[Gnu-arch-users] [FOSDEM Substitue] Loss, Corruption, Thievery, Confusion and Constraint (A Global Hypertext Strawman and Its Problems)

[Thomas Lord]

Since I won't be at FOSDEM I'm not preparing a talk or slides.  I am,
however, writing a brief series of short essays about the topic I
meant to speak about.

This is the fourth essay in the series.  The "cliff's notes" so far are:


1. The King's English: What Do Links Point At?

   Today's links point at property.

   We don't really have hypertext.   All our links point to
   Internet real estate, not to documents.


2. The Literature Shelf is Not Literature

   True names would enable real hypertext.

   True names don't really exist but can be approximated.


3. Who Owns the Author?

   Authorship is the right to modulate a document's contents.

   Documents require access protection.  The right to modify
   can be shared, transferred, and subdivided.


4. "The Document" is Just a Hypothesis

  Documents are never tangible -- only signals about documents are
  "real".  We can possess evidence of a document but not the document
  itself.

  Software systems for hypertext should be about managing signals --
  not storing documents.


5. Loss, Corruption, Thievery, Confusion and Constraint

  Usenet counts as a global hypertext strawman, but points
  up some problems to solve.


-----------------



         Loss, Corruption, Thievery, Confusion and Constraint
             A Global Hypertext Straw-man and Its Problems


  If you squint your eyes a little bit, it looks like global 
  hypertext was more or less invented around 1979 by Tom Truscott
  and Jim Ellis in the form of (what would become) Usenet news.
  Let me describe a bit how to *use* Usenet news as a global 
  hypertext -- I'm not claiming that all uses count directly as
  global hypertext.

  (Earlier examples of non-distributed but still group hypertext, such
  as Plato can also be found.  Of course, the idea and prototype goes
  back even further, with the like of Douglas Englebart and Vannevar
  Bush -- other materials cover the history better than I can.)


* Straw-man

** A Usenet Author

  Anyone who can transmit NNTP to a Usenet host, including at least
  the `POST' protocol, can be a hypertext author.

  An author's "frequency" -- their unique ID -- is simply their user
  name.  Of course, people are free to share user names and subdivide
  the rights to make posts under a given name.


** A Usenet Document

  We can assert that, by convention, a document is created when a 
  given author (name) posts a message with a subject of the form:

        Subject: [DOC ___(ID)___] ____(title)____

  The ID can be filled in with anything but should remain constant
  throughout the lifetime of the document.   The title can be filled
  with anything and can be freely changed over time .... the title
  is just a convenience.

  Each message of that form is a signal -- a snapshot of the contents
  of the document at some point in time.   Other header fields, such
  as the time-stamp, author name, and message id, help to distinguish
  a particular transmission from all others.

  Of course the body of the message is the contents of the document
  in that snapshot.   It's free to use mime, rich-text, etc.   

  The body can contain location-independent hyperlinks such as:

        <sender="address@hidden" id="xyzzy">joe's cool document</>

  or, to refer to a particular snapshot:

        <sender="address@hidden" id="xyzzy" msg-id="2347234...">
           the best version of joe's cool document
        </>

  Of course, we presume clients that recognize and handle this style
  of link.

  A real world example that comes very close to this proposal are the
  various FAQ posts on various news groups.  The "comp.lang.c FAQ" is
  a living, referencable document with a controlled authorship.


** Distribution

  Usenet distributes articles on a simple P2P network.  If someone
  gives me a link and says "you should take a look at msg-id
  "2347234..." I don't have to necessarily go to `google.com': 
  I can find that at the closest, most convenient, possibly 
  off-line place.


** Half a Wiki

  It's worth noting a convenient feature built-in to this mapping
  of hypertext to Usenet:  it's trivial for people to post to 
  discussion threads *about* any document -- just hit the `reply'
  button.

  If a wiki is "open editing" + "discussion" we're half way there.

** Full Wiki

  A slight variation let's *anyone* post with a subject like:

        Subject: [WDOC ___(ID)___] ____(title)____

  and allows links like:

        <id="xyzzy">Usenetpedia article on T. Jefferson</>

  The key point in mentioning that these wiki features "come for
  cheap" is that the structuring -- the relationship between 
  competing edits and commentary and authorship -- is all implicit
  in the content of the snapshots of the documents.   There is
  no obvious need for a central server controlling an authoritative
  version of a document.


* Problems

** Loss

  The Usenet system as it stands makes no formal provision for
  long-term archival of valuable materials.   It would be a mistake,
  for example, to run Wikipedia on Usenet because then infrequently
  updated articles would too quickly expire and become forgotten.

** Corruption

  The Usenet system includes no built-in integrity checks to make
  sure that saved and transmitted articles are faithful copies.
  It relies on lower-level transport and storage for that.

** Thievery

  As anyone who has had a popular Usenet post forged under their
  name knows, the integrity of authorship on UUNET is weak.

** Confusion

  Are we looking at the latest version of that "[DOC...]" or
  "[FAQ...]"?   What are we to make of it when conflicting, apparently
  roughly concurrent versions appear from different sources?

  These problems may ultimately be unresolvable but Usenet has no
  internal mechanism for at least managing the confusion.

** Constraint

  With length and format limitations, Usenet imposes rather 
  severe constraints on document format.   Sure we could split 
  a large or complex document into many messages, but then we
  are talking about a lot of infrastructure built *over* not *of*
  Usenet.

-t