This change seems unnecessary; a template parameter can be used in a non-defining declaration too.

But the example doesn't do that. The example should help make things clear, not list unrelated, though un-wrong, statements. Should we maybe delete the sentence?

Or we could say "in the declaration or definition".

All definitions are also declarations...?

@timsong-cpp: Yes, but not all declarations are definitions. To repeat, I think the example should be useful, otherwise we can just remove it. This change is an attempt to make the sentence relevant to the example and help the illustraion, rather than be a general statement that happens to be true.

@timsong-cpp: Let me try again. I think that the value of an example lies in its power of illustration. That is, specification should be general, and illustration should be specific. You want to give the reader as much insight into the details as possible. For example, a declaration is valid C++, so you could also say, "The following code might appear in a C++ program." That would also be correct, but it would illustrate less. I'd like the explanation of example code to point out as specific a level of detail as is relevant to the subject it's an example of. Does that make sense?

The "WILL" part seems dubious to me. I'd go for "MAY be used in the declaration", because the sentence does sound like a general-ish description of the meaning of the "template <class T>" prefix to me, even though it appears in an example.

@Eelis: How about "is", since the example is indeed using it?

It is indeed using it, but that fact is not specified by the prefix. The prefix merely specifies the possibility. So if the goal of the sentence is to describe the meaning of the prefix, that's what it ought to say. Just my interpretation :)

@Eelis: Yes, I agree. "May" sounds good.

This sentence seems strange; it appears to be suggesting that a List template might only work for some finite set of specified types. Maybe we could remove this (or rephrase it) rather than incorporating it into the somewhat-more-reasonable Array example.

How about:

A List template or an Array template might provide a common definition for, respectively, a dynamic linked list or a contiguous array of elements of an arbitrary type. The Array template might be [...]

@zygoloid: Ping

I think it's lacking something -- I think we want to more explicitly say that a single List template can simultaneously provide definitions for lists of type T, for all possible types T.

@zygoloid: How about:

A single class template List might provide an unbounded set of class definitions: one class List<T> for each type T, describing a linked list of elements of type T. Similarly, a class template Array describing a contiguous, dynamic array might be defined as follows:

Sure, let's go with that.

Done, now it looks like this:

"specifies that a template is being declared and that a type-name T may be used in the definition" simply doesn't make much sense to me, sorry. Similar to the argument for "may" here, at this point in time we don't know if the prefix is being followed by a definition or a (non-defining) declaration yet, so I still think "declaration" makes more sense. But at least the two highlighted words should match.

OK, you're right, I'll back those changes out for now. It seems like the status quo is fine, and this isn't making it better.