Meta-problems in general

Concerning n : 1

Problem edit

In the beginning there was one Wikipedia in English. Wikipedias in other languages followed. Each project was set up in an independent Wiki.

Edgar F. Codd proposed to store some information only once instead of many times (database normalization, relational model) if it is the same information.

Each position where this information is needed just contains a reference to the place where the information is stored.

In reality we find some redundancies now. I try to submit solutions as well.

User names edit

Every user was asked to create a new account in each project. Consequences are:

  • the same user might have different user names in the projects
  • the same user name might belong to different users

It is discussed in Single login specifications.

My proposal is to create super accounts. For each local project 1 local account can join (validation by local password, realisation by cookie, here).

Images edit

It would be better to store an image with 100 kB only once (100 kB or better quality) instead of 100 times (10 MB).

  • the same picture can have many names in the local projects
  • the same name might be contain different pics (or versions)

Still, new pictures are uploaded locally every day. The local databases are still growing and this problem of redundancy and inconsistency is increasing.

I know no solution execept closing national wikipedias for new uploads.

Interwiki links edit

We need 1 common name space for interwiki links.

Experiment:

  • There are articles about Big Ben in 100 projects.
  • Each article contains 99 interwiki links (in different order, 9,900 all together).
  • Now there is an article in an another 101th Wiki.
  • So 100 additions have to be made in each Wiki.

If all 100 articles would relate to one commons field, there would be 1 addition.

Geographical coordinates edit

Stefan Kühn from Germany is extracting the coordinates (34 thousands in English WP, 22 in German language) and creating KML files for Google Earth. You can download them here.

Experiment:

  • The article Big Ben will get a coordinate in each project.
  • Result: 100 times a coordinate is created
  • Some coordinates are different.
  • A collection for all projects would be "inconsistent" because several signs would mark the same object.

Proposal: to deposite 1 coordinate in a common edit field. 100 references in the local projects to this field would do.

Geographic Names edit

The vast majority of names of places, countries, rivers, valleys, mountains, streets, glaciers, lakes, sees, oceans, and continents are identical or rule base transcriptions from one language to another - only few famous ones or near ones sometimes are not. Automatic or semiautomatic linking between those entities might assist editors in finding more data, pictures, and information quickly, whe they work on those names.

Dates edit

Most articles on persons tell when and where they were born or died, became e.g. mayer or prime minister of ..., etc.. Many places have been funded, mentioned the 1st time in a historic paper, 'discovered', 1st climbed, etc., at a certain date. A common data base of such date informations, how ever it would have to be built, might make new data, e.g. of deaths, or dismissals from office, accessible more quickly and reliably and economically (n:1), and if set up to do so, would foster automated conversion from/to each calendar of each culture worldwide including ancient ones for wikies in historic languages.

Wikidata edit

In the same way hidden tables with personal data and various other objects should be centrally stored and managed (consolidated to "n:1").

I think, this should include interwiki links and geographical coordinates.

How about linking to a [[wikidata:Big Ben]] and its content could be included into the article like a template?

See: Wikidata

Comments edit

Your opinion is appreciated. -- Simplicius 12:38, 3 February 2006 (UTC)[reply]