Mostly I want to reconcile the differences in how and what I was counting along the way so far, which changed several times (though not radically), but also to enumerate the various kinds of things that the people in the photographs are doing, or at least appear to be doing.
And to try to arrive at some more general result, though I'm not sure how well that's going to work out.
In the meantime, though, before I forget, I want to put forward a few of the lessons learned thus far, especially since the detailed run-through from 73rd Street up to 110th Street this afternoon has brought them home to me with somewhat unexpected force.
The lessons learned are of two kinds: those pertaining to the actual photography, and those pertaining to the frequency, to the time series.
The photography lessons are straightforward enough:
Settle on a more consistent horizontal field of view, measured in absolute rather than relative distance, i.e., the corner ±125–150 feet, say, rather than a constant ±25º–30º or so would be helpful. Wide enough that a pair of photos, e.g., the NW corner of xth street and the SW corner of x+1th street would cover the whole block. And certainly a much greater consistency than I was aiming for in 2006.
Get a more consistent white/color balance, together with a more consistent exposure, if necessary obtained by a second shot at each corner with a calibrated gray-white-black card, would make time series comparisons easier by eliminating or at least reducing these sources of irrelevant variation.
In relation to #2, then use a much more consistent processing workflow, with much less cropping.
Take the extra time at each corner to ensure that the camera is level, or nearly so, to reduce the amount of perspective correction to be done in post-processing (which adds some other geometrical distortion while eliminating the converging verticals, something I'd like to maintain).
Possibly also make a much larger lens shade to reduce the amount of flare that sometimes creeps in later in the day.
Administratively, keep closer track of the shots during each shoot to reduce the likelihood of accidentally skipping a corner.
And don't hesitate to take as many "ID" shots as possible, to avoid subsequent confusion.
shoot everything that can justifiably be called a corner, whether it meets my narrower criteria for a "valid" corner or not — inclusive is the word during subsequent shoots.
The time series related lessons (insofar as they're not already covered above) are:
Since it is unlikely that I'll be able to shoot the whole island more than once every few years — though I'd like to — to select some key areas and shoot those much more frequently, say quarterly, over a span of five to ten years — good grief, I hope I live that long!
If, as before, all daytime Manhattan, and bright sunny days at that, then all in addition all weekday afternoons, no weekends, no holidays, for the sake of uniformity in vehicular and pedestrian traffic.
I'm sure there are even more lessons, but the above list is a tall enough order for now already!