Hi Art: I'm trying to understand your letter. Your design sounds very intrigueing. In fact, I've often thought of carving a map of the state of Arizona onto the dial face, with Tucson at the center of the dial. All the hour lines would radiate out from Tucson. With my vertical pointer in the center and the correctly oriented map on the face, you could point to any place within the state, like a boyscout does with his map and compass. I could even put latitude and longitude lines on the map.
But I was going to use my cable coaxial gnomon, with a sphere on the cable to serve as the nodus for date readings instead of a vertical style with nodus sphere at tip. Your design would place Tucson north of center, and the hour lines would radiate from a point south of Tucson (probably in Mexico!). So with a vertical gnomon you would lose the ability to use the poiner to take a compass bearing. The following I understand: >A "normal" sundial has the gnomon coaxial with the Earth. This is >done to keep the errors with respect to clock time to a minimum during >the course of the year. If we have the ambition to make our sundial >read clock time to better than +/- 15 minutes, then we have to correct >for the Equation of Time. There have been many public discussions >here and private ones in my head about the best way to do this. >Simply reading a table or graph is inelegant and subject to errors of >sign. But when you say: >Methods which use the declination of the sun, either by using a >specially shaped gnomon or by observing the shadow of a nodus, rather >than an edge, are perhaps more esthetic, but they are inherently >ambiguous at the solstices and double-valued the rest of the time. What do you mean by "inherently ambiguous at the solstices and double-valued the rest of the time? This part I understand: >One way of thinking about the nodus methods (which has come up here in >discussions of the EoT with respect to leap years) is that the >declination tells you what the date is, and the figure-eight-analemma >allows you to find (with the restrictions mentioned above) the EoT for >that date. It seems reasonable to suppose that everybody has a pretty >good idea of the date already, so we are making the sundial do >unnecessary work. But this I don't: >If we make the user do this work instead of the >nodus, the figure-eight can be unfolded and made unambiguous. Are you talking about a moveable nodus? And from here on I'm completely lost! I can't imagine what the face might look like, or the gnomon, let alone how you would calculate such a dial. Wish I could see a picture! >(I am >sure I have seem such a dial design somewhere, but I can't remember >where.) For example, the date-lines can be made concentric circles, >from Jan 1 innermost to Dec 31 outermost, and the EoT for each hour >marked as a nearly radial wavy line. We could even trivially >accommodate the change between standard-time and daylight-saving-time. >As a practical matter, I think it would be easier to read clock time >(corrected for EoT) from such a dial than from any alternative. The >freedom opened up by this arrangement is astounding: The date-lines >can have (nearly) any shape, and they could all have different shapes >(as long as they don't cross). The gnomon need no longer be parallel >to the Earth's axis; it doesn't even have to be straight! > >I could envision a Salvador Dali sundial, but maybe I should start >with something for John Carmichael: Draw the outline of Arizona many >times at different scales and put them inside of one another, but so >that all the Tucsons overlap, and of course properly oriented with >respect to the compass. Put an obelisk at the location of the >Tucsons. (How can an obelisk be movable?) Label each outline with a date and calculate (the hard >part!) where the shadow of the obelisk will fall across that outline >for that date and each hour of the day. For each hour, connect the >points for all the dates and label the resulting wavy line with the >hour. Voila! > >I'd love to do the design myself, but realistically I know I won't >find the time any time soon, so I'd rather through the idea out to the >world. Is the description clear enough? (The idea is probably >between 500 and 2000 years old anyway.) > >Regards, > >Art Carlson > If you design it, I'll build it (if it works!) John
