home *** CD-ROM | disk | FTP | other *** search
Text File | 1989-10-20 | 2.6 KB | 53 lines | [TEXT/GEOL] |
- Item 1826427 19-Oct-89 20:24
-
- From: BURBECK.S Burbeck, Steve
-
- To: D2078 AS & M, Alan Dail,PRT
-
- cc: MACAPP.TECH$ MACAPP Tech
-
- Sub: Re Next IB vs ViewEdit
-
- Alan,
-
- ViewEdit is both more and less powerful than InterfaceBuilder (IB). In
- ViewEdit's favor, it manipulates resources so you can add to and modify the
- user interface without rebuilding your program. IB, like Prototyper in the
- Macintosh world, puts out source code. This approach is one-directional: once
- you export the code and edit it to add your behavior, there is little you can
- do to modify it with IB. The resource description approach is far more
- flexible. And flexibility is very important because commercial user interfaces
- typically take many iterations to get right (and users may continue to suggest
- changes through the alpha and beta testing phases). However ViewEdit is less
- powerful than IB in one important respect: it only describes the appearance of
- the user interface. You must write your own code to connect its active
- components (e.g., buttons) to your program's behavior. IB lets you describe
- both the appearance of the interface and connections to behavior.
-
- As you correctly point out, a fundamental stumbling block to describing
- behavior in ViewEdit is that Object Pascal is missing some basic power that is
- available in Objective-C. It needs the ability at runtime to send a message to
- an object where the message name is specified by data (e.g., a string in from
- 'view' resource or -- consider the possibilities -- a string entered by the
- user!!).
-
- As Geoff Pascoe explained, this is not a language feature, it is a runtime
- metadata issue. In effect you need information at runtime that, at present,
- the MPW linker has thrown away. You and Geoff note that Objective-C has (some
- of) this, whereas Object Pascal does not. We have been looking into this issue
- with the MPW engineers. No promises, but there is reason for optimism!
-
- I believe that the architectural decision to base ViewEdit on the Macintosh
- resource mechanism will, in the long run, give MacApp a fundamental advantage
- over IB. The IB scheme provides a single purpose mechanism for helping the
- programmer to build the original interface. It is of little use for subsequent
- maintainers of the program or for internationalizing or customizing the
- application after it is built. The MacApp approach, while lacking an important
- feature today, is more extensible and generalizable. We intend to extend and
- generalize it.
-
- Steve Burbeck,
- MacApp Product Manager
-
-
-