensive digital library of everything published... last month. Going back only a few years, when you look at e-mail archives from 2000, many of the HTTP links have ceased to work.
How are we supposed to "stand on the shoulders of giants" if more of these shoulders slip out of existence every day?
Point in case: I'm working on a project that has had two trademark-induced name changes since then. Our old web page, http://gzigzag.org, has in the meantime been acquired by a domain grabber and turned into an advertisement for "teen porn." Sure, the homepage is accessible through the Internet Archive, but even that only takes you so far.
The problem can be alleviated inside the existing infrastructure, as Tim BL points out, but this means that pages will only be preserved if there's some big organization interested enough in them to put up the servers, do the paperwork, acquire the domains, etc. That interesting personal home page with notes about taking the train to Morocco which I linked to in my private diary will have gone "out of print" when I look at the diary again in five years.
My stance is that a web page should remain accessible as long as there is somebody who think it's worth keeping a copy around on their own hard disk, and not keeping it private. Links to it should work without using some special "archive" command or web site. With p2p and location-independent identifiers, that's possible.
Yes, there's file:, but there's no way to make a link to a file that I have and that you have-- but in different places of our respective file systems.
I want to be able to make links between desktop documents that continue to work when the documents are--
Web linking is used on the Web, but not on the desktop. I believe that one reason is the above-- when you make a link, it only works on your own desktop, not if you publish the document or send it by mail.
When I send a mail with a link, the link should work if the receiver can find a copy of the linked to document, no matter whether it's--
A related issue is linking to e-mail; I should be able to link from one e-mail to another so that the receiver can follow the link if they either have the target e-mail in their own inbox, or if it is archived somewhere on the Web.
This needs location-independent URIs; P2P makes these resolvable on the public Web.
We need something better than the web to build more advanced hypermedia structures on top of it.
As a simple example, why doesn't the Web show which other pages link to the page you are viewing? In the client/server architecture, there is no way to do this globally. For some thoughts about this from the dawntime of the Web, see Building back-links by TBL.
More ambitiously, we are implementing Xanalogical storage, hypertext as envisioned by the group of Ted Nelson, who introduced the word in 1965. In Xanalogical storage, each character you type (or pixel you photograph, etc.) has a global identity number. When you copy some characters to a different document, they retain their identities. Links aren't between documents, they are externally attached to characters (and thus to the documents where these characters appear). This allows you to--
This system needs two lookup primitives:
To do this in a decentral, scalable, and reliable way, a p2p lookup scheme is necessary.
Not everybody may agree that the above are reason enough to replace the current Web, but they are concerns that are important enough in my own use of the Web that I would be happy about having a replacement.
- Benja