Is THEADS the same as FIRST?

Question

I'm reading the paper "A Practical General Method for Constructing LR(k) Parsers" by David Pager, and it contains the following paragraph:

We will first of all briefly review the LR(1) parsing and parser construction algorithms. In this paper the symbols of a grammar are denoted by Roman letters while Greek letters are employed to denote strings. $\epsilon$ is the null-string. $\alpha \Rightarrow^* \beta$ means $\beta$ is derivable from $\alpha$ (which is considered to be true if $\beta = \alpha$). By THEADS($\alpha$) we refer to the set $\{b \mid b \text{ is the head of a terminal string derivable from $\alpha$}\}$.

From what I can see THEADS here is what is usually called FIRST. Is this correct? Does THEADS predate FIRST?

Answer 1

I recently had the same question in mind when reading Pager's work. I found the paper The Edge-Pushing LR(k) Algorithm by Chen and Pager, which seems to say that the two terms are indeed the same. Quoting that paper:

The concepts of state, configuration and theads(α, k) used here are equivalent to “item set”, “item” and FIRST_k(α) correspondingly in some other literature.

Note that a lot of good references on Pager (and Chen's) work can be found at the HYACC web site.

As for why Pager uses his own terminology in his papers, I don't have an adequate explanation, but it does indeed appear that THEADS and FIRST are the same.