Shall we mark important topics?

[Martin Baldwin-Edwards]

How about a random"developed article"? There are not yet enough approved articles to warrant randomness, but maybe later...

[Robert_W_King]

Why not just label it "Pseudo-random" (actually, that's true in part anyway) and put a disclaimer somewhere that says "this effect is based on the number of articles per catagories and is representative of a pseudorandom sample of the distribution of article types"

[Joe Quick]

Why not just label it "Pseudo-random" (actually, that's true in part anyway) and put a disclaimer somewhere that says "this effect is based on the number of articles per catagories and is representative of a pseudorandom sample of the distribution of article types"

If we are going to weight the results in the interest of resulting in something interesting, why not give extra weight based on page views?  Popular destinations are generally interesting destinations, right?  Lots of people go to Acapulco because its a lot of fun.  There's a minor chance that this will result in a few articles getting way out in front of the crowd because they start with higher page views and then continue to get more and more page views as a result of the weighting, but this can be easily addressed by keeping the added weight low.

If it were up to me, though, I would leave the random page function alone.  When I use the random page links, I'm not looking for a new passion in life, I'm just looking for something I didn't know.  Plus, we want to avoid looking like we're trying to make ourselves look like something we're not.

[Jitse Niesen]

I think you will find that all scientists and mathematicians will object to a random page which is not random, but rigged to present a non-random view of something. How about another name for it?

This mathematician has no problems with it. "Random" does not mean that every page has an equal chance to be chosen; it only means that the page is chosen according to some probability law.

Nevertheless, I guess most people would expect "random page" to pick every page with the same probability, and that it would thus be misleading rig "random page".

[RJensen]

There is a more serious issue here. I suggest CZ have a main branch for our usual articles, and a separate branch for large collections of standardized list-like articles (like 10,000 genomes, 5000 cricket players, or whatever).  People interested in genomes will go to the right place.  The "random page" function is exactly like picking up a new book and flipping through the pages.

[Joe Quick]

There is a more serious issue here. I suggest CZ have a main branch for our usual articles, and a separate branch for large collections of standardized list-like articles (like 10,000 genomes, 5000 cricket players, or whatever).  People interested in genomes will go to the right place.  The "random page" function is exactly like picking up a new book and flipping through the pages.

I would be all for this if I thought that we could define and draw a line between important and not-so-important articles.  I realize that paper encyclopedias do this, but they do it arbitrarily and due to lack of space.  We don't have to think about space, so I don't think we're justified in drawing and arbitrary line and I don't think there exists a natural line - some articles are clearly above the line and some are clearly below it, but it is a mirage that disappears as you approach it.

[Derek Harkness]

Richard Jensen and Stephen Ewen comparisons to search on Google and Amazon contain a major flaw. The search results form Google and amazon list a large number of articles that are actually the same or very similar. The are not significantly different. Where as a search on wikipedia or citizendium will produce a large number of articles that are all different.

Google will happily bring up 2000 articles on apple pie. CZ will have one article on apple pie, an article on apples and an articles on pies and then some other types of pie. Search for 'true vipers' on Google an you will get hundreds of articles about the general category of 'true vipers'. Do the same on CZ and you will get just one article on the general subject, all the others will be individual varieties of viper.

Ultimately, the most important article is the one the user is looking for. All the other articles are irrelevant. Google has to try to do the job of determining which article is most important out of several that seem to be an equally good match to the query. We don't have to do that, we will have very few articles that closely match any given search query.

Richard example of the research librarian is a close match to CZ. The librarian has a huge number of books but knows that few of them will match the users needs. The Librarian filters the selection of books and presents the user with the few that seem to match the requirements. The librarian's task is similar to that of our search engine. It tries to weed out the irrelevant articles and supply the user with the best match. However, the librarian must have the huge library of books to match potential enquiries. If you try to narrow the selection of books available to the librarian, then they will soon become frustrated as they are unable to find matches for the library's users.

To give an illustration that Richard may find pertinent. If you were researching the war of 1812 and the librarian came back with a book about World War II you'd be pretty annoyed to be told that they don't stock anything about 1812 as WWII was a more important war. What is important is decided by you as the user based on what you want at that moment in time. Importance based on the 'grand scheme of all knowledge' has little to add to the relevance or irrelevance to the question asked. WWII is important, but if you want to learn about 1812 the WWII is utterly useless.

We need two things. First a better stocked library. Second a better librarian to search the library. (actually I think our current search function does quite a good job at matching articles to queries. Where it falls down is in the speed of the search. We need a slightly more nimble search rather than smarter)

Larry's suggestions reminded me a bit of the computer Deep Thought in the Hitch Hikers Guide to the Universe by Douglas Adams (a golly funny read if you have read it yet.) Deep thought declares that the answer to the ultimate question of life the universe and, well, everything is in fact "42". In reaction to peoples surprise, the computer goes on to explain that it's users didn't understand what the question was but that the answer was definitely 42.

[RJensen]

Anyone who cannot tell the difference between what is important and what is not on topic XYZ should not be writing for CZ on that topic. Send them to Wikipedia! 

[Joe Quick]

Anyone who cannot tell the difference between what is important and what is not on topic XYZ should not be writing for CZ on that topic. Send them to Wikipedia! 

Maybe that's true for people who are trained in a field of study, but there will be constant conflicts over how to categorize the sort-of-important topics.  You might be willing to play the politics, but I have no interest in creating areas of conflict where we don't need to.

[Argyriou]

One use for the "random article", which works best when it is truly random, is to pick a random article, and see if there is any way you can improve it. Most short articles have *something* wrong with them, even if it's just a typo or two, and can be improved by someone who is not an expert in the particular subject.

It would be interesting to find out how people use the random article function at Wikipedia, to see what the real value proposition is for its existence.

[Aleta Curry]

I've got a different issue with "random topic".

Can the random topic thing be set to EXCLUDE some of the subpages, like Green Spotted Snail/Approval?

At Jitse's suggestion I used "random" to find something to edit--amazing the number of non-editable pages that came up.

Or can we maybe ADD a button to be used just for finding stuff you can edit which excludes the subpages?

[Martin Baldwin-Edwards]

Yes, the random feature obviously should not normally look for subpages. I think it should be like flicking through a book in the bookshop, just to get a general idea of how bad it is before you waste your time and money on it 

[Derek Harkness]

The random pages and also the search function need tweeked for subpages. The default set of subpages should probably be excluded form the search results. There may be instances where you will want to search subpages, for this there should be some sort of checkbox like we have just now for searching within other namespaces.