Semantic MediaWiki
1Liner
: Semantic MediaWiki (SMW) helps organize and retrieve data in a Wikipedia-like website.
|
|
Order from Chaos[edit | edit source]
The english language is not very structured. There are some rules, but many exceptions and need to have a deep understanding of context and intent to gleen what is meant. It's basically just a pile of words (that can be said in many orders) in a stream of consciousness, in order to get a point across. And like this paragraph, it can be somewhat hard to parse... especially for computers.
What Semantic MediaWiki (or other sematic tools) can do, is they allow you to tag key fragments of data -- to give little hints as to context for the computer/software. That allows you get create new articles, or glean insights from all the data fragments you've already added -- without having to do it manualy.
That's a bit abstract, so let's go a litte more real world: I use a bunch of quotes on the site, and I create a page for each. Normally, that's just a bunch of page quotes that you can search for. But with Semantic hinting, I noted the author, the citation (where and when it was done), and some tags on what topics it belongs to. That means that if I create an author page, all the quotes (and books, articles, etc) by that author, can be gathered up. The same by date, or by a particular topic. The tags, parameters allow many more ways to find or display the same information. Instead of someone having to manually have to create one index for every possible way the reader might want to find things, the hints allow the computer to automate it.
So SMW turns the chaos of freeform articles, into the order of semi-structured data. Of course I have to add the structure (categorization), and add the data (keys and their values) to every article. But SMW lets me extract those nuggets that I put in there -- both from my site and other sites that use SMW.
Others[edit | edit source]
| 🗒️ NOTE |
|---|
| I'd done the first Wiki version of this site, using DPL (and LST) plug-ins -- which allowed me to put some structured data in, and did a lot of the organization. But display of results was limited. And the structures were inherently fragile. In coding lingo, I'd say it was more like procedural programming than object oriented -- if you named everything right, it could work similarly. But it could be fragile, and hard to debug when something went wrong. |
Are there other data organizers? Sure. Using Wikipedia's engine (called MediaWiki) there are the following:
- SMW - The most full featured, with the largest support, most features, and is widely used -- with a lot of German and European enthusiasm keeping it alive.
- Cargo - A lighter weight and simpler to understand version, created by a legend in the community: Yaron Koren. But it doesn't quite have the following/scale
- DPL3 - (Dynamic Page List 3) - this is basically a way to grab fragments from other pages. But with a judicous use of consistency in naming, you can do a lot.
- And a few like WikiDB and Wikibase - that more allow you to retrieve data from tables that you manage... instead of allowing you to use articles to structure information.
That's before looking at other website engines or content management systems.
Some of the concepts of structured data aren't that hard. Like tagging to find content is widely used in many apps. Putting it together to dynamically design pages? in that, you've narrowed the pool to a lot fewer contenders -- and the vast majority of them are designed in a way that you must structure the data the way they want, and display the results with the same limitations. SMW and MediaWiki are basically more like giving you an engine and letting you build the rest of the car... instead of giving you a car and just allowing you to pick paint color. If you like the car they delivered, the latter is great. If you need to design things that no other car has... then an off-the-shelf car, with a few options and colors, just won't be enough.
