Talk:Box the compass
General
The naming sequence was more complex than I originally thought. I started by coding up a generic quarter of the points but North and South take precedence over East and West in some of the orderings and so I had to 'reflect' a quarter, quarter1 to make quarter2 in the Python solution.
P.S. Couldn't help the Pirate speak :-)
--Paddy3118 21:43, 27 March 2011 (UTC)
- I chose to just put in a list of all the packed representations (which are very slightly modified from the official abbreviations to add hyphens). The fact that I don't have to do a lot of quoting of the strings makes this a very attractive option in Tcl. –Donal Fellows 10:10, 28 March 2011 (UTC)
- I like the way you've done it. Coding all 32 points is done so compactly and you skip all the logic for choosing quarters. Neat! --Paddy3118 12:53, 28 March 2011 (UTC)
- In my VB.net implementation I chose to build the list of points into an array beforehand and keep it separate from the calculation subroutine. It was implemented in a similar way to the Python routine otherwise. Admittedly, it'd probably have been more efficient just to write the list of points straight into the array, but would have been far less interesting that way! --TheMadman 13:31, 28 March 2011 (UTC)
- For J, I decided to do the simple thing, and then I fix the result by preventing the text of the first word from appearing multiple times in multi-word names. --Rdm 15:35, 28 March 2011 (UTC)
Capitalization?
The Wikipedia article capitalizes, for example, "Southwest by south". However, a google search shows that as a general rule the capitalization will be either "southwest by south" or "Southwest by South". The Wikipedia capitalization could be rationalized as the lower case variation as it would appear at the beginning of a sentence... Anyways, perhaps the task description should either explicitly require one form of capitalization (and I do not much care which, though the Wikipedia form seems awkward) or should explicitly state that capitalization is a non-issue? --Rdm 14:50, 28 March 2011 (UTC)
- The current J implementation abuses the information contained in the wikipedia capitalization to deal with one of the wording quirks. (I would prefer to get rid of that abuse but do not want to code so generally that I am explicitly parameterizing and supporting all three forms of capitalization -- at least not if the task does not require it.) --Rdm 15:10, 28 March 2011 (UTC)
- I would argue that the capitalisation is a bit of a non-issue. Strictly speaking, the cardinal directions are not proper nouns therefore shouldn't really be capitalised. I did capitalise mine in the same way as the Wikipedia entry simply because it just wouldn't have looked right otherwise, but I did make sure that the capitalisation subroutine acted on the direction names at display time, not on insertion into the array. --TheMadman 15:32, 28 March 2011 (UTC)
- It should appear as in the wp table. Other capitalisations wouldn't be following the task description. --Paddy3118 15:39, 28 March 2011 (UTC)
- Absolutely, although if you look at the example at the bottom of the page, the Wikipedia entry does put "east by north" in lower case, suggesting that it's done in the table purely for aesthetics (which, in those circumstances, is probably the correct way to write it), hence my argument of non-issue. --TheMadman 15:50, 28 March 2011 (UTC)
- I would also argue that capitalization should not be an issue. If you get concerned about something like that, you're starting to worry about the brand of mustard on the hamburger instead of the actual meat. --Mwn3d 16:11, 28 March 2011 (UTC)
- It should appear as in the wp table. Other capitalisations wouldn't be following the task description. --Paddy3118 15:39, 28 March 2011 (UTC)
J and task specifics
What should I do about the J solution. No doubt, as the comment seems to suggest, J makes sure that the compass point for all three angles in its fuller row of three headings are all the same, but I am concerned that:
- Others will just look at J's solution and not check.
- It is not strictly comparable to other entries.
- The full wp table misses out -5.62 and 365.62
For these reasons, I am inclined to stick an incomplete tag on the J example and require it to produce what was asked for; or add to their current entry another version that does strictly what was asked for. Would this be OK Rdm? --Paddy3118 15:54, 28 March 2011 (UTC)
- I think I understand your concerns, but I do not think that these are currently task requirements (I do not see anything in the task requirements specifying that the 33 numbers be isolated from other numbers -- the description does not use words or phrases like "only" or "limited to" or anything like that). Marking the J approach "incomplete", with the current task description, would be analogous to marking down a task for including extra examples. Of course, if you feel it's necessary to change the task description, I will go along with that. However, for now, I am going to add some additional examples for J to show how it works, and hopefully to prevent people from getting the wrong idea. --Rdm 16:02, 28 March 2011 (UTC)
- Hi Rdm, the task has the note on computing the index and headings to convert that isn't in your current example but is in others. --Paddy3118 19:48, 28 March 2011 (UTC)
- I believe I have the compass index in the J implementation. I believe I have included all the headings asked for in the task description. I believe the headings I use are in the same order that they are specified in the task description. So I am not sure what you meant here. Is it possible you were looking at a cached version of the page when you wrote this? --Rdm 20:26, 28 March 2011 (UTC)
- Hi Rdm, the task has the note on computing the index and headings to convert that isn't in your current example but is in others. --Paddy3118 19:48, 28 March 2011 (UTC)