Chapter Two

Proposal For A Universal Electronic Publishing System And Archive

This chapter is ”the heart of this book.“

2.1. An Electronic Literary System

On a computer screen one should see the true structure of text, of information.

We are describing here how to represent and store this true structure, called literature.

Virtuality is the design of interactive systems. One starts with the intended user's experience, after that one begins programming.

2.2. What Is Literature?

Literature (in broad sense: belles lettres and scientific literature) is ”a system of interconnected writings.“

For example in science there are implicit and explicit links in literature (references, quotations).

LITERATURE IS DEBUGGED.

By that I mean that the system of writing and publication is a well-worked-out method and structure with deep and subtle workings. [p. 2/11]

Perspective and fashion are ever-changing but the worked up mechanisms of written text endure.

In the world of business the linkage structure is very similar. Documents contain references to other documents as ”In reply to your letter of the 13th....“

Because links exist in all kind of writings, we have endeavored to create a system for editing and retrieving documents with links, ”to mirror, and replicate, and extend, existing literary structure.“

2.3. A True Storage System For Text And Other Evolving Structures

Apart from backup copies for safety reasons, which will become soon automated, one copies the work regularly for backtracking, too. It is superfluous to store each time the full material, since one has done just small changes.

A file management system should store every change and keep track of it. One should be able to get a previous version of each fragment and see its evolution.

This assumes you are reading on the screen, not on paper. We call pounce the effect of constructing the view when the user calls it.

The file storage system can be extended to handle alternative versions with variations of the document. It should be possible to compare versions word for word.

This storage system works not only for text but for any forms of data structure. Example: architecture, airplane design.

Computer Lib proposes the term ”thinkertoy“ for a computer screen facility for detailed comparison of complex alternatives.

The unified structure containing all versions (history, alternatives) may be called ”docuplex.“ The storage system for documents with parts, versions and links among them may be called ”hyperfile.“

2.4. A Linking System For Text And Other Data

A link is a connection one puts between parts of text or other material as pathways for the reader's exploration. They are part of the writing.

Link types:

Jump Link. The writer can set a jump opportunity, similar to an asterisk of a conventional footnote. The reader points to it and the related place appears on the screen, one can jump back.

Marginal notes, Side-by-side writing. Marginal notes (not necessarily rendered on the margins) are a particular kind of side-by-side connections. One can set links between passages of two texts and when reading one of them one can recall the other one.

This can be used for ”virtual yellow stickers,“ electronic counterpart of Post-It Notes etc.

Hypertext is fully non-sequential writing based on links.

A user should have the possibility to set any types of link. Links are point-to-point, point-to-span or span-to-span. A link can have multiple endpoints. Links can be directional or not, depending on the type of link.

The Front-End is responsible for showing the links in an adequate manner, for keeping track of the places the reader visits and remembering the favorite ones.

Having links presupposes having historical backtrack and alternative versions. The source one has linked to may be updated by its author and the link must still point to the right place. One needs to compare alternative versions in depth.

The jump-link leads to all sorts of new text forms (scholarship, teaching, fiction) and facilitates serendipitous browsing.

Links can be set between all data types, not just text but also music, pictures, etc., including links themselves.

2.5. The Document Convention

If anybody could change anywhere anything, the whole would end up as a chaotic blur.

The solution is to have separate documents, as traditionally done in literature. Somebody designates a unit consisting of text, graphics, links and window-links to be a document. This unit is proposed as replacement of the file.

Everything in the system is a document and has an owner. The owner is the rightful copyright holder or someone that has permission from the copyright holder and pays for storage. The owner is the only one who can remove or change the document.

A link between or within documents resides in one document. It is part of the document that creates it and has an owner. Ordinarily a document consists of contents and out-links pointing to other documents, these being under control of the document's owner, whereas in-links are set elsewhere and owned by someone else. Other cases are possible, such as links between documents residing in yet others.

2.6. Compound Windowing Documents

A document can quote as much as desired from another document (text, pictures, etc.), including it in a quote-window or quote-link. A window of a windowing document is itself a link, not a copy, and does not alter the ownership of the original document.

A compound document is a regular document and the ownership conventions apply to it, too. It can again be quoted in another document.

A reader might ”step through the window“ to explore the origin of a quotation. Then the reader can go back to the first document or follow other paths.

Examples. Annual report of a corporation with a brief paragraph about each division quoted from other, more detailed documents. Children's story with pictures being part of larger pictures, each with its own story attached. A scholar interpretation of ancient Greek society quoting original writings. In all these examples the user can navigate through the inclusions and explore the source documents at will.

Advantages of this method of storage: it saves storage space and there are no copy operations, thus avoiding update problems.

There are two types of inclusion links: a) floating in time, b) fixed in time. The first one is seen by the reader as the latest version, the second one as a particular one set by the writer. There are front-end functions available to switch between these.

Links are two-way: the reader can see documents that include the current document, too.

Derivative documents are a particular form of compound document consisting of an original document and changes to it. The original and the derivative document can have different authors.

Alternative versions of a document can be published by the owner as part of the same document (versioning by descent), or by the owner or somebody else as a separate document (versioning by inclusion).

An intercomparison document consists of relations among other documents.

Inclusions may nest. Links may be quoted.

We have defined here an interesting world with rich document relations and a simple design that relieves us from the complications of conventional computing.

2.7. Electronic Publishing: Making The Literary System Universal

The storage and delivering of compound documents and hypertexts can be extended into a full publishing system, provided it includes provisions for privacy, copyright, royalty and accounting.

A document is by default private. Publishing it must be a conscious act, probably signing a contract. Publishing is irreversible and allows every reader to link to the document. No withdrawal is possible for a published document. But the owner can allow unrestricted distribution of a private document (privashing) and remove it later on.

Royalties are automatically payed to the document's owner. A fix amount, say a thousandth of a cent per byte.

Incoming links from other documents can be asked for, filtering (sieving) by some criteria.

Books, magazines etc. remain existing and relevant. They are not being substituted but enhanced with this new system.

Categories are not system level but user directories (documents, too).

2.8. Distribution And Networking

The system must be distributed in multiple computers around the nation and the world to be able to grow without size limit. The data is stored at multiple places but the response for the user is immediate (some seconds).

2.9. Vital Issues

Vital issues are user privacy, freedom of the press, legal good behavior.

The interconnections are as important as the words; and following the continuing variations and re-uses demands a simple but subtle system for instantaneous access to any part of any document and any possible connection between documents, and for the unending creation of new variants and connections, forever. In the light of these concerns we offer this system, and believe it to be the vital next step for humanity and civilization.