Alex Rice wrote: > > On Feb 9, 2004, at 4:16 AM, Frank Leahy wrote: > > Richard wrote: >>> What is the argument against the xTalk messaging model? > > I'll keep repeating it: the argument is you don't know at until runtime > if the message goes where you think it will go. One way to state it is > "throwing a message out for some object to catch, hopefully".
The problem has been discussed in detail on the Improve-Rev list. My question was: Can you think of a way to address those concerns while retaining the essential flavor of the language? How would you like it to work, and have you submitted an enhancement request for it? > Frank wrote: >> 1) There is no convention for whether one needs to pass messages up to >> the system. For example, >> 2) It's hard to guarantee the message passing context. >> 3) The messaging model does not always work correctly. And it >> sometimes works differently depending on where the code resides. >> 4) Race conditions. > > Thank you Frank you are hitting the nail on the head. He did but looking at the explanations he provided for each, none of them were unique to xTalk. One could rephrase those as: 1. Transcript is flexible and requires me to make choices. 2. It's possible to make suboptimal choices for object references. 3. Software sometimes has bugs. (FWIW I rarely see messaging bugs) 4. The potential for race conditions exists in multi-tasking systems. -- Richard Gaskin Fourth World Media Corporation ___________________________________________________________ [EMAIL PROTECTED] http://www.FourthWorld.com _______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
