Popularity of compiler tools, was LRgen

"Paul B Mann" <paul@paulbmann.com> writes:
>The current status for LRGen is:
>"Not enough interest. No feedback. No money. Not worth my time".


That seems to be a common problem for compiler tools. I think there
are several reasons for this:


- There are many tools on offer, and not that many compilers being
    developed (compared, to, say the relative numbers for compilers and
    programs that use them).


- As is usually the case in computing, a few tools get most the
    interest/users/feedback/etc. In the compiler front end area these
    are yacc/bison, and maybe JavaCC or ANTLR.


- Finally, many compiler writers seem to dislike tools (or maybe none
    of the tools are good enough or something).


    In particular, while I know of several tools for instruction
    selection using tree parsing, none of them seems to be widely-used;
    many compilers use hand-written instruction selectors, and of those
    where I have heard that they use generated tree-parsing instruction
    selectors, the generator was developed or extended in-house.


    One explanation I have heard is that the compiler writers don't like
    to make themselves dependent on a tool that may go away. OTOH, gcc
    reverted from using bison-generated parsers to hand-written ones (at
    least for C++ and C), and I very much doubt that the future of bison
    was the reason for that.


Maybe some other posters can provide additional insights into the use
or non-use of compiler tools and the reasons for this.


>If anyone has generated more interest in their product by making it
>open source, let me know.


OpenOffice (nee StarOffice) and Mozilla/Firefox are two products that
come to mind. However, they also did a lot of development in addition
to unchaining their software.


If a tool is free software, that would at least be a partial answer to
the fears of the tool going away: if the author or developing company
drops it, the compiler writer can take over maintenance (possibly with
other users), so one would not be worse off than when maintaining an
in-house tool (apart from the transition cost).


>Also let me know how to get money for
>working on open source code.


Work for Red Hat, SuSe/Novell, MySQL AB etc. They probably won't pay
you for working on a product that is uninteresting to their paying
customers, though.


- anton
--
M. Anton Ertl
anton@mips.complang.tuwien.ac.at
http://www.complang.tuwien.ac.at/anton/

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