Sideland stability

[brittanikedge]

Problems that have arisen with sidelands:

• Trial plots appear outside of field boundary when sidelands are very narrow
• Sidelands cannot be larger than headlands (not common use case, but we found when testing that the headlands will be forced as the same side as the headlands)
• Some farmers want to be able to specify that one of the sidelands will be a certain size. This is a challenge presenting this option in a way that isn't confusing to those who don't want it on the website.

[tmieno2]

@brittanikedge, can you give me an example code of the first case?

[brittanikedge]

@tmieno2 Of course, first I need to see if I can replicate this issue with the package.

[brittanikedge]

@tmieno2 I was able to recreate the errors I saw before with the following combinations:

1. When the sideland is larger than the headland, the code makes everything the size of the sideland. For example, a 30-foot headland and 60-foot sideland will become all 60 feet in the trial design. I don't know that this is something people need, but Scott brought this to my attention.
2. I used 60-foot plots, 60-foot headlands, a 60-foot applicator, and a 30-foot harvester. I found that any sideland less than 9 feet resulted in plots outside of the boundary area.

[tmieno2]

@brittanikedge, the problems are fixed. But, several things to note.

First, side and head lands are clearly defined only when the field is perfectly rectangular. On the website, I think it is important to let the users know somehow it is a rough (or approximate) length on the sides and top/bottom. They typically should pick a number slightly larger than the acceptable length. If some farmers are really picky about where the edge of the experiment boundary is, then specifying with a single number is not possible. In such case, they need to give us the exact line that represents the edge of the field on ONE side (two sides cannot be respected as plot_with governs the other end when one end is fixed). You should discuss this with Dave. If necessary, I can code that in.

Second, we have never had exactly the side_length and head_length specified by the user (especially the side land, headland should be very close as long as a field is rectangular) reflected on the trial design map.

For now, please double check if the problems above do not happen and close this issue if there is no problem. I only checked with the cases you gave me. You can create a new issue on the feature above if it turns out to be useful.

[tmieno2]

@brittanikedge, have you had a chance to look at this?

[brittanikedge]

@tmieno2 Yep, I actually tested out another case today to see what would happen with minimal headlands and an angled field. This all looks good. I feel comfortable closing it now.