CFG for lexical analysis

Question

A context free grammar can be used to model the lexical concerns of a HLL. This is not normally done and a regular grammar is used for the structure of lexemes because
(A) The cfg will blow up unnecessarily
(B) The structure of lexemes can be described by the simpler regular grammar
(C) There is a separation of concerns which controls complexity.
(D) All of the above

Answer 1

Ans should be D) All the above

A) yes it is true , as we know regular languages are implemented by finite automata which is more efficient to implement than context free languages implementation (using push down automata which is used stack ) .

B) structure of lexeme usually  be in letter(letter+digit )* and regular grammar describe it .

c) same reason as option a) .

for more detail ref: http://www.iith.ac.in/~ramakrishna/Compilers-Aug15/slides/02-lexical-analysis-part-1.pdf

Comments

Plz describe (C) option. What is the separation here?

---

Separation means- Lexical analysis phase takes care of tokenization and parsing phase takes care of syntax checking. i.e., the two tasks of tokenization and parsing are separated - makes software maintenance and development easy- in object oriented model we get 2 objects here, one for lexical analysis and one for parsing. 

In option A, it says that if we don't use a separate lex phase, our CFG will be much bigger and this can increase the parsing time. CYK algorithm for parsing runs in $\Theta \left(n^3 . |G|\right)$, where $n$ is the length of the parsing string and $|G|$ is the size of the grammar in CNF form. https://en.wikipedia.org/wiki/CYK_algorithm

---

Thank  you sir! :)

Answer 2

D) All of the above