User:History2007

From Wikipedia, the free encyclopedia

First thought

I keep wondering why those who edit Wikipedia are called users, for they are in fact unpaid workers. Be that as it may, this is "user" History2007.

Me

If you have clicked on this page, assuming that you did not misclick, you obviously want to know about me. Briefly speaking, I grew up as a young mathematician and physicist, but my PhD was in computer science. I used to edit science wiki-pages using a different user name, but I got bored with that and I now edit humanities type pages. I still go back and edit pages on theoretical and scientific topics once in a while, but not too often any more. I have actually learned a lot about humanities and religion as I have edited these new pages. It has been fun. I like to learn as I write.

Current Favorite

My current favorite joke:

Rene Descartes goes to a bar and asks for a beer. Once he has finished his beer, the bar tender asks him: "would you like another beer?" Descartes answers: "I think not" at which point Descartes just evaporates away…..

By the way, did you know that poor Descartes practically froze to death at age 53 because he eventually accepted a third invitation (he refused twice) from Queen Christina of Sweden who then demanded early morning (say 6 am) lessons with the windows open, during February in Stockholm!

When she sent the invitation, the 6 foot tall Christina neglected to mention to Descartes that she often rode horses bareback during the winter in the snow in Sweden and liked to keep the windows open a lot. Note that in the painting on the right, while sitting she is almost as tall as Descartes (far right) standing up. I guess Descartes was used to French customs and royalty, and he used to usually stay in bed himself until noon and work and think there. The Wikipedia entry on Descartes mentions this episode very briefly, I should edit that page one day. Anyway, may he rest in peace....

Current Thought

I wonder if Wordnet's lexicon and ontology hierarchy can be used to automate relevant link addition in Wikipedia. Currently the category structure is fully determined by Wikepedia users, and quite often leaves a lot to be desired. And there are often several pages that duplicate information and do not even have links to each other.

There are a number of really simple bots that roam Wikipedia and do simple things. The use of a rich semantic lexicon like Wordnet would open a new dimension here and would also require some more serious programming. Yet, it would be an interesting artificial intelligence project that could suggest useful links and would then lead to more clever bots that roam Wikepedia. The current bots are just too simple to be interesting.

The interesting thing is that one does not have to have any specific administration rights to write this program. It can just run by reading Wikipages and use the Wordnet database. The link relevance determination can both compare Wikipedia categories with Wordnet and also use some rudimentary fuzzy logic scoring based on partial matches to determine the relevance scores. I will have to think of this a little more....

An interesting, and more immediate, application would be the use of affinity analysis on Wikipedia pages to determine interesting links, somewhat like the Amazon.com feature that says "People who bought book A also bought book B". Here one would get: "People who clicked on Leo Dupont also clicked on Marie Louise Trichet". Although the computational costs may seem daunting at first, with suitable algorithms, this is totally feasible. Indeed, Amazon has over 1 million books, and such a feature for the top 200,000 Wiki pages could follow a similar algorithmic approach. And it does not need to update internal tables for every visit, but just enough so some interesting page affinities emerge.

Ongoing thought on Accuracy

Wikepedia has clearly demonstrated that the centralization of knowledge by the users for the users can have substantial benefits. Yet, as I look at the more technical pages, I am often alarmed by the glaring errors that exist in plain sight. In many cases, it appears that well meaning students may be contributing text by copying paragraphs from various online sources, but the end result is less than correct. And these are not just obscure pages, but pages that are designated as vital.

Wikepedia pages that deal with city locations and major sights, or with literary figures and their list of books seem to be mostly error free and can be relied upon. But on more complex topics (such as biochemistry) where I am not an expert, I hesitate to rely on Wikepedia. I wonder if expert biologists get as alarmed when they read those pages, as I am when I read an advanced page on mathematics or computer science. The key problem is that the ratio of experts to general users is low, so on scientific topics we do run the risk of having an encyclopedia written by the students for the students. Clearly, one could not run a university that way and an encyclopedia should not either. The eventual solution may be to designate vital scientific pages as semi-protected so they can only be edited by science-administrators. I think that day will eventually arrive.