triangle: incorrect test in some tracks

[kytrinyx]

Please check if there's a test that states that a triangle with sides 2, 4, 2 is invalid. The triangle inequality states that for any triangle, the sum of the lengths of any two sides must be greater than or equal to the length of the remaining side. If this doesn't affect this track, go ahead and just close the issue.

[NobbZ]

There is such a test which demands a failure. AFAIR I just took this from JSON or another track. Is it enough to just remove that test or shall I rewrite it to test for a specific kind?

I will tackle this after #84 is finished and merged.

[kytrinyx]

Oh, good question. I think we should have it return an isosceles triangle -- it's a good edge case to have.

[NobbZ]

Hmm, I just asked Wolfram alpha, it does classify as degenerated.

So I'd go for either taking further research when a triangle is degenerated
exactly and add it as another possible return value for kind. Or leaving as
is (error) or do classify as isosceles, because this is what comes when you
simply change the inequality.

Since we have 1 o'clock in the morning here, I'll go to bed now, but I'll
do some more research tomorrow and open a ticket in xcommon (or chime into
an ongoing one if there is one).

Katrina Owen notifications@github.com schrieb am Sa., 30. Jan. 2016 00:58:

Oh, good question. I think we should have it return an isosceles triangle
-- it's a good edge case to have.

—
Reply to this email directly or view it on GitHub
#85 (comment).

[masters3d]

This is very interesting. You are right.
2,4,2 is not a triangle because since 2 + 2 equals the length of the 4, then the 2 + 2 sides merge into the same plane so you end up with only a line.. (Technically is is three lines that occupy the same space of the 4 length)

[kytrinyx]

So I'd go for either taking further research when a triangle is degenerated exactly and add it as another possible return value for kind.

So wikipedia is wrong, huh. Imagine that! OK, let's do some more research.

[hankturowski]

Wikipedia is not wrong. The 2,2,4 is a degenerate triangle, but still a triangle. In either case, we should call out our definition clearly in the .md.

[behrtam]

So for a degenerate triangle (e.g. 2,2,4) which looks like a line segment do we return "isosceles" as kind or "degenerate"?

[hankturowski]

My opinion is that we either add degenerate as a case, or specifically say in our problem definition that degenerate triangles should be rejected.

To be honest, rejecting degenerate triangles feels like a cleaner solution. I'd just like the problem description to get an update.

[behrtam]

triangle.md could really need some improvement. I think it should also mention the 3 kinds.

+1 rejecting degenerate triangles

[masters3d]

++ for adding degenerate as a case

We can then separate the math definition from the notion of accepting or rejecting as a triangle, thus moving the conversation away from declaring what is or not a triangle, to just stating the type of case.
(I personally think degenerate triangles are not triangles and should be rejected! they are just lines! ;) "the emperor has no clothes" yo! ).

[yurrriq]

It seems like the general vibe here is to reject degenerate triangles and update the .md accordingly, yeah?

[yurrriq]

👍

[NobbZ]

So actually, we leave all the "faulty" tests as they are and repair the
"correct" ones?

Which again would mean I don't need to do anything here?

Katrina Owen notifications@github.com schrieb am Sa., 30. Jan. 2016 20:25:

yeah, agreed

—
Reply to this email directly or view it on GitHub
#85 (comment).

[yurrriq]

I'm not sure I understand your double quotes, @NobbZ, but the xerlang triangle tests look good as they are.

[kotp]

I say that we do exclude examples that stretch the definition to an absurd[1] degree. We keep to general position not special position geometry, and keep this where it should be in the context of the program here. If the assumption is that Triangles have area, this goes against that assumption, and is suitable for our purposes, and keeps everything at a non-college level, and is reasonable.

We don't test for the other more commonly known triangles, which is to say "right, acute or obtuse" why are we jumping into special position before we talk about all the general position triangles?

This is (as far as I can see) meant to be a simple exercise, even more so since we test for only 3 specific triangles.

[kytrinyx]

Yeah, having slept on this I think that the right approach is to avoid the edge cases and stick to the basics (having things that obviously violate the triangle inequality thing is great, the distinction between not-a-triangle and yes-but-degenerate seems like a tangent that isn't useful here.

[kytrinyx]

I've summarized this issue here: exercism/problem-specifications#202

I am closing this issue. If necessary I will open a new issue once we have come to an agreement about how to resolve the inconsistencies and what exact steps to take.