WikiPOBia talk:Community Portal

From WikiPOBia

Revision as of 04:10, 29 December 2008

TOC Project

What exactly are you planning with a table of contents project? --LadyShelley 00:59, 2 July 2008 (BST)

Speaking of Being Perplexed

After reading the first few titles of the Aubrey-Maturin series by Patrick O'Brian, I went in search of a WIKI that almost certainly must exist for it and found this one. The first thing that I usually consult in a new-found WIKI is its Statistics Page. That of this WIKI reports an incredible disproportion of total pages (over 6,500) to content pages (just 250)—a rate of less than 4% actual content to total pages. (The English Wikipedia has a 17% rate with over 2.5 million articles, and the Lord of the Rings WIKI—a smallish WIKI with just over 3,000 content pages—boasts a 40% rate.) After probing around a little, I realized that there may be a problem with the WIKI's namespace array. In order for pages in a WIKI to count toward content, the namespaces in which they reside must be included in the MediaWiki canonical namespaces and content namespaces arrays. Perhaps someone needs to make additions to the $wgContentNamespaces variable in the LocalSettings.php script. (This variable can be added if missing.) What suggests this to be a problem is 1) the low content page count, and 2) the absence of many pages from the All Articles list. None of the Lexicon pages, for example, appear in this list, and these I understand to be in a separate namespace.

Yes, the Lexicon articles are in a separate namespace. This was done because the number of Lexicon articles, around 3000, simply overwhelmed the 'content articles', especially at WikiPOBia's startup. (The current Lexicon articles, most of them, at least, if not all, were assembled by script, from data from other sources, and uploaded by bot before WikiPOBia was officially opened.) So, it was though desirable that they not appear, for example, in the All Articles list. It's been a while since I've looked through all of the related settings variables, but if they can be counted as content without overwhelming the other articles in the All Articles list, that would be fine. Otherwise, I think it might be best to leave things as is. -- Jblumel 15:53, 28 December 2008 (GMT)

Was a substantial amount of the material here compiled in this way? What were the sources? Guide for the Perplexed, I can see, but what were the others? Opus 16:13, 28 December 2008 (GMT)

All of it, originally. The other source was Maturin's Medicine. There's a link at the bottom of each entry that indicates the source, see, for example Lexicon:Abdomen -- Jblumel 16:27, 28 December 2008 (GMT)

I also find that there is a bewildering number and complexity of categories here. There are almost 2,000 categories in this WIKI. I wonder whether the actual number of content pages is in a desirable ratio to the number of categories. The previously mentioned Lord of the Rings WIKI has 360 categories for its 3,000 content pages. For the same ratio to exist in this WIKI, there would have to be about 16,500 content pages. Now, I do not hold up LOTR WIKI as a particular standard—although not an active contributor, I do regard it as a particularly well organized WIKI—but it does seem to me that there is an extraordinary use of categories here. I notice that there is an odd multiplying of categories with many pages appearing in multiple category trees. This is not particularly unusual, but in the words of William of Ockham, "Entities must not be multiplied beyond necessity." Almost 1,500 of the almost 2,000 categories here are part of a multitude of Guide for the Perplexed category trees, and pages are cross-linked in these trees in a myriad of ways. One entry, Lexicon:Quackeens is a leaf in five trees in which four of these trees it is deeply nested. The English Wikipedia has an enormous number of categories, but many of them, perhaps most, do not quite satisfy the site's own guidelines for categorization and overcategorization. (Many of its categories have only one page, for example. Is the Fourth Council of the Lateran the only thing of note that happened in Italy in 1215. If so, then perhaps there should not be a 1215 in Italy category. Wouldn't it be better to roll it up under the 1215 category, from which this important event happens to be strangely missing.) Of course, that site is somewhat of a free-for-all. Wouldn't things be much easier to find here if there was a simplification of the categories? Couldn't the forest be felled and replaced with articles that arrange things together neatly?

Regards, Opus 13:46, 28 December 2008 (GMT)

First, thanks for volunteering to create some articles that would provide a guide to the Lexicon! Saves us the trouble of pressing you!

There are a lot of categories associated with Lexicon articles, particularly those from Guide for the Perplexed (GFTP). There are a few reasons for this, but the overall category/subcategory structure was dictated by an issue with the MediaWiki software -- https://bugzilla.wikimedia.org/show_bug.cgi?id=1211 -- that presents a problem for categories with a large number of entries and that also contain subcategories: in short, some or all of the subcategories often don't appear on the first page of entries, so are easy to miss.

Because of this issue and the large number of entries, and the desire to organize them by book/chapter and language (when appropriate), the current categorization scheme was developed. Many of the entries occur multiple times throughout the books, and the book/chapter categorization results in them being assigned to multiple categories, sometimes quite a few. This does overwhelm the All Pages, Category listing, so a category hierarchy was also developed so that they could more easily be 'browsed' by readers of the wiki.

There is a top level Category:WikiPOBia which is meant to contain all the higher level categories, as subcategories, and then each of these contains other subcategories, so that one can 'drill down' to more specific levels of categorization. The articles themselves are mostly contained in the lower level categories, although, there is an 'All Articles' category that contains all of the Lexicon articles in that section. There are also categories for each book and chapter that were created as places to categorize other 'content' articles, and which contain the appropriate Lexicon subcategories.

So, even though there are a large number of categories related to the Lexicon, this is necessary to allow a wiki reader to locate these by language or book chapter, by browsing the category hierarchy. Of course, they can also search the Lexicon (it is included in the default search namespaces). -- Jblumel 16:58, 28 December 2008 (GMT)

This arrangement renders many of the best tools in the MediaWiki toolbox unusable. I almost passed over your WIKI by the appearance of its Statistics and All Articles reports alone. It gives the appearance of an empty ship that is not worth the taking, so to speak. (I had already been put off by the slovenly character of the Patrick O'Brian Wiki in Wikia Entertainment.) It was the Categories list that shifted the wind. It was clear from this that there was treasure here even if most of your sails, rigging, and stays were below decks. The well known MediaWiki bug is usually avoided by not overly multiplying the number and complexity of categories. I wonder whether something more elegant and usable could have been achieved by a much simpler design without the enormous multiplication of categories. This reminds me very much of a Banyan tree in which the central bole can become completely obscured by a huge number of "prop roots". Well, it is your ship, and you sail her as it pleases you.

Note: I do not like to pad my indents because of the inevitable crowding to the right, but if you insist then, again, it is your ship.

Cheers, Opus 18:51, 28 December 2008 (GMT)

Oh, well, as for the indenting, I guess it really doesn't matter, seems to be something of a convention (or at least it was on Wikipedia at the time we got started) to sort out who said what, but, you don't have to do it. They do tend to get a bit crowded to the right after a few replies.

As far as tools rendered unusable, please let us know exactly what you mean, and we'll see if there's a way to work around those issues. If you have some ideas on how to rework the category structure so that it works better, let us know that too. Since the entire category tree was created by bot, we could restructure them the same way, so it probably wouldn't require any massive human effort. We try not to be dogmatic, and there's usually more than one way to accomplish something. It seemed like a good idea at the time, but, if there's a better way to do this, not all innovations are to be abhorred. -- Jblumel 19:18, 28 December 2008 (GMT)

It is possible to see what can be done without making any changes to the existing arrangement. If something more effective can be found, then a few SQL scripts can rearrange things as you like. I am not one, however, to come in and take over another's house and shouldn't like to try even when asked.

I originally came here looking for a definitive catalog of nautical terms having already found the GFTP before encountering your WIKI. I was confounded by the Lexicon being so oddly arranged that it could not be found by looking in the usual places—Special:Categories and Special:Allpages. These and other tools like them are extremely useful because they are standard in all WIKIs that use MediaWiki. The arrangement of things here makes these of little use and seriously skews the Special:Statistics report to make it appear that it is all structure and no content—like a house with empty rooms—when in fact there is a great deal worth getting at.

I must confess that I was stymied until I saw a comment somewhere that the Lexicon was in a separate namespace. (To old programmers like myself, this explains all. Namespaces are all about scoping, and it was clear that the Lexicon lay outside the main scope, as it were—either accidently or intentionally—hence, my suggestion about the content namespaces variable.) All MediaWiki reports and special pages that are scoped by namespaces—and I suspect that virtually all of them are—will be affected in some way by the exclusion of crucial content namespaces. It would be far better to find a way to bring them in play. It is always desirable to have all content within reach of reports and pages that are intended to reveal content and structure as well as to accurately compute statistics about them. Doing so makes the information that is present more accessible and can better reveal where any deficiencies lie. Opus 04:10, 29 December 2008 (GMT)