revisit multimethods vs protocols in hoptree namespace

[robinkraft]

See discussion of multimethods and protocols here in the hoptree namespace for calculating pixel neighbors, particularly Sam's comment and Dan's reply.

[danhammer]

I will wait to do this until we start to work on updating the neighborhood feature collection workflow; it's a little irrelevant until then.

[robinkraft]

Takes about 6 minutes on my machine, mostly due to the neighborhood processing.

On Jul 17, 2012, at 2:50 PM, Dan wrote:

I will wait to do this until we start to work on updating the neighborhood feature collection workflow; it's a little irrelevant until then.

---

Reply to this email directly or view it on GitHub:
#69 (comment)

[danhammer]

What takes 6 minutes? This is a good benchmark, since the hoptree namespace isn't used anywhere yet. The idea is to speed up the process, and deal with the edge pixels. The importance of the edge pixel problem cannot be overstated.

[robinkraft]

Oh, whoops, I'm not sure why this was added to this comment stream. I thought I was replying to you directly. The thing that takes ~6 minutes is the full workflow in formarunner with 2 pixels.

On Jul 17, 2012, at 3:40 PM, Dan wrote:

What takes 6 minutes? This is a good benchmark, since the hoptree namespace isn't used anywhere yet. The idea is to speed up the process, and deal with the edge pixels. The importance of the edge pixel problem cannot be overstated.

---

Reply to this email directly or view it on GitHub:
#69 (comment)

[danhammer]

Oh, I see. The tiny data set gets hung up on the neighborhood processing. Very slow. We'll speed this up in the next couple of weeks.