Here are some of my personal Wikipedia philosophies and ideas for improving the site. Discussion is welcome, although I may move it to a separate page after some time.


Improving Wikipedia edit

Scope: dictionary definitions edit

I believe that Wikipedia currently has an appropriate scope, and that the policy for inclusion are about right. However, there is one policy that I would like to see changed: I believe that Wikipedia should allow the inclusion of dictionary definitions. First, this is often used as grounds for deletion of short articles that could—in my opinion—become high quality encyclopedia articles by current standards. More controversially, I believe that etymology and usage of words are encyclopedic topics, that is, that words themselves are worthy subjects of encyclopedia articles. Words are very important, and there are a reasonably limited number of them (even the extremely thorough OED has fewer entries than there are articles already in Wikipedia; most of these words are not noteworthy enough to warrant articles).

I am of course aware of Wiktionary, but it simply does not have the level of activity that Wikipedia has, and I see no reason (other than arbitrary historic divisions) to make the distinction between "articles about anything important that's not a word" and "articles about words."

Database dumps edit

I strongly believe that the off-line database dumps should be performed much more frequently. These used to be released weekly, which would be tolerable, but in 2005 they slowed to a pace of one per month or less. Database dumps are critical for many reasons:

  • They allow for distributed backups, in case something happens to Wikimedia (although hopefully they are doing backups more frequently than once per month)
  • They may be required for GFDL compliance (?)
  • They are used to build wikipedia mirrors (although I believe these are somewhat pointless)
  • They are the primary source for a number of off-line analysis projects (like my own), which become rapidly less useful the further out-of-sync their data gets with the current state of WP.

Dumping only the cur tables from the most active projects (ie, en.wikipedia at least) more frequently would satisfy at least some of the above goals, while keeping the burden on the servers down.

Update: In 2009 the dumps appear to be updated pretty regularly, though a three-week period is still a bit longer than ideal.

Random pages edit

Wikipedia is immense and interesting; it is an excellent way to kill time. The random article feature is great for finding a few jumping off points. However, when I use the random article button, I usually need to open about 20 articles before finding one that is interesting, and the reason is that a large number of articles that come up are articles about specific towns, or articles about specific historic figures. These certainly deserve articles, but are pretty pointless in random page results, because:

  • They are generally uninteresting
  • The potential for improvement by the reader is low, since the reader is unlikely to know anything about the specific area (most of these articles are just lists of facts).

Friends of mine have independently observed this same fact. I would love to see the "random" option improved; some ideas:

  • Allow for the exclusion of stubs, places, and/or biographies
  • Provide a random button when looking at a category, that takes the reader to an article from that category at random.

TO-DO edit

I haven't written about these ideas yet.

  • client-side rendering
  • better image quality
  • formatting problems ([edit] links)
  • confusing placement anyway
  • automatic style/syntax checks at edit time