![[spacer]](/schemes/blue/line.gif){kind=small}
|
Andrew Skalet HW5 DLC 1.1 I thought the observation that computer critics were more effective in domain-oriented design environments was interesting. This makes sense to me, because since computer-based critics have a tendency to be disruptive, they had better have something useful to say when they do interrupt. Applying a system to a specific domain makes this much easier, I imagine. In general, I appreciated the general direction of the paper. I believe in using the strengths of people and computers together, and less in replacing people with computers. As such, I enjoy the direction of this paper as compared with the many more AI papers one sees (which aren't all bad, but mostly have the latter philosophy). 1.2 It took me a while to realize how the hypermedia element was coming into play. I was hung up on how the hypermedia system enabled experts to formalize their knowledge in a way that the critic could use it directly. Later on in reading the document, I realized that the hypermedia system was to supply rationale to the designer. Once I understood this, I really liked the idea. 2. The main message of the article to me was that computer-based critics can help both amateur and professional designers make better design decisions in domain-oriented design enviroments. Several specific features (e.g. user customization of level of critiquing, different design perspectives) can make this type of system much more useful. 3. The main thing I would like to learn more about is how later research, mentioned in the "Reflections on..." at the end, attempts to alleviate the requirements for a software expert to address each specific domain, in particular the end-user scripting language that Tamara Sumner worked on. 4.1 I really don't know anything else about DODEs. 4.2 This is also the first time I have been exposed to computer based critiquing, although I can see the paper's point that critiquing is present in all kinds of design environments. 5.1 The article says that carefully constructed software systems can help humand design more effectively. 5.2 The article shows that learning can take place directly from coded rules, and also through informal justifications for those rules. 5.3 The most interesting collaboration element of the article is that it focuses on a human collaborating directly with a computer, using the strengths of each. 5.4 The article shows that through creative use of innovative media, design-oriented information can be presented to a user in such a way to stimulate design ideas in the user. 6. I would be very curious how this type of system might be applied to the design and evolution of software. |
