JBQ’s blog
Next Page...
Hacking geohashing part 2: Centihashing
One of the issues with geohashing is that graticules are large (70 miles north-to-south, a bit less east-to-west depending on where you live: e.g. my graticule is 56 miles east-to-west). While that's a good size for an all-afternoon trip during the week-end, it's definitely too large for most after-work weekday expeditions or even casual week-end trips.

Centihashing uses the same principle as plain geohashing, but shrunk to centicules, which reduces the distances by a factor of 10, making the hashes a lot more practical to reach. The basic geohashing algorithm is trivially extended to centihashing, by preserving an extra decimal digit from the original coordinates. I'd recommend dropping a decimal digit from the hash instead of shifting digits, such that the geohash for a day is also the centihash for its centicule.

Along the same lines, I also propose a geohashing achievement for reaching centihashes in all the centicules of a graticule (as long as those centicules contain less than 50% water).
Hacking geohashing part 1: Minimal Retrohashing
This post is the first in a series that explores different meta-subjects around geohashing.

I live in the San Francisco graticule, which contains 60% water. As a result, relatively few geohashes are on dry land (it could be worse, see e.g. Santa Cruz). To take a recent example, all 6 geohashes in the San Francisco graticule between 12/31/2009 and 1/5/2010 were on water. Someone who'd hope to geohash during an extended new year holiday would be entirely out of luck. One option for such cases is to go in neighboring graticules, but that can quickly create long drives (e.g. from home, the nearest dry geohash yesterday 1/5/2010 was 75 miles away). Personally, I think that retrohashing (i.e. reaching past geohashes) is a viable approach. I do understand the concerns that retrohashing is a way to "cheat" since it theoretically allows to pick in a back-catalog of about 30000 geohashes for each graticule. Therefore, I propose the concept of minimal retrohashing, where the retrohashes must be considered in order starting with the most recent. Therefore, if all the geohashes for San Francisco for the next 3 days are wet (a 22% probability) and if I go geohashing on Saturday, I'll be attempting the San Francisco geohash for 1/6/2010 (i.e. today).
What I look for in a camera
Echoing Eugenia's post about important features in a video-recording camera, here are my priorities when looking at a still camera, in this specific order.


  1. How large is the sensor?
  2. Does it have fast prime lenses that match the sensor size?
  3. Does it have aperture control?
  4. Does it have sensitivity control?
  5. How many pixels does it have?


The first two are really important: they are the ones that most fundamentally influence the final image quality (in terms of sharpness, low-light characteristics, and background blur).

Notice how the pixel count is at the bottom of the list. Yup. It doesn't matter. If you have a camera with a large sensor, a good prime lens, aperture control and sensitivity control, you're usually fine. Get your hands on a used 5D, stick a 50/1.8 on it, set the aperture around f/7, use the lowest sensitivity you can get away with, and you'll have files that can print to just about any size.
JBQ: now on twitter
Since I obviously can't get myself to write long blog posts any more, I decided to open my own twitter account to get a taste for twitter. So far so good.

Looking back at internet fashion, I skipped Hotmail, did Geocities, skipped Myspace, Delicious and Facebook, hopped on Gmail, and now I'm on Twitter. I'm probably missing some in the list.
It's not the economy, stupid!
Today was warm in tha area, and quite a few people showed up at the local Mexican joint to have a taste of summer in the evening, on the patio by the water.

The restaurant stupidly closes at 9pm, and barely after 8pm they were turning customers away as they didn't have enough waiters left on duty (even though neither the patio nor the room was full)

Based on the number of customers they turned away, I'd guess that thye missed several hundred dollars of sales just in the 10 minutes that they had us waiting.

Next time, plan better, and you'll be helping the economy!
Next Page...
MoBits mobile web design