This post originated from an RSS feed registered with Agile Buzz
by James Robertson.
Original Post: Subsystem Exception Handling in Smalltalk
Feed Title: Richard Demers Blog
Feed URL: http://www.cincomsmalltalk.com/rssBlog/rademers-rss.xml
Feed Description: Richard Demers on Smalltalk
In the late 1970's, my job was to design the message and exception handling facilities of the operating system for the IBM System/38. The system's Virtual Machine already had support for exception handling, but the operating system and language development teams had almost no idea what to do with it. Clearly, we had to be able to handle all the exceptions raised by VM instructions, everything from computation to process management and database support. But should more exceptions be raised by operating system programs. Should applications be able to monitor and handle exceptions, or even raise new ones? And how did all of this fit into plans to support cross-language debuggers and service tools? Needless to say, I had several very interesting years working on all these questions.
Now skip ahead a decade. I finally got a chance to do some Smalltalk programming, and there was my old friend Exception Handling, but in a totally different context where it made little sense to me. No longer was I dealing with huge procedural programs that implemented whole components. Instead, there were all these little classes and their hoards of tiny methods. The questions I had previously answered in another time for another system now needed different answers. Or did they?
"Exception Handling in Smalltalk" remained a back-burner issue for me for several years. I only began to arrive at answers when I looked at the problem from a different point of view; that of design in terms of subsystems. You can read about my conclusions in a paper on
Subsystem Exception Handling in Smalltalk
that I just published in Why Smalltalk.
I hope this paper will stimulate further discussion on both exception handling and subsystems.
Previous Topic
