I was thinking today about how to create a future Internet by using something like Xanadu Hypertext as an overlay network.
Xanadu is the name of Ted Nelson's 1960 invention of the Xanadu Hypertext system. Ted's a great guy, and he foresaw most of what would be needed for a Memex-like Knowledge Navigator. The World Wide Web is a drastically watered-down subset of Ted's original idea.
I realized that the page moving or document renaming problems in the
underlying internet are still a problem. I don't see anything
addressing it that could be deployed in the forseable future. So, why
not have my Xanadu bi-directional transclusion links use a
content-addressable link like IPFS document addresses. I don't know
how/if you can address a bookmark inside an IPFS document, which could
be in any document format. Perhaps search and format conversion would
be required.
I'm envisioning a document creation, editing, and reading tool that
has Xanadu's bi-directional links, micropayments, and a document
format that the Xanadu network understands. Hmmm, I'm starting to
think that this should be an IPFS-native application which can operate
over the overlay network to reach anything on the public internet.
So, imagine a full-featured document editor such as LibreOffice
Writer, or perhaps Microsoft Word. It has the ability to embed
sections of other documents via rich bi-directional links. It doesn't
work with ordinary filesystems on storage disks. It's 100% IPFS file
access (with content-addressability) all the time. The editor doesn't
create "files." It creates an IPFS filesystem in a workspace file or
connects to a MyXanadu server, which can be local or remote. This
editor can import and export files in the most-used document formats.
I don't see why not to set this up as a project. I think the primary
client document editor should do something more reasonable than the
current document representation file types. Perhaps text and structure
are represented in a markup language and displaying or converting a
document is delegated to Latex or similar. Hmmm, that implies being able
to edit a Latex document in a WYSIWYG fashion. That's not a trivial idea.
Since I need a new document editing application, why not support simultaneous
multiple-user editing via CRDTs. That's a bit trendy in some circles.
Pros:
All of this could be an infrastructure with a fully-documented protocol, to allow for the development of alternative client software. I think that mind maps and other outline-adjacent thought aids could be included.
Cons:
WYSIWYG editing of a formatted, camera-ready, final form
document. It seems overblown to create a new editable-in-final form,
so I'd like to just have final.
Dependency on IPFS features. It would be nice to work on any overlay
network.