Proposal talk:Semantic wiki-web
Is this something like an advanced search or a dashboard? Or a web scraper? Joey the Mango 17:56, 13 August 2009 (UTC)
- I think what is proposed is a way to type in questions, like "How can humans live longer?" and get back relevant results. Search engines have been trying to do this for years, with varying levels of success. -kotra 23:19, 13 August 2009 (UTC)
Don't understand but there is Semantic MediaWiki
I don't understand the proposal at all. The example in the proposal "How to extend the life of homo sapiens" sounds more like a desire for topic management like en:SKOS and/or natural language querying.
However, there already is a "semantic wiki-web" working great with MediaWiki. The en:Semantic MediaWiki extension lets wiki markup indicate semantic relationships:
- why this page links to another
- [[is spouse of|Hillary Rodham Clinton]]
- what some number or value represents
- [[has current population|3152350]], [[has torque|240 Nm]]
You can convert existing templates to make the annotations, and over time start annotating links and values on wiki pages.
With these annotations, humans and tools can make smarter queries that are impossible currently and are poorly simulated by creating static categories and lists:
- list the spouse of every female politician
- show me the city with the highest population
- list the highest-torque front-wheel drive cars.
(Those queries may not be as sexy as getting a wiki to answer vague questions, but such annotations are a requirement to make science fiction tractable.)
- Thank you for help. I see now, that before any formal structures we need to develope very hierachic, exactly structure all information om WP - language. Then all user, who know that terms, may quckly find what he needs. And no super-semantic extra-tools Fractaler 11:29, 17 August 2009 (UTC)
In first sorry for my bad English.
I mean search tools (for searching in Wikipedia) like combobox of search string Google, Nigma and of course Wikipedia, but more advanced (realy, advanced search and "Semantic wiki-web" is bad name): search not only in string, but and in combinations of title of templates and fields, categories and subcategories/articles, in disambiguation pages. Now we have only static string of titles. Step by step search - search disambiguation pages, look for some special word (if you know it), go to page with that special word, reading, get next special word, go to next links and so on. And no tools for a choose of sequence of "semantic" links of objects one to other.
I add image for illustration the proposal
About Semantic MediaWiki i did't know, thanks. Also about Semantic MediaWiki said on Proposal talk of my second proposal - Proposal talk:API (article=class) (communication "human-wikipedia", "wikipedia-program", "search-tools" for out of Wikipeida devices?). Thank you for your questions and information. Fractaler 08:12, 14 August 2009 (UTC)
- I think you probably want http://freebase.com -- is that the sort of thing you were looking for? 220.127.116.11 18:46, 14 August 2009 (UTC)
- Thanks for usefull information, i'll check it. about details my aim - here in ==The main problem== Fractaler 11:29, 17 August 2009 (UTC)
Turning on the semantic extension
Is that part of your proposal? Hillgentleman 14:38, 15 August 2009 (UTC)
- Yes. 1 step - to do hierarchic structure (HS) of all information (like in infobox) - to exclude synonymes (disambigous). 2 - use such HS not only in WP, but and external devices - about it on Proposal talk:API (article=class) Fractaler 11:29, 17 August 2009 (UTC)
What do you think of answer engine en:Wolfram Alpha? Wikipedia says that Wolfram Alpha has many parallels with Cyc (Proposal:A 'common knowledge' database - like 'Cyc'). Well, Wikipedia says that Wolfram Alpha runs on 10,000 CPUs. That´s a lot. --Goldzahn 03:08, 16 August 2009 (UTC)
- A lot of technologies finally trying make answer's services (while only of human) the most quickly. But WP-database World fills. If WP will have API like OOP (for example: WP.AI.Technology.ListAllFields() ) and we can get info into our program, then success I think garanted. But if Wolfram Alpha is double of WP - for why it? Fractaler 11:29, 17 August 2009 (UTC)
- I tested Wolfram Alpha, but found its answers are of the BBS scripting type, i.e. a huge list of data that is pretty relevant, but where I had to find the answer myself. This could simply be achieved by sorting word according to a numeric significance level, choose the ones over a certain threshold value and then (by set intersection) select answers that pertain to this selected word set.
- Nevertheless, since the proposal question was about semantic web, one might speculate in a manual or an automated data tag-up of wikipedia articles. A full tag-up would probably be superhuman whether manual or automated, but since many articles uses infobox templates, one might consider automagically perusing the infobox templates to get a shallow and preliminary tagup. Rursus 11:07, 20 August 2009 (UTC)
One of the parts where many articles already provide structured information are infoboxes. It would be great to have advanced search that would first allow to pick between existing infoboxes and then do search based on the infobox fields. Audriusa 11:22, 16 August 2009 (UTC)
- You are right. And infoboxes is forlam structure. It migth be even page of WikipediaNextGeneration - fully formal page wtih entirely structured information not only for numan (see about it on Proposal talk:API (article=class) for program) Fractaler 11:29, 17 August 2009 (UTC)
The main problem
I think the main problem for all (user, WP) is united languaпу (words, spacename, terms). Baby use baby languages, scientists - science, religion - religion words, philosophies - philisophies name space (terms), WP - WP-words. So, if WP want user has understood it information, WP must have something like theacher (wizards), wich know all languages (namespace) and help to find what user needs. Tools for that wizard - media (images, sounds, video), disambigous and so on. So, after X year everybody - knows WP-term and talk with WP (and with other users) on united language. Fractaler 11:29, 17 August 2009 (UTC)
Some proposals will have massive impact on end-users, including non-editors. Some will have minimal impact. What will be the impact of this proposal on our end-users? -- Philippe 00:16, 3 September 2009 (UTC)
Cooperate with Wolfram Alpha, Powerset and/or universities
I'd recommend to make this an actual research project. --Fasten (Wikinews: Aktion Deutschland Hilft asks for donations after the earthquake in Indonesia) 14:57, 11 October 2009 (UTC)
- And, of course, one shouldn't ignore the Theseus research project (in English) in this context. --Fasten (Wikinews: Aktion Deutschland Hilft asks for donations after the earthquake in Indonesia) 15:00, 11 October 2009 (UTC)
examples not the simple
Input →Output usually doesn't work without considering all other relevant factors. Even in what was intended as a trivial example, Process: H2+O2 Results of that process: H2O This does not take place without a catalyst, typically an electric spark
In the other example, increasing the lifetime of humans can be done with other things than changing their DNA, and an arbitrary change in the DNA is exceedingly unlikely to have a beneficial effect. There are the necessary intermediates of finding what genes that can be changed to cause an increased lifetime , finding out what changes will do it, and finding out how to make the changes. It will be a long time until we can fill in these steps, so long as to make it fairly certain that if I wish to affect my own lifetime, I should consider environmental rather than genetic factors. The change in our knowledge since about 1850, is our realization that if we are interested in the much longer term, this is a sensible direction to pursue, rather than being pure fantasy. .
To take a well-known problem in enWP, a football player is notable if he played for a fully professional team. This seems trivial: you find a roster, an look for the name. As interpreted in enWP, it is necessary however to show that he actually played, not was merely on the roster, and that the team was fully professional at the time he played on it. In unfamiliar geographic areas, that second step in particular can be difficult to determine.
Things happen in the context of the entire world. The only closed system is the universe as a whole--and even that is an assumption. The sciences are devoted to selected out of all the possible things that might affect one another, those things that are likely to have a significant effect & then testing to see that they account for all the variation.
This does not mean we should not explore the semantic wiki--I strongly favor a much greater degree of structure in our articles. It does mean we need to limit ourselves to those aspects of information where we can construct an adequate structure. Viewed from a slightly different direction, the labor of preindexing fields for searching needs to be less than the labor it would take to do unsophisticated searches. DGG 18:40, 11 October 2009 (UTC)
Education search engine and web filter
One could use semantic web technologies for an education-specific search engine and web filter that could be used by parents and schools. An interesting model for licensing a web filter could be to invite schools to license it and to make parents their customers. The filtering system could be combined from options selected by each school individually, including the addition or removal of individual web pages. --Fasten (Wikinews: Aktion Deutschland Hilft asks for donations after the earthquake in Indonesia) 13:00, 13 October 2009 (UTC)
- See also: Education Wikia Search, NSDL Search, http://discovered.creativecommons.org/search/, Google for Educators
- Now a proposal: Proposal:Education search engine and web filter
sounds very nice
Hummm, I dont know if I've got the whole thing, but sounds nice. Nevinho 23:25, 2 November 2009 (UTC)