Talk:Frame (artificial intelligence)

Different frames
The description of frame on Frame appears to refer to a different kind of frame: that kind of frames are typically used in expert systems but the description in this article doesn't seem to describe the same frames. Are they the same? Otherwise a new article should be written about the frames mentioned on Frame. - Simeon (talk) 22:16, 29 June 2008 (UTC)
 * As far as I can tell it's the same construct. Frames weren't nearly as important to commercial expert systems as rules. Also, Frames -- at least originally -- didn't come from expert system research as much as from more generic AI, the early work on Frames was primarily in natural language and common sense reasoning, things like ordering at a restaurant (an early classic frame example) hardly qualify as "expert" tasks. The article on Frame language is a lot more comprehensive. I think perhaps this article should be merged with Frame languages. MadScientistX11 (talk) 20:15, 9 December 2013 (UTC)

Merge with Frame Language?
I propose merging this article with Frame language. It's like having an article for Object class and object-oriented analysis and design, they are really the same topic IMO. MadScientistX11 (talk) 20:16, 9 December 2013 (UTC)
 * Note: I formalized this using the appropriate tags (I hope I did it right, first time I've done this). I think the convention is you discuss it on the talk page of the article that is to be merged into (the one we keep) which in my proposal is the Frame language article. Here is a link to the discussion section topic on that page: Talk:Frame_language MadScientistX11 (talk) 03:50, 10 December 2013 (UTC)

Sorry, but I don't agree. Yes if the article would stay as it is. But I think the article actually describes and is concerned with one implementation of the Frames framework. Although it is understandable as it is the one from the creator and salient scientist, Frames in AI can be considered a cognitive framework, which does not imply talking about any particular language. 189.217.23.52 (talk) 03:36, 30 January 2014 (UTC)

I've undone the merger. The concept of frames is quite distinct from particular frame languages, and, if anything, the page on frames is more fundamental. A merge in the other direction would not be implausible, but the concept of frames is the central idea. It is both older than particular frame languages, and is philosophically distinct, as frames serve to address particular questions in AI and philosophy of mind (e.g. the frame problem) which are not tied in any particular sense to practical issues of knowledge representation and frame languages. Furthermore, the idea of frames shows up within linguistic semantics as well as in AI, in a way that has no connection to frame languages, and so it seems reasonable to have a page for the AI-analog of linguistic frames. Plausibly, one might want to merge the AI frame article and linguistic frame article. Either way, merging this page into the frame language article resulted in a loss of useful information on this page, and that alone is enough to warrant retaining it as a separate page, at least for the time being. Augur (talk) 23:04, 20 July 2014 (UTC)
 * If you are talking about "Frame" as a concept in linguistics that is a different matter. But if you are talking AI then you are wrong. Frame and Frame Language are the same thing. And the article that I merged was definitely talking all about Frames in the AI sense. Going back to the earliest uses of Frame by Schank there was always some language involved. It's no different than an object and object-oriented programming in computer science, you don't have an object without an object-oriented language (even if that language is an analysis language like UML) and you don't have frames without some language. Everything in the article you tried to bring back was talking about Frames in the AI sense and everything that was in that article was said better in the Frame language article. If you want to create a new article Frame (linguistics) feel free otherwise, unless you make a better case than what you have so far. --MadScientistX11 (talk) 13:24, 21 July 2014 (UTC)
 * You ignored objections to the merge and went ahead with it anyway. The fact that you improved the Frame Language page instead of this one does not justify the merge. It just means that you improved the wrong page. I'm going to request protection on this page. Augur (talk) 18:31, 21 July 2014 (UTC)
 * Your analogy to object oriented programming doesn't hold water, given that there are currently *three* distinct pages for objects in CS, each serving a validly different purpose. Object (computer science), Object-based language, and Object-oriented programming language. At any rate, the concept of frames is a theoretical framework that exists apart from any particular way of implementing it, just as you can have object-oriented programming in languages with no native support for objects, as is often done in C or Scheme through libraries. As already mentioned by the anonymous use in January, the frame model is a theoretical model independent of implementation. In fact, your very good edits to the Frame Language page outline a long history of frames independent of frame languages. There is value in distinguishing the concept of frames from the particulars of how frames get used for particular projects. Augur (talk) 21:43, 21 July 2014 (UTC)

Reverse merge
I agree with the above comments that this article is the more general subject, of which frame language is a subtopic. There's a lot of overlap in the articles, such as history and explaining the general idea. Frame languages are in fact a major subtopic of semantic frames, and to write down an example, we either have to use a real frame language or make one up in the form of a consistent notation. So, I've proposed merging frame language into this article rather than the other way around. -- Beland (talk) 18:20, 3 March 2019 (UTC)
 * ✅ Klbrain (talk) 10:29, 15 June 2020 (UTC)