Loading ...
Sorry, an error occurred while loading the content.

Re: Problem sets.

Expand Messages
  • William M. Shubert
    While I was away from my e-mail on vacation it looks like a discussion of problem sets in SGF showed up. It looks like it s over, but in any case, I have lots
    Message 1 of 5 , Jan 5, 1999
    • 0 Attachment
      While I was away from my e-mail on vacation it looks like a discussion
      of problem sets in SGF showed up. It looks like it's over, but in any
      case, I have lots of problems stored in SGF and a set of shell scripts
      to manipulate them (sorry, they're copyright by other people so I can't
      distribute them). But in any case, I thought I'd include a few to show
      how I've been doing my SGF problem sets - it really does work out great.

      The first attachment is five life or death problems in a single SGF file
      from the problem set on floppy that Kiseido sells; it is important that
      your SGF reader does NOT show the "next move" in variations! Showing
      that will make the problem too easy. The second SGF file attached is the
      first problem from "Whole-Board Thinking in Joseki" (a great book, by
      the way). This shows a problem with lots of commentary, sometimes
      several nodes in a row with no moves but just information, so you can
      step through it and each node will give you another hint as to the
      proper answer.

      Hopefully these examples will be useful to anybody trying to figure out
      how you can nicely put problems into SGF files.
      --
      Bill Shubert (wms@...)
      http://www.hevanet.com/wms/
    • David Fotland
      Why not put the position in the root node? What does the SY[] property mean? Many Faces doesn t recognize it. You don t have a machine readable indication
      Message 2 of 5 , Jan 5, 1999
      • 0 Attachment
        Why not put the position in the root node? What does the SY[] property
        mean? Many Faces doesn't recognize it. You don't have a machine
        readable indication of the correct answer (like TE[2]). I think it
        will be confusing to people that they can't just click where
        they think the answer is. They have to step forward a move to
        get the A-D labels, then click on the point where the answer is.

        I prefer to put the starting position in the root node, so they
        don't see an empty board first. If they want to see labels
        for choices they can turn on next move labelling, so I don't
        put labels in the problem. I prefer to have the answers as
        direct followers of the root node, without another node
        in between. I like having an property to indicate correct
        answer, so the computer can keep track of which problems
        have been answered correctly already. This enables the
        program to show the percent correct, and to avoid repeating
        problems that have already been solved correctly.

        David Fotland

        >
        >Hopefully these examples will be useful to anybody trying to figure out
        >how you can nicely put problems into SGF files.
        >--
        > Bill Shubert (wms@...)
        > http://www.hevanet.com/wms/
        >
        >
        >
        >Attachment Converted: "c:\eudora\attach\tsumego-1,5.sgf"
        >
        >Attachment Converted: "c:\eudora\attach\1-1.sgf"
        >
        >Attachment Converted: "c:\eudora\attach\wms3.vcf"
        >
      • William M. Shubert
        SY was once part of the standard. You are right, I should get rid of it since it is no longer in the standard. Also, you are right, a machine readable
        Message 3 of 5 , Jan 5, 1999
        • 0 Attachment
          SY was once part of the standard. You are right, I should get rid of it
          since it is no longer in the standard. Also, you are right, a machine
          readable indication of correct answer would be nice.

          The rest I think is just personal preference. In the tsumego problem I
          don't put anything in the root node because there are several problems in a
          single SGF file. I do it this way because I like it this way, and I wrote
          the problem-format-converter for my own use. The joseki problem has nothing
          in the root node because it is essentially a game frozen a few moves in,
          and games usually start with an empty board. Again, I did it this way
          because I like it - sometimes I want to back up and see the order of the
          moves as they appeared, because sometimes I can't figure out how the board
          got into this position. As for having to step forward to get the labels, I
          like that because I pick my move first without the labels, then if there is
          no label where I picked, I know I am way off and re-think the problem
          knowing that the solution is one of the labelled points. I never click
          where the answer is, when I do these problems I never use the mouse at all.
          With cgoban you can navigate the variations very easily with just a
          keyboard. Again, it is this way because I did it for myself and I like it
          this way.

          Basically, except for SY and TE, I don't think you can call the way I like
          it or the way you like it "wrong" or "right". Both ways you have a problem
          and a solution. Little things like whether the board appears in the root
          node, and whether you step to see the options or see them right away, are
          rather nitpicky.

          Maybe on the SGF web pages we could have a collection of problems, not
          really to solve, but just to see the different ways that problems may be
          presented with SGF? That could be useful both as templates for people
          trying to build problem sets, and also as test cases for SGF viewers so
          that programmers could make sure that their viewer/editor performs as the
          SGF author intended for every case. Does this sound good? If enough other
          people like this (and will contribute their own SGF problems done the way
          they like), then I'll clean up mine to get rid of SY and add TE and repost.

          David Fotland wrote:

          > From: David Fotland <fotland@...>
          >
          > Why not put the position in the root node? What does the SY[] property
          > mean? Many Faces doesn't recognize it. You don't have a machine
          > readable indication of the correct answer (like TE[2]). I think it
          > will be confusing to people that they can't just click where
          > they think the answer is. They have to step forward a move to
          > get the A-D labels, then click on the point where the answer is.
          >
          > I prefer to put the starting position in the root node, so they
          > don't see an empty board first. If they want to see labels
          > for choices they can turn on next move labelling, so I don't
          > put labels in the problem. I prefer to have the answers as
          > direct followers of the root node, without another node
          > in between. I like having an property to indicate correct
          > answer, so the computer can keep track of which problems
          > have been answered correctly already. This enables the
          > program to show the percent correct, and to avoid repeating
          > problems that have already been solved correctly.
          >
          > David Fotland

          --
          Bill Shubert (wms@...)
          http://www.hevanet.com/wms/
        Your message has been successfully submitted and would be delivered to recipients shortly.