mcgregor@hemlock.Atherton.COM (Scott McGregor) (04/19/91)
>>1) Is, as is the authors claimed above, a pictorial programming language >>"*equally* accessible to users from all linguistic backgrounds"? I wrote the above in response to an extended quote from Cox & Pietryzkowski's "Using a Pictorial Representation to Combine DataFlow and Object Orientation in a Language-Independent Programming Mechanism". I am now thinking that it was perhaps unfortunate that this got mixed in with other discussion on whether text strokes or pictures were more universal. Perhaps in fact the quote was insufficient context for a useful discussion. Maybe you have to read the paper, and see their pictures (unfortunately, their pictures are not reproducible in this medium which is too bad since then everything is obvious :-) C&P were not discussing whether a red octagon was understood in more cultures that the series of strokes "stop" is. Nor is this about whether languages such as Pascal which have keywords such as "while" would be more quickly learned than languages by non-ENGLISH speakers vs. languages such as APL which do not. Rather they were making a claim more like whether reverse-polish notation is easier to understand than algebraic notation. Their claim is that textual languages are biased in favor of using variables to represent dataflows, and that this use of variables is bad for non-Indo-European language users, because non-Indo-European language users are not used to the ability to compose "meaningless strings". The claim that by chosing a pictorial representation, they can avoid use of the notion of variable to indicate a dataflow. They claim that this is better for the non-Indo-European (non-variable) user, but also that it is no worse for Indo-European users, since variables are not critical to Indo-European language usage. It is this attribute that they raise which they argue for the advantage of pictorial languages. They make no claims about the correctness of the graphical items that their language uses--they assume that all users will have to learn them from scratch just as they would have to learn the primitive operators in APL or C from scratch. The closest argument of this kind that I am aware of in the past is over whether RPN or algebraic notation is more natural for calcualtor users. I am interested in comments on the C&P thesis. A copy of their paper is in Ephraim P. Glinert's "Viual Programming Environments, Paradigms and Systems" published by IEEE Computer Society Press Tutorial Is their premise that variables are a difficult concept for non-Indo-European language users correct? Is the reason because meaningless strings can be created more easily in alphabetic languages? Or can variables be understood in terms of in terms of only meaningful strings? In fact, in most computer uses of variables, aren't meaningul names prefered? Are there important examples of textual computer languages which don't use variables? Finally, if C&Ps thesis is true, does a nonvariable based language really equalize things for all cultures? Or does it tilt the playing field against Indo-European users? Would you want to use such a non-variable system? Scott McGregor Atherton Technology mcgregor@atherton.com