Re: Why are LR parsers faster than using if conditions

shripal.meghani@philips.com writes:


> Well intuitively I know that LR parsers are fast, they SHOULD be
> faster than using if then conditions etc. But I was hoping to get
> some more solid academic reasons to convince a group of people here at
> my organization...


Just describing the alternative as "using if-then conditions" doesn't
give us much to work on. Well-written parsers using if-then-else can
well be faster than LR parsers, but they can also be worse. It all
depends on how they are written.


LR parsers are a bit slower than well-written recursive descent
parsers, which is the most common systematic way of writing parsers by
hand. The main advantage of LR parsers are that they don't require as
much mangling (left-recursion removal, left factoring, etc.) of the
grammars as LL(1) parser (including recursive descent) do.


But hand-written parsers written in a non-systematic way may well be
slower than LR parsers.


Torben

Return to the comp.compilers page.
Search the comp.compilers archives again.