Student Solves Sol Lewitt Drawing Assignment in Pseudocode | Hacker News Reader

Student Solves Sol Lewitt Drawing Assignment in Pseudocode

cmuems.com

76 pointsqdot7636725 comments

He's really quite shameless and condescending about singling her out for praise and elevating her above all of the other students:

> Melanie Nailed It

> (diagrams showing wrong answers of other students)

> (diagram labeled with "Lewitt" and "Melanie")

> If I were Pace Gallery, I know whom I would hire.

> each of the corners of Lewitt’s quadrilateral was reliably located by about a quarter of the students

> marvelous pseudocode system

> As far as I’m concerned, Melanie has earned the right to title her pseudocode as she did [SOLLEWITTFUCKYOU]. This document is really a gem

> (entire homework answer)

ick ick ick. It's like when a teacher photocopies your essay and hands it out to everybody else in the class as an example of stellar writing, except now it's on the internet.

What pedagogical purpose does this serve? I'm a fan of tough love from teachers. I'd rather get a low grade and learn a great deal than a high grade and learn nothing. For some instructors, part of that is being willing to say the equivalent of "this is complete shit" when your work is, in fact good, but not up to standards.

But the best ones do it in private, or in written feedback, or generally to everyone as a motivational tactic. All of that I'm willing to tolerate, and I've seen some (not many) user acerbic brutal feedback to achieve results out of me I'd never find possible.

But never have I seen a teacher signal out particular students, compare their work openly, while holding up other particular students as examples of what not to do.

Is that good or bad? The principle of rewarding individual excellence is good, and the attitude could even be forgiven in certain circumstances.

But there's a way to do that without being a dick. And if you're going to be a dick, you better b the extremely rare person that knows how to be a dick properly and effectively, usually with an undercurrent of good humor and in the pursuit of ultra-high standards.

Keep in mind: This class is part of a conservatory-style BFA program. These students are already, or will soon be, custom to public critiques. And learning how to accept public criticism, positive and negative, is explicitly a goal of many such programs.

My two cents: While I would have hated such a thing in an engineering class, I really appreciated public discussion and criticism of my work in the arts. I found that such transparency, even when painful, was hugely valuable to me as I grew as an artist.

Yes, that's a good point. I've certainly never examined the problem through the lens of a BFA conservatory-style program. If a necessary part of the skill-set a student wishes to learn is the ability to accept and process public criticism, it certainly makes sense to emulate that in the classroom.

I ran OCR over the text (using the first website I could find that takes image urls directly), so you don't have to.

A quadrangle which is formed and enclosed by four lines, the first of which is drawn from a point halfway between a point halfway between the center of the wall and the upper left corner and the midpoint of the left side and the upper left cor- ner to a point halfway between the midpoint of the top side and the upper right corner, the second line from a point halfway between the start of the first line and a point halfway between the midpoint of the top side and the upper left corner to a point halfway between a point halfway between the center of the wall and the lower left corner and the mid- point of the bottom side, the third line from a point halfway between a point halfway between the start of the first line and the end of the second line and a point halfway between the midpoint of the left side and the lower left corner to a point which is on an axis between the lower left corner to a point halfway between the midpoint of the right side and the upper right corner where a line drawn from the center of the wall to a point halfway between the midpoint of the right side and the lower right corner would cross that axis, the fourth line from a point equidistant from the end of the third line, the end of the second line and a point halfway between a point halfway between the center of the wall and the midpoint of the bottom side and a point halfway between the midpoint of the bottom side and the lower right corner to a point halfway between the start of the second line and a point where a line would cross the first line if it were drawn from the midpoint of the right side to a point halfway between the midpoint of the top side and the upper left corner.

-Sol LeWitt, 1974

BTW, the text as given on the page is incorrect and cannot be parsed. It also doesn't match the solution.

more readable image: http://cmuems.com/2012/a/files/2012/08/lewitt-instructions.j...

I can't fully parse the first point of the first line, the part in quotes below. I have added parenthesis on the parts that I can parse:

... from "a point halfway between (a point halfway between (the center of the wall) and (the upper left corner)) and (the midpoint of the left side) and (the upper left corner)" to ...

There are 4 points and only two operators, i.e., "a point halfway between". The pseudocode given has used three "halfway between" operators on these four points (on lines 5, 7 and 9). I don't know how that interpretation was arrived at.

The first line is given in the text as: "... first of which is drawn from a point halfway between a point halfway between the center of the wall and the upper left corner and the midpoint of the left side and the upper left corner to a point halfway between the midpoint of the top side and the upper right corner".

Given that the line has a FROM and TO point, we can extract those:

  FROM
    a point halfway between a point halfway between the center of 
    the wall and the upper left corner and the midpoint of the left
    side and the upper left corner
  TO
    a point halfway between the midpoint of the top side and the 
    upper right corner

The "halfway between" in each FROM and TO part must specify two points to be between, so we can extract those:

  FROM
    a point halfway between:
      1. a point halfway between the center of the wall and the upper 
         left corner and the midpoint of the left side
      2. the upper left corner
  TO
    a point halfway between:
      1. the midpoint of the top side
      2. the upper right corner

We do this once again for the nested "halfway between":

  FROM
    a point halfway between:
      1. a point halfway between:
         a. the center of the wall
         b. the upper left corner
         c. the midpoint of the left side
      2. the upper left corner
  TO
    a point halfway between:
      1. the midpoint of the top side
      2. the upper right corner

This results in a nonsensical parsing ("between" can't have three parameters), so we need to backtrack and reconsider other options. This is English and unfortunately the specification isn't precise. The only one that makes sense is therefore:

  FROM
    a point halfway between:
      1. a point halfway between the center of the wall and the upper 
         left corner
      2. [a point halfway between] the midpoint of the left side and 
         the upper left corner
  TO
    a point halfway between:
      1. the midpoint of the top side
      2. the upper right corner

This yields the final definition of the first line:

  FROM
    a point halfway between:
      1. a point halfway between:
         a) the center of the wall 
         b) the upper left corner
      2. a point halfway between:
         a) the midpoint of the left side
         b) the upper left corner
  TO
    a point halfway between:
      1. the midpoint of the top side
      2. the upper right corner

I leave this as an exercise to the reader to parse the English using greedy vs. non-greedy regex statements and see which one results in the correct outcome.

I came to the same conclusion.

Now I'm expecting someone to make a parser/interpreter and a Sol Lewitt image encoding format using the genetic algorythm that turn images into semi transparent polygons :).

Well I was a bit borred today, I made the Parser/Interpreter part, I can turn the instructions to a SVG document:

https://github.com/guillaume86/SolLewitt

I came to this conclusion at first, but I'm pretty sure it's incorrect (blue in [1]). See Camillo's comment[2] about "a point halfway between" being missing, and yohui's reply. I modified my construction to match, and now it seems right (red).

[1] http://imgur.com/U0sIlNE (based on my own parsing)

[2] https://news.ycombinator.com/item?id=6319414

I love the filename: SOLLEWITTFUCKYOU

http://i.imgur.com/ZprOYsV.png

FOURTH LINE is missing a colon and there are dashes missing all over the place! That pseudocode would never compile!

Well, as I like to say, pseudocode is for pseudoprogrammers

But in this case it it appropriate

*pseudocompile

That's just a normal school day at Carnegie Mellon University. :)

She just analyzed the syntax tree of the sentence. Where I'm from, that's a normal school day in middle school. Of course this is a long and convoluted sentence, but the process is not more complicated, it's just longer.

Anything sounds trivial if you start the sentence with '(S)he just ...'. And you say that as though a few dozen bright students didn't fail to accomplish the same task.

They failed because the instructions are incorrect and cannot be parsed (assuming, as is reasonable, that they were given the instructions seen in the blog post). It's quite possible that the only student who succeeded did so because she misread the first sentence in just the right way. Note that the assignment has "the first of which is drawn from a point halfway between a point halfway between the center of the wall", while the student's version repeats "halfway between a point" twice, which makes the sentence parseable.

What really surprises me is that nobody reported the mistake to the professor.

> What really surprises me is that nobody reported the mistake to the professor.

I think someone else in the class did:

> Michelle has asserted that the statement, as it appears in the Pace catalogue, only makes sense when an additional phrase is added.

Which linked to: http://cmuems.com/2013/a/michelle/08/29/quadrangle/

Worth noting that Lewitt had draughters construct his work from instructions, as the students have done in this task.

I've stood in a gallery and followed a different Sol Lewitt location drawing before and it had a similar error - an instruction missing.

I wonder, then, if these are deliberate omissions on the part of Lewitt?

Here's an Irony Grammar [1] that can parse the proposed problem if the missing "a point halfway between" is added. And a screenshot of the Grammar Explorer [2] for the lazy ones.

It was fun to write :).

[1] https://gist.github.com/guillaume86/6423775

[2] https://dl.dropboxusercontent.com/u/177896/SolLewitt.PNG

Could someone who understands art elaborate on why both the original work and this are important?

The classic description of this kind of thing is "the idea becomes a machine that makes the art", which is what Sol Levitt said about his work. It's a conceptual art strategy: instead of being about subjectivity, skill or the personal will of the artist, this kind of artwork is about rationality, objectivity and a neutral, abstract kind of mental process. Nothing particularly special about the pseudocode but it comes from Golan Levin's class: he is a a leading algorithmic artist (using code to make visual art). Lewitt's work could be seen as a forerunner or incunabulum in this field. http://www.moma.org/learn/moma_learning/themes/conceptual-ar...

Reply on news.ycombinator.com