Re: Why separate Lexical & Parser Generators

John Heron <heronj@smtplink.NGC.COM> writes:
>Pardon me if this question is naive. Why have a separate parser generator
>and lexical analyzer generator? ...


adrian@platon.cs.rhbnc.ac.uk (A Johnstone) writes:
>2. Efficiency: recursive descent parsers like the ones generated by
>PCCTS, PRECC, RDP and CoCo tend to do a lot of function calling.
>That's OK at phrase level, but down at the character-by-character
>level a monolithic lump of code is what you need using iteration rather
                                                                                                                ^^^^^^^^^
>than recursion. Anyone who doubts the importance of efficient scanning
            ^^^^^^^^^
>should profile their favorite compiler. You'd be surprised how much of
>the time many compilers spend in the scanner.


Haven't compiler people ever heard of tail call optimizations which
automatically turn recursions of this sort into iterations?


>5. Cleverness on the part of the language designer: anybody looking
>for a really frustrating student project might consider building a
>parser for Occam, in which there are no explicit brackets for compound
>statements but nesting is indicated by indentation.


Cut yourself on Occam's razor, I see. :-)


            Henry Baker
            Read ftp.netcom.com:/pub/hbaker/README for info on ftp-able papers.
--

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