[also see ...]
[also see our article in the June 1998 Byte.]
Peter Salus explained the "History of Scripting" for the Sixth Tcl/Tk Conference.
Nicholas Petrely declared scripting fashionable in September 1998.
[ Python vs. Tcl; Scheme vs. Tcl; Java vs. Tcl (EET) ...]
[events vs. threads]
[languages; modes; DCOM; CORBA; ...]
Nat Pryce makes available a delightful work-in-progress, Patterns for Scripted Applications. "Patterns", here, is of course a keyword in the tradition of ...
The perl-porters mailing list spawned a thread in reaction to the article. Readers also sent a pile of letters to the SunWorld editors. We reply to several publicly.
A couple of readers complained that the article gave no direct reference to more information about Expect. I apologize. We also failed to give a pointer to Lua's home page, which it certainly deserves.
One gentleman argued that Haskell isn't a scripting language. Perhaps so; I understand the point. On the other hand, with availability of the Hugs implementation, it looks like scripting to me. I leave it to readers to decide for themselves. The same reader asked for links to Limbo and Taos.
Another reader made a strong pitch for the importance of FORTH and NetREXX. The former, for example, is the monitor language of a naked SPARCstation, some Macintoshes, and so on, while the latter anticipates and even surpasses in some regards the more recent Jacl. He's right; still, their use in 1997 is more specialized than for Perl, Tcl, and Python, and we made a distinction where we thought it would most benefit our readers.
Scott L. Raney of MetaCard Corporation wrote a long and generous response to the article. Among the points he asserted is that the MetaCard equivalent of the stopwatch example
is 22 lines (vs 19 for the Tcl version), largely because of the "on message, end message" handlers. But [it] is *much* more intelligible to non-MetaTalk users, which transfers in to a much shorter learning curve. Note too, that no "syntactic gymnastics" are required in the MetaTalk version as are required in the Tcl version (especially the use of the "format" function, some references to variables require a $ and some don't, implicit globals, and the tricky distinction between references to globals that *do* require global declarations vs those that don't).Mr. Raney goes on to detail several other features he has built into MetaCard which he asserts make it superior to Tcl and the other languages featured in the article.
Also note that there is no code to construct the UI: this is all done with MetaCard's IDT (you can write scripts to do UI construction, but you almost never have to). . . .
Mr. Raney is a mildly controversial figure in comp.lang.tcl where he advertises the benefits of MetaCard ("Scripting, without all the syntactic gymnastics") too aggressively for the tastes of many. On the other hand, I appreciate his industry in making available to readers the MetaCard example above, so that you can do the comparison for yourselves.