Last modified: 2006-10-21 16:38:31 UTC
Similar to how we allow escaping of raw html, javascript, php, et al. Implementation: LocalSettings.php would have a switch which would allow tiny bits of xml to be stuck in the content of pages. ( $wgAllowRawXML=true||false; ) Example content formatting IN the articles might look like this: <aritcle_title>MediaWiki</aritcle_title> <definition>A fantastic piece of software.</definition> <type>Wiktionary Entry</type> <version>1.0</version> OR <subject_title>Spider-Man</Subject_title> <real_name>Peter Parker</real_name> <height>5'11"</height> <timestamp>12:34:56-12/25/06)</timestamp> Reasoning: For allowing external applications (which could be web, or windows/mac/linux client software) to not only pull full content out of MediaWiki (Similar to functionality in API Bug 208 & XML export), but once they have the contents of the article, they can pull detailed, specific data out of the article itself as well. The above situation(s) could be universally useful to Wiktionary and Wikipedia. Currently, a third-party company called Webaroo(.com) takes full database dumps of Wikipedia to provide offline content to PDAs and laptops. Imagine the power of being able to dynamically pull an entire dictionary, or parts thereof, to your PDA (offline) and 'synchronize' it periodically? Surely, these are only a few of the possibilites... P.S. Is there already some sort of functionality/API/methodology that might allow this, or something similar?
This can already be done in any number of ways, such as storing them in template parameters that don't show up as anything. Templates by themselves are pretty easy for anyone to parse; see also [[Wikipedia:Persondata]]. What advantage would we get by disappearing anything unrecognized that looks like XML? And where is this XML to be output, if anywhere?
You can do this trivially as a custom extension. Of course this would never ever be allowed on a public site like Wikipedia since, being identical to $wgRawHtml, it would be a huge security risk.
