As you know, I take the process of REBOL naming quite serously. Ironically, coming up with the best name for a function or other structure can take more time than writing the entire function in REBOL.
The typemap! datatype is a good example, and we should discuss it right now.
Originally, I thought a better name was typeset! to be more consistent with the bitset! datatype. But, I was concerned about the fact that the word typeset has already has a firm technical meaning in English (even more to those of us who have published books before, which is my own bias).
However, that name conflict really quite irrelevant, isn't it? After all, how likely is it that typesetting will ever have anything to do with REBOL?
So, let's hear your comments on this. I'm strongly leaning toward changing the name to typeset! and now is the time to do it.
Although non natively speaking English, I like typeset! more. typemap! sound like keyboard abstraction related, and as we want better keyboard support for Rebol 3.0, let's use typeset! :)
Volker Nitsch 9-Apr-2006 16:02
I prefer typemap, because it maps values to types. but thats my german-english. With a set i expect union, itnersect etc. Thats partly the case with charset (union digits letters). But with our few types i think that is secondary. Would be different if we had user-defined types and could use it for something like subclassing (union baseclass my-type)
Ridcully 9-Apr-2006 16:22
I would prefer typeset! On one hand, as you said, it would match the bitset! on the other hand a set is an unordered bunch of items while a map is more like a key-value mapping - like a HashMap in java.
Carl Sassenrath 9-Apr-2006 16:53
They are actually more like sets than maps.
Also, it is likely R3 will support user-defined types (UDT), but not in the same way as primitive datatypes. That is, a UDT would not provide unique identification at the primitive type level. (All UDT's would be of user-type!) This remains an open issue.
maximo 9-Apr-2006 17:34
well, a list of exclusive things is a set, by definition and in many computer related softwares. a map implies two values with a mutual dependency. so I don't see why to use typemap, especially when there may eventually be a proper use for typemap!
Andreas Bolka 9-Apr-2006 17:51
+1 for typeset!
Brian Hawley 9-Apr-2006 18:05
I vote for typeset!, as it is more descriptive of its semantics. The printing industry term is archaic anyways - who uses moveable type anymore?
Volker Nitsch 9-Apr-2006 18:14
After some thinking +1 for typeset too. And a +0.5 for typefilter, as this is what they are used for: filtering what passes as arguments.
Christian Langreiter 9-Apr-2006 18:14
I'm all for typeset!.
Carl Sassenrath 9-Apr-2006 19:02
Typeset! it is. Article #0002 updated. REBOL 3.0 modification complete.
Thanks for all your comments.
Tamás Herman 10-Apr-2006 21:19
+1 for typeset! - as a non-native english speaker