> Let us be friends and hug each other.

Thanks.  I apologize to everyone (Of course,  James, Ondřej, and you are
included) for being too offensive
against Ondřej's argument and wasting your time.  I'll do more constructive
discussion.

2013/4/2 James Pike <jp...@chilon.net>

> Let us be friends and hug each other.
>
> On Mon, Apr 01, 2013 at 09:44:02am -0430, Juancarlo Añez wrote:
> >    So...
> >    You're just long-time enemies,
> >    and I'm wasting my time in this forum.
> >    I'm translating 600K lines of COBOL into C# in under 4min per CPU
> core,
> >    and that means nothing.
> >    Onegai shimasu,
> >
> >    On Mon, Apr 1, 2013 at 9:14 AM, Kota Mizushima <[1]mizuk...@gmail.com
> >
> >    wrote:
> >
> >    > And that was precisely my point. You did not mention that this is
> >    most
> >    > important part (Name of section is "3.3 properties").
> >    > Also you did not mentioned counter in paper.
> >
> >    Yes. This is not most important part in our paper since the notion of
> >    our cut operator
> >
> >    and its effectiveness were described in our earlier work. Â However,
> >    our earlier paper
> >
> >    has been written in Japanese. Â Then, the overview of cut was needed.
> >    Â That's all.
> >
> >    And I think that it's not need to mention counters in our paper since
> >    it's **trivial** optimization
> >
> >    and the all implementation detail cannot be written in our paper that
> >    has page limit.
> >
> >    > It looks that main idea of paper is how to insert cut.
> >
> >    It's not correct in a precise sense.  the main idea of the paper is
> >    how to insert cut **automatically* without
> >
> >    changing grammar meanings.
> >
> >    > It looks that main idea of paper is how to insert cut.
> >    > Actually that section is completely useless. When one incorporated
> >    > first, follow optimization and does not increase counter when it is
> >    > predicated then for any cut that you may insert counter is already
> >    zero.
> >    > You only increased possibility of bugs without any real benefit.
> >
> >    What do you mean? Certainly, hand-written cut operators may increase
> >
> >    bugs. Â Then, we explored the way to decrease space complexity without
> >
> >    hand-written cut operators in our paper. Â Of course, our methods is
> >    too
> >
> >    conservative to cover all practical grammars.  But, we don't
> >    say such things
> >
> >    and I've written as the followings:
> >
> >    " We believe that this problem can be solved by some extensions to
> our
> >    methods
> >
> >     (e.g. increasing the number of lookahead nonterminal expressions
> >    like LL(k)).Â
> >
> >    We intend to address the problem in future work." Â (Sorry for not
> >    continuing thisÂ
> >
> >    research since I work as a software engineer in a company).
> >
> >    Although forms of writing paper maybe poor, it seems that you
> >    misunderstand about
> >
> >    our paper and the proposed methods.
> >
> >    2013/03/31 18:19 "OndÅej BÃlka" <[2]nel...@seznam.cz>:
> >
> >      On Sun, Mar 31, 2013 at 03:06:30PM +0900, Kota Mizushima wrote:
> >      >    Hi, OndÅej.
> >      > Â  Â > I found that paper poorly writen, main idea is:
> >      > Â  Â > Create counter c. In c maintain number of pending ordered
> >      choice
> >      > Â  Â branches in
> >      > Â  Â > current derivation.
> >      > Â  Â > When c==0 then you can obviously discard entries before
> >      current
> >      > Â  Â > position.
> >      > Â  Â > This idea is hidden in paper and mentioned only implicitly.
> >      > Â  Â The idea you said is the same thing that we said in our
> >      paper. Â In our
> >      > Â  Â paper[1] Section 3.3,
> >      > Â  Â We explained our idea using "backtracking stack" and how cut
> >      operators
> >      > Â  Â reduce size of "backtracking stack".
> >      > Â  Â Counter-based approach you said is just a trivial
> >      optimization in real
> >      > Â  Â parser generator implementation.
> >      > Â  Â Actually, in our parser generator (Yapp) implementation,
> >      since preparing
> >      > Â  Â "real" backtracking stack explicitly sacrifices execution
> >      speed because of
> >      > Â  Â too much allocation and GC,
> >      > Â  Â We implemented our idea using one counter (32-bit integer) as
> >      you said. Â
> >      And that was precisely my point. You did not mention that this is
> >      most
> >      important part (Name of section is "3.3 properties").
> >      Also you did not mentioned counter in paper.
> >      It looks that main idea of paper is how to insert cut.
> >      Actually that section is completely useless. When one incorporated
> >      first, follow optimization and does not increase counter when it is
> >      predicated then for any cut that you may insert counter is already
> >      zero.
> >      You only increased possibility of bugs without any real benefit.
> >      As for real parsers do you know implementation other at yapp that
> >      implements [1]?
> >      Also experiments in [1] are java and xml. Java grammar was designed
> >      to
> >      minimize stack usage, and xml is simple. How does your results
> >      change
> >      when you parse c for example?
> >      > Â  Â [1]Â [1]Packrat Parsers Can Handle Practical Grammars in
> >      Mostly Constant
> >      > Â  Â Space (PDF). Kota Mizushima, Atusi Maeda, and Yoshinori
> >      >    Yamaguchi. [2]PASTE, June
> >      > Â
> >      Â 2010.[3][3]
> http://ialab.cs.tsukuba.ac.jp/~mizusima/publications/pa
> >      ste513-mizushima.pdf
> >      >
> >      > Â  Â 2013/3/29 OndÅej BÃlka <[4][4]nel...@seznam.cz>
> >      >
> >      > Â  Â  Â On Thu, Mar 28, 2013 at 03:55:28PM -0430, Juancarlo Añez
> >      wrote:
> >      > Â  Â  Â > Â  Â Hello Kota,
> >      > Â  Â  Â > Â  Â I'm reading your paper more carefully now (Packrat
> >      Parsers can
> >      > Â  Â  Â Handle
> >      > Â  Â  Â > Â  Â Practical Grammars in Mostly Constant Space).
> >      > Â  Â  Â > Â  Â If I understand correctly, if a cut is seen at
> >      position P in the
> >      > Â  Â  Â input,
> >      > Â  Â  Â > Â  Â then all memoization for positions prior to P can
> >      be discarded?
> >      > Â  Â  Â > Â  Â Thanks in advance!
> >      > Â  Â  Â > Â  Â Cheers,
> >      > Â  Â  Â > Â  Â --
> >      >
> >      > Â  Â  Â Not exactly,
> >      > Â  Â  Â I found that paper poorly writen, main idea is:
> >      > Â  Â  Â Create counter c. In c maintain number of pending ordered
> >      choice
> >      > Â  Â  Â branches in
> >      > Â  Â  Â current derivation.
> >      > Â  Â  Â When c==0 then you can obviously discard entries before
> >      current
> >      > Â  Â  Â position.
> >      > Â  Â  Â This idea is hidden in paper and mentioned only
> >      implicitly.
> >      >
> >      > Â  Â  Â This is interwined with transformation that allow us say
> >      that some
> >      > Â  Â  Â branches are not possible.
> >      >
> >      > Â  Â  Â How they write paper it forces you to use certain
> >      representation to
> >      > Â  Â  Â carry transformations. This is problem as your
> >      representation is almost
> >      > Â  Â  Â certainly incompatible.
> >      >
> >      > Â  Â  Â It is better to use generic optimizations (predication
> >      gives most) and
> >      > Â  Â  Â get
> >      > Â  Â  Â better space utilization for free
> >      >
> >      > Â  Â  Â See my thesis for other improvements, chapter 2.6 most of
> >      ideas apply
> >      > Â  Â  Â [5][5]http://kam.mff.cuni.cz/~ondra/thesis.pdf
> >      >
> >      > Â  Â  Â Ondra
> >      >
> >      > Â  Â  Â > Â  Â Juancarlo Añez
> >      >
> >      > Â  Â  Â _______________________________________________
> >      > Â  Â  Â PEG mailing list
> >      > Â  Â  Â [6][6]PEG@lists.csail.mit.edu
> >      > Â  Â  Â [7][7]https://lists.csail.mit.edu/mailman/listinfo/peg
> >      >
> >      > Â  Â --
> >      > Â  Â Kota Mizushima
> >      > Â  Â e-mail: [8][8]mizuk...@gmail.com
> >      >
> >      > References
> >      >
> >      > Â  Â Visible links
> >      > Â  Â 1.
> >      [9]
> http://ialab.cs.tsukuba.ac.jp/~mizusima/publications/paste513-miz
> >      ushima.pdf
> >      > Â  Â 2. [10]http://cseweb.ucsd.edu/paste2010/program.html
> >      > Â  Â 3.
> >      [11]
> http://ialab.cs.tsukuba.ac.jp/~mizusima/publications/paste513-mi
> >      zushima.pdf
> >      > Â  Â 4. mailto:[12]nel...@seznam.cz
> >      > Â  Â 5. [13]http://kam.mff.cuni.cz/~ondra/thesis.pdf
> >      > Â  Â 6. mailto:[14]PEG@lists.csail.mit.edu
> >      > Â  Â 7. [15]https://lists.csail.mit.edu/mailman/listinfo/peg
> >      > Â  Â 8. [GMCP] Compose a new mail to [16]mizuk...@gmail.com
> >      > Â  Â  Â
> >      [17]
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=mizukota@
> >      gmail.com
> >      --
> >      Interference from lunar radiation
> >
> >    --
> >    Juancarlo Añez
> >
> > References
> >
> >    1. mailto:mizuk...@gmail.com
> >    2.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=nel...@seznam.cz
> >    3.
> http://ialab.cs.tsukuba.ac.jp/~mizusima/publications/paste513-mizushima.pdf
> >    4.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=nel...@seznam.cz
> >    5. http://kam.mff.cuni.cz/~ondra/thesis.pdf
> >    6.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=PEG@lists.csail.mit.edu
> >    7. https://lists.csail.mit.edu/mailman/listinfo/peg
> >    8.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=mizuk...@gmail.com
> >    9.
> http://ialab.cs.tsukuba.ac.jp/~mizusima/publications/paste513-mizushima.pdf
> >   10. http://cseweb.ucsd.edu/paste2010/program.html
> >   11.
> http://ialab.cs.tsukuba.ac.jp/~mizusima/publications/paste513-mizushima.pdf
> >   12.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=nel...@seznam.cz
> >   13. http://kam.mff.cuni.cz/~ondra/thesis.pdf
> >   14.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=PEG@lists.csail.mit.edu
> >   15. https://lists.csail.mit.edu/mailman/listinfo/peg
> >   16.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=mizuk...@gmail.com
> >   17.
> https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=mizuk...@gmail.com
>
> > _______________________________________________
> > PEG mailing list
> > PEG@lists.csail.mit.edu
> > https://lists.csail.mit.edu/mailman/listinfo/peg
>
>
> --
> http://chilon.net
>



-- 
Kota Mizushima
e-mail: 
mizuk...@gmail.com<https://mail.google.com/mail/u/0/?view=cm&fs=1&tf=1&to=mizuk...@gmail.com>
_______________________________________________
PEG mailing list
PEG@lists.csail.mit.edu
https://lists.csail.mit.edu/mailman/listinfo/peg

Reply via email to