User talk:El Sandifer/Branching

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Initial thoughts[edit]

My instinctive reaction is that this is not a good idea. The Derrida example, for one, is something where branching is not needed. Books by an author should either have their own articles, or be mentioned in the article about the author. The lengthy biographies example is even worse. What you want there is a navigational box to provide a timeline. Subpages imply hierarchy, and that should not be confused with chronology. In my view, the current systems of: wikilinks, categorisation, summary style, and navigational boxes is sufficient. Adding branching to this has the potential to make things far too confusing and tangled. Also, people are used to seeing whole titles. "Thailand/History" for example, looks and reads badly compared to "History of Thailand". Having said that, I do like Jooperscoopers 'tabbed' article idea to have various "extra" bits or bits with different layouts, placed in a different tab (eg. stats, infoboxes, navboxes), but that would be a big change as well. See Wikipedia:Categories, lists, and navigation templates for more on related stuff. One more thing, the little line of "subpage" links across the top is precisely what some of smaller navboxes do. Not the gargantuan monsters at the bottom of articles (eg. Template:Health in the People's Republic of China), or the ones at the side of an article (eg. Template:History of China), but really small ones to allow someone to navigate around a small topic, or go to the articles closest in relevance (rather than overwhelm someone with links to everything - a strategy that also overwhelms "what links here" which gets polluted by links from footer templates). See Template:After Tolkien navbox and Template:HoME navbox and Template:History of Arda for examples. Those sort of templates are doing similar things, but putting the links at the side of the article, instead of across the top. The footer boxes (gargantuan monsters that they are) claim to do similar things. Category trees sort of provide a hierarchy if that exists. See here for example. Category trees can be placed on portal pages, see Portal:Arctic for an example. Carcharoth (talk) 14:15, 6 September 2008 (UTC)[reply]

I'm fine with retaining normal article names as we split - History of Thailand is a better and more readable title than Thailand/History. I do not think we need to use the / syntax for branched articles, and can see strong arguments for not doing so.
But I'm looking at, say, George W. Bush, and that article looks to me like it is already arranged with de facto sub-articles in the manner I'm describing here - there is a part of the article where George W. Bush as governor of Texas goes. The entire article is clearly a sub-article that drops in at a specicific point. And George W. Bush substance abuse controversy is just as clearly an expansion of Early life of George W. Bush. Perhaps the problem here is with the term branching - I see your complaint about hierarchy, but on the other hand, there is a nesting structure of sub-articles already in place here.
The Derrida example is trickier. But looking at what we want to cover there, we want to cover his biography, his philosophical thought, and criticisms/reactions to that thought. That is too much for one 60k article - covering that is necessarily going to be split over multiple articles. The question is how best to do that. I am open to other suggestions, but I think the overall question is a real one. Templates work to a limited degree. But the real problem here is that "article" for us is defined as a 60k chunk of text. And 60k chunks of text don't have a 1:1 correspondence with "independently notable aspects of a topic that are necessary to cover as part of encyclopedic coverage." It's telling that EB, for instance, does not maintain articles on books by major thinkers - Critique of Pure Reason effectively redirects to Kant, with some notes of where it appears elsewhere. But at 88kb, their Kant page is already longer than we would want ours to be. So when we start splitting off by book, we're already splitting off of what most encyclopedias would. I agree, individual books by Kant and Derrida should be covered in detail. But we should be aware that we're, in doing so, dividing what most encyclopedias consider a single topic.
I'm familiar with the categories, lists, and navigational templates page, having written the original version. I think it is telling that, despite the clear advice of the page, templates and navigational templates took off in weird and largely useless directions, and categories became so bloated as to be useless as well. As a reader, I virtually never use categories, and rarely use templates. In any case, I think there are still large problems to solve in organizing the disparity between topic and article for us. Phil Sandifer (talk) 14:42, 6 September 2008 (UTC)[reply]
Well, I use categories a lot, footer templates not a lot, and lists only sometimes (but when I do use lists they are very useful). I think a better mental model is of a central article, with topics branching off in many directions (not just "downwards"). I "spider graph". But then this is essentially what categories do already. I forget the exact term for the type of branching and interconnected network that is formed. Note that this only becomes apparent with a well-organised, mature and appropriately trimmed category structure. Early on, it can look llke a complete mess. The main gripe I have with the summmary style approach is the need to keep things up-to-date. In some ways, some system of templates would help there, so that when you edit a section that is a summary (or a page containing a summary section), the software opens up the main article and asks if you would like to update that as well. The problem comes when reverting such "double changes". How would your system handle synchronisation between summaries and main articles? Carcharoth (talk) 15:06, 6 September 2008 (UTC)[reply]
Well, the sub-articles feature could be used for a spider graph as well. Perhaps the problem comes in the excessive focus on article reconstructability? I am not married to that point. Phil Sandifer (talk) 15:13, 6 September 2008 (UTC)[reply]
Reconstructability? You've lost me there. :-) BTW, have you seen Wikipedia:Article series? Carcharoth (talk) 15:18, 6 September 2008 (UTC)[reply]
The focus on being able to reassemble the sub-articles into the main article to produce a coherent whole. And no - I'd not seen article series. Phil Sandifer (talk) 15:31, 6 September 2008 (UTC)[reply]