Path: utzoo!attcan!uunet!mcsun!ukc!warwick!expya!jtr From: jtr@cs.exeter.ac.uk (Jason Trenouth) Newsgroups: comp.object Subject: Re: is-part-of relationships Message-ID: Date: 20 Nov 89 21:50:51 GMT References: <469f709a.20b6d@apollo.HP.COM> <31417@news.Think.COM> Sender: news@cs.exeter.ac.uk Followup-To: comp.object Organization: Computer Science Dept. - University of Exeter. UK Lines: 27 In-reply-to: barmar@leander.think.com's message of 4 Nov 89 01:57:17 GMT >>>>> On 4 Nov 89 01:57:17 GMT, barmar@leander.think.com (Barry Margolin) said: Barry> You'd probably want to have a component of the component object that Barry> points back to the object of which it is a part, though. Sometimes Barry> component objects need to affect their containers. For instance, a gas Barry> tank object might want to send a message to the car's gas gauge object. Barry> The component should be prepared for that slot to be empty, Barry> interpreting that to mean that it hasn't been installed. Barry Barry> Margolin, Thinking Machines Corp. This isn't a good example in general. Monitoring facilities are probably better implemented transparently with respect to monitored objects. They should not have to "know" about that sort of thing. _Active values_ are a way of doing this. For example, see: Stefik, M. J., Bobrow, D. G., and Kahn, K. M. (1986) "Integrating Access-Oriented Programming into a Multiparadigm Environment," in IEEE Software, January 1986, pages 10-18. JT -- ______________________________________________________________________________ | Jason Trenouth, | JANET: jtr@uk.ac.exeter.cs | | Computer Science Dept, | UUCP: jtr@expya.uucp | | Exeter University, Devon, EX4 4PT, UK. | BITNET: jtr%uk.ac.exeter.cs@ukacrl|