winkler@Apple.COM (Dan Winkler) (03/07/88)
Any change I make to HyperTalk from now on is guaranteed to be upward compatible. I might add new features, but I'll never break old scripts. Therefore, I'll never require variables to be declared. I may provide an optional way to declare them, perhaps with a type, so that I can do type checking for you or, when I get to writing a compiler, generate better code, but to uniformly require it would break existing scripts. Similarly, I'll never disallow unquoted literals, although I may provide a way for you to turn them off when you want to.
edwards@bgsuvax.UUCP (Bruce Edwards) (03/09/88)
In article <7574@apple.Apple.Com>, winkler@Apple.COM (Dan Winkler) writes: > Any change I make to HyperTalk from now on is guaranteed to be upward > compatible. I might add new features, but I'll never break old > scripts. Therefore, I'll never require variables to be declared. I > may provide an optional way to declare them, perhaps with a type, so > that I can do type checking for you or, when I get to writing a > compiler, generate better code, but to uniformly require it would break > existing scripts. Similarly, I'll never disallow unquoted literals, > although I may provide a way for you to turn them off when you want > to. I am perfectly willing to accept a language will a little looser way of dealing with my human tendency to leave things out, like quotes and variable definitions and I applaud Dan's stick-to-ed-ness on keeping things upward compatible. This isn't FORTRAN guys....its something new, let's give it a chance to work on us as we work with it.