Hi, John Pane (one of the authors) has a few interesting comments. I'll put the most relevant on top:
> As for what precedence rules to adopt, despite [Spohrer 1986b], I would not > advocate for left-to-right precedence. I think it is better to be consistent > with what novice know from mathematics (if they know mathematics) (the > principle of "Consistency with External Knowledge"). Well, this is not a study, but it's an expert opinion. So, being consistent within the PL, but inconsistent with external knowledge (math as used on paper, in Excel, calculators, ...) can lead to more errors and decrease learnability. If this is not convincing enough I'll continue to post replies from other experts and profs as I receive them, but I think you can already guess what they will say. Here is the complete reply that also mentions a study about novices' understanding of parenthesis: On Dec 12, 2007 2:02 PM, John Pane <[EMAIL PROTECTED]> wrote: > Hi, > > I'm glad you find that paper useful. Unfortunately, I have no current plan > to update it. > > Here are a few thoughts on operator precedence. I assume we are talking > about novices. That's what the paper (and all of my Ph.D. research) was > about. > > Optimizing the solution to operator precedence for novices might have > negative side effects. For example, requiring all expressions to be entered > with explicit precedence could become very tedious. Thus, the solution must > be balanced with other potential usability problems. > > A modern programming environment might display the evaluation order, but not > require the user to enter it explicitly (unless the user desires an > evaluation order that is different than the built-in rules). This could be > done by inserting parentheses, or using other methods to indicate > precedence, such as tiles. > > However, novices did not understand parentheses in a study I conducted. > (J.F. Pane and B.A. Myers, "Tabular and Textual Methods for Selecting > Objects from a Group," Proceedings of VL 2000: IEEE International Symposium > on Visual Languages, Seattle, WA: IEEE Computer Society, September 10-13 > 2000, pp. 157-164. http://www.cs.cmu.edu/~pane/VL2000.html). I view this as > a very important finding that suggests that programming language/environment > designers should seek other solutions. In this study, I used a > two-dimensional way to indicate precedence, but that solution uses a lot of > space on a page of program text. > > As for what precedence rules to adopt,[...posted above...] > > Best, > John _______________________________________________ fonc mailing list [email protected] http://vpri.org/mailman/listinfo/fonc
