This is a proposal for the use of tags as the main document classification mechanism
This is a proposal for the use of tags (rather than a hierarchical containment tree) to classify documents. Under this proposal, the collection of documents are simply a linked graph, and we can have many Containment orders by simply maintaining index pages.
Formally speaking, a Containment is simply a partial spanning trees defined by index pages, which is how we'd have done it anyway. The thrust here is that Containments are too restrictive, since an article may fit, for example, into both the "Climber" and "Ribbon" subsystems (non exclusivity) and that in different scenarios, different orders of containment makes sense (e.g. subsystem first, or status, etc.)
Using tags eliminates both these problems.
Implementation and Presentation
The simplest "in-band" mechanism to implement tags is based on an Infobox-like templates.
The tags, along with other information such as title, cover image, author, date (and perhaps an article abstract) are displayed in an Infobox on the top right.
We can potentially use page formatting templates to make baseline tags actually appear different, though this should be used sparsely, maybe only for the status tag.
Privileged (Administrative) Tags
Most tags are passed to the system through an infobox instantiation box.
Privilaged tags, in cotrast, are set by the powers that be. In this proposed implementation, they are enforced by having the infobox control them based on the article name. A good example are status tags (Baseline/Peer Reviewed/..) since we do not want the author controlling them.
This means the Infobox has access to the list of article names that are "Baseline", and present the status line accordingly. The template for the infobox has to be protected.
Once articles are tagged, several auto navigation tools become trivial
Automatic Index Pages
As part of the navigation system, tags should be associated with auto-generated index pages. Already on the infobox, tags will be linked to their index pages, so that if you're looking at an article marked "politics" you can go to the "politics" index.
Perhaps the strongest tool, a filter takes a subset of tag values and returns a list of relevant articles. (perhaps with abstracts, if the list is short enough)
For example, the user interaction might look like this:
- "There are 2414 articles currently in the DB"
- Select tags: Peer Reviewed, Ribbon, Deployment
- "There are 132 articles that match these tags"
- Select date
- Select author
- "There are 4 articles that match your query, listed below…"
The filter page should be accessible from the infobox, perhaps from a "Tags" link about the tag list.
- First thing's first, the system has to be tried out statically on the first 100 articles or so, to see that it works out.
- Write the infobox, if that's indeed the mechanism we use.
- Need to work out the mechanisms for the filter and auto-index pages.
List of Tags
The tags below are grouped into "somewhat exclusive" groups - for example, usually an article is exactly one of engineering/legal/humor/art ... but I can think of exceptions...
- Baseline (Has been accepted into the baseline group of articles)
- Peer Reviewed (Has been peer reviewed and has comments by the reviewers)
- In Progress (Author acknoledgement that this is half-done work)
- Collaborative (Moderator encouraged many hands in the pot)
- Archived (Author likes it as it is, thank you very much)
- Not Reviewed (New stuff)
- Contentious (Has been deemed contentious yet acceptable)
- Front Page (Article should be linked from the front page)
In progress means it's not baseline or peer reviewed.
Baseline means it's archived
Collaborative means it's in progress (and author becomes moderator)
...So no need to over-tag.
- Has 3D
- Has calculator
- Has media
- Engineering (An engineering/science article)
- Politics (A political article)
- Environmental (Article has to do with doom and gloom)
- SciFi (Articles with a plot!)
- Index (Article is mostly pointers to other articles)
- Tutorial (Article is introductory text)
- FAQ (Article is a frequently asked question)
- Earth (Article is specific to a terrestrial Elevator)
- Mars (Article is specific to a Martian Elevator)
- Moon (Article is specific to a Lunar Elevator)
- Asteroid (Article is specific to an Elevator on an asteroid)
- Orbital Mechanics
- CNT technology
- PV technology
- Ribbon (Article pertains to the ribbon design)
- Climber (Article pertains to the climber design)
- Power Beaming (Article pertains to the power beaming system)
- Anchor (Article pertains to the anchor station)
- Architecture (Article addresses global architecture issues)
- Deployment (Article is relevant to the deployment phase)
- Operations (Article is relevant to steady state operation)
- Applications (Article is about applications enabled by Space Elevator)
- Manufacturing (zero-g manufacturing processes)
- SBSP (Space Based Solar Power)
Status flags behave in special ways:
- The Baseline tag causes the addition of a "baseline" icon, and implicitly means archived
- The Peer-Reviewed tag causes the addition of a "peer reviewed" icon, and implicitly means archived
- The Archived tag allows only the author to make edits
- The In Progress (Collaborative?) tag encourages editing, and changed the person-in-charge from "Author" to "Moderator"