[HN Gopher] Semantic MediaWiki
___________________________________________________________________
Semantic MediaWiki
Author : Tomte
Score : 35 points
Date : 2023-06-24 20:22 UTC (2 hours ago)
(HTM) web link (www.semantic-mediawiki.org)
(TXT) w3m dump (www.semantic-mediawiki.org)
| nologic01 wrote:
| Semantic technologies, linked data never really delivered. The
| "Semantic" aspect was in sense a hyped, exaggerated term, long
| before ML/AI hypes became regular abusers of linguistic precision
| at the service of marketing.
|
| But had they been adopted for what they are: well defined
| metadata according to vocabularies and ontologies that following
| accepted, interoperable standards, the development of ground
| truth and knowledge graphs would have been much more advanced
| that the current state.
|
| So my guess is that projects like semantic mediawiki, wikibase
| etc will eventually see renewed roles and gradually link with the
| more statistical and algorithmic processing of information.
| facu17y wrote:
| This project reminds me of this now-ancient blog post by
| @marcfawzi from the Evolving Trends blog (does not seem to be
| maintained):
|
| https://evolvingtrends.wordpress.com/2006/06/26/wikipedia-30...
|
| It took more decades and the arrival of Transformers and self-
| attention (see Attention is All You Need) for the "End of Google"
| (the search engine, not the company) to become a reality.
|
| I was able to find Marc on Twitter: https://twitter.com/marcfawzi
|
| Seems to have picked up his old idea of a geek-run VC fund, but
| using AI: https://evolvingtrends.wordpress.com/2006/06/21/geek-
| run-gee...
| neilv wrote:
| When I was first working on distributed "semantic" Web+HCI+AI
| stuff, in the late '90s, nobody was concerned with monetizing --
| you just did useful or interesting things, and the money will be
| figured out eventually.
|
| Today, the first thoughts that comes to my mind aren't what my
| work could do with this information, but more about
| sustainability, accuracy, and power.
|
| A bunch of huge companies are going to hoover up all the data
| you've compiled, and use it in such a way that most people using
| it will never see your Web site.
|
| So won't be prompted to contribute back wiki-style, won't get all
| the benefit of your quality control, won't see your donation
| appeals, won't know you exist, won't see the other things that
| you would like to see, won't give you analytics on your impact,
| won't fund you with ad impressions, etc.
|
| Which might all be OK for your use case, but over time it could
| break most of the current useful percentage of the Web.
| detaro wrote:
| Would love to hear people's experience with it, especially for
| more complex things. Seen it used a few times to add consistent
| metadata to pages (through forms), but never much built on top of
| such basics.
| dankilman wrote:
| The guild wars 2 wiki makes great use of it imo
| duskwuff wrote:
| Perhaps surprisingly, it isn't particularly useful for general
| encyclopedias like Wikipedia. Real life is messy; it doesn't
| always reduce cleanly to a set of key/value pairs.
|
| Where extensions like SMW really shine is for wikis which
| discuss highly structured content, like content in video games.
| Another extension which I've seen used in this space is Cargo
| [1].
|
| [1]: https://www.mediawiki.org/wiki/Extension:Cargo
| RicoElectrico wrote:
| Check out WikiChip for an example of it.
| https://en.wikichip.org/wiki/WikiChip
| resolutebat wrote:
| Around 10 years ago I experimented with this to try to create a
| geographical hierarchy of articles in a wiki.
|
| It was, in a nutshell, not great: there's a fundamental
| contradiction between wikis, which are very loosely structured
| and edited by humans who do all sorts of wacky unpredictable
| shit, and hierarchies, which need to be carefully pruned and
| maintained.
|
| So this use case didn't work, and I think it's telling that
| (AFAIK) not a single Wikipedia project has adopted it.
| detaro wrote:
| as duskwuff said, wikipedia projects are a very open scope,
| that'd be hard to model. So I really wonder if there are any
| complex (but not as-complex-as-wikipedia) expamples people
| had success with.
| hirundo wrote:
| "The purpose of SMW is to allow users to improve the structure
| and organization of the knowledge in a wiki by adding simple,
| machine-processable information to wiki articles."
|
| With LLMs those tags have suddenly become a lot cheaper to make.
| Arguably their cost was the largest barrier to the semantic web
| vision. But now it may already be practical to semantically
| markup wikipedia content automatically as submitted. But does AI
| moot the semantic web, or enable it? Is the semantic layer
| redundant with an AI interface? I can't predict.
___________________________________________________________________
(page generated 2023-06-24 23:00 UTC)