[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Bug-gnupedia] Publishing processing solution
From: |
Peter Verrey |
Subject: |
[Bug-gnupedia] Publishing processing solution |
Date: |
Thu, 18 Jan 2001 16:57:03 -0800 |
After reading the posts from the past two days, here are some suggestions.
Remember, if we want a relatively easy-to-maintain system, we need to
capsulize
our ideas.
Here is a conveyor belt system for dealing with article submissions:
1. Article is created by author at some point.
2. Article is converted to one of many possible suggested formats for
submission,
like LaTeX, Html. Each format has its own guidelines to be an
appropriate
submission.
3. The authority responsible for first putting the article on the web
receives the article.
If the article meets the syntactic guidelines for submission (proper
title, author,
working links,etc) and as long as it meets a VERY basic semantic ruleset
(not immediately and obviously spam) then the article will be
'published'.
4. The publishing authority converts the submitted article into an
appropriate storage
format, like XML, LaTeX, (perhaps PDF), etc.
5. At the end of each day, the publishing authority puts all the articles
selected to be
published that day on their online server, alerting all mirrors that new
articles are
ready to be distributed. Each article has its own unique identifier,
like the title
followed by the publish date, and possibly the format if it is stored in
multiple
ways.
6. Category authorities look at the articles and determine their place in
the Dewey
Decimal and other categorical systems. I suggest that each category
type should
have its own authority. I'll give my reason in just a second.
A proper category could be granted in one of three possible ways (though
I am not
suggesting each way or just one way should be implemented).
A) The author notifies the category authority that they want to be
included. The
author may or may not recommend one or more subjects to be placed
under.
B) It is up to the authority to notice the new article is published and
ready to be
'indexed'.
C) The published document may include instructions on how it should be
categorised, either put there by the author, or by the publishing
authority.
Now, as to why each filing system should have its own authority, I think
it would
be great for backwards compatibility. There are many many different
category
systems in the world, some very popular, some for specific fields, and
this allows
all of them to operate at the same time, and be added at any point,
without
throwing off the system of publishing. If you want your own category
system to
be compatible with the Yahoo directory tree, then go for it.
7. Finally, review authorities can look at the articles. This can work in
many
of the same ways as the category authorities. There should not just be
one, there
should be one for every key concept. The author may wish to send it to
a
specific reviewing authority, or some authorities may simply review only
articles
that are in a specific subject in a category authority's system, such as
a reviewing
authority involved specifically in physics subjects only. Ultimately,
most important
articles could be handed to a fact-checking authority. A certain
category authority
may require that an article go through a peer review authority before
being added.
Regardless, the fact is that the article would still be 'published' by
the publishing authority,
regardless of quality, but if it really IS bad, or factually incorrect,
it will simply be
unendorsed.
Ultimately, I think a system like this would provide fairness to authors who
feel they might
be unjustly treated. There is always an opportunity to get another review
by different
authorities.
This also clearly delineates in what role someone can become involved in,
all of the roles
mentioned above, and the programmers necessary to maintain the system in
which it works,
from format conversion to search engines based on the different categorical
sytems.
I realize there are important things in the system I have not covered, like
revisions.
As a starting point, the Dublin system is very promising. Revisions should
be handled
like Accounting, in that something should never be made to 'not have
happened'
(unless you enjoy gooking up the books).
There can only be new articles which make note of the fact that they are
designed to replace
older articles, with proper linking and such, not overwritten replacements.
I hope this helps to provide a clearer direction, and I welcome any
arguments
against or revisions to this system.
Best,
Peter Verrey