Salt in Your Coffee

Chile vs Haiti earthquakes: is not 8.8 vs 7.0. Donate money to Haiti

jimmyhchan — Mon, 01 Mar 2010 18:49:30 +0000

Q: Two countries were struck by large earthquakes: one of magnitude 8.8 and one of magnitude 7.0. You are the Red Cross, who would you give aid to?

A: WHAT?! That's a ridiculous question! You want me to decide using only a mathematical number? What about telling me how many people died? or how many building collapsed? or how adequate are the hospitals—hell … have the hospitals collapsed in the quake… ahhh! were there hospitals to begin with??!?

8.8 > 7.0 right? What's the problem?

Short answer: Scientists like scien-terrific numbers but you like useful numbers.

The number in question is the Moment Magnitude, M_w (http://en.wikipedia.org/wiki/Moment_magnitude). Chile had an 8.8 magnitude earthquake. Haiti had a 7.0 magnitude earthquake. It's a beautiful number (for scientists) — every earthquake can be reduced to a single number and that number is in units of energy, the same units as nukes, car engines and chemical reactions. M_w can be generated instantly using seismographs (read the peak acceleration, velocity; move some bits; carry the 1 ==>output Magnitude). Easy.

Unfortunately, the number is completely useless to you.

The moment magnitude (8.8 for Chile and 7.0 for Haiti) is only a number which relates to the energy the moving earth can generate. This would be a useful number for harnessing "earthquake power!" or for answering "how strong is god" but not as useful for what you care about — Should I give aid to country A or country B… you know straight effects-on-buildings-and-p

eople comparisons.

**sidenote: The old Richter scale is similar but uses a different mathematical formula (That is they are different numbers). Today, all magnitudes are moment magnitudes but Richter magnitudes are still around… just to add confusion.

The problem with looking only at Magnitudes

The magnitude is just one number: the energy. More specifically the energy released along the entire fault. The magnitude of an earthquake does not tell you anything about Aunty Betsy's apartment complex 500 miles from the epicenter or Uncle Bob's high rise office built on the old landfill 20 miles southeast of the fault. Everything from the earthquakes travel to the surface will affect what it does. This includes the direction of the fault, the quality of the soil, the quality of the building, the preparedness of the people and the amount of support post-quake. Again, we care about Aunt Betsy and Uncle Bob, not alternate energy from tectonic plates.

What should we look at instead? Mercalli Intensity

The maths for comparing magnitudes is crazy and altogether not quite useful. There is another number though which gives you the effects of an earthquake: the Modified Mecalli Intensity (MMI) http://en.wikipedia.org/wiki/Mercalli_intensity_scale . MMI is a subjective human description of the earthquake from different areas around the epicenter. This more "human scale" accounts for proximity to the epicenter, soil conditions, building construction, population density etc. using a very complex machine: You. You, or people in general, determine this number by feeling the earthquake and looking around and making the analysis of "OMG this place is destroyed who dropped the nuke (XII)" or "WTF my favorite tea cup fell down (IV)" .

It's the same as the USGS's "did you feel it" maps.

So this scale make it easier to compare earthquakes on a human scale. Awesome! ….but it's subjective. According to wikipedia, Santiago, Chile was a VII (very strong) but closer cities were an VIII (Destructive). Haiti's Port-au-Prince was a IX(Ruinous) and there were large population areas in areas of X (Disastrous). Whereas Moment magnitude is a scientific number which accepted by Americans, Chileans and Haitians seismologists alike, the Mercalli Intensity if different depending on who you ask (on an individual, neighborhood and country scale, Chile's scientists think the MMI is higher than the USGS.)

Also, the only way to get MMI maps is by having people fill out a survey and that not easy for big earthquakes in rural areas. So the MMI comparisons are useful but not quite scien-terrific.

Other things to consider?

Earthquakes in the US today will not cause the number of fatalities as an earthquake in Haiti. Why? Building requirements are more stringent and hospitals are readily available.

We can't make earthquakes stop so we need to fix governments and people. Promote preparedness, push awareness and provide support. Even small earthquakes can be devastating in societies not ready for it.

Action

Even though the magnitude scale is a silly bunch of maths, we should still act. Please be prepared, be aware and donate to red cross http://www.redcross.org/.

Filed under: Uncategorized

Big list of FAQ for programmers from Reddit Programming: I’m job interviewing and need a bunch of resources

jimmyhchan — Thu, 11 Feb 2010 19:24:18 +0000

The subreddit programming on reddit.com has always been an treasure trove of wonderful, sometimes quirky but always useful and interesting information. Unfortunately, the same stuff keeps popping up and it's hard to sift through the junk.

The FAQ for programming.reddit http://www.reddit.com/help/faqs/programming A large list of awesome

In currently looking at web development / programming jobs so here's a list of things I found interesting:

Blogs:

A List Apart - http://www.alistapart.com/ on the design, development, and meaning of web content
Joel's blog and Jeff's blog — but I think the stackoverflow podcast is more interesting

Interviews/Job:

List of programming interview questions – http://www.reddit.com/r/programming/comments/95nux/ask_proggit_what_are_your_favorite_programming/
Google interview questions – http://courses.csail.mit.edu/iap/interview/materials.php
Reverse phone interviews (interviewing your interviewer) http://www.martincmartin.com/blog/?p=144
The joel test –http://www.joelonsoftware.com/articles/fog0000000043.html
(not listed but) jeff's idea of a sniff test. That is when you hear about the company does it pass "the sniff test"

Web Hosting Alternatives

Nobody enjoys being with godaddy — for ethical reasons (they support republicans, they shutdown sites for random reasons, they take your domain and auction it) and for service reasons (limited software and limited support)

http://www.reddit.com/r/programming/comments/8bkm3/ask_proggit_recomendations_for_free_or_cheap/
http://www.reddit.com/r/programming/comments/747gk/alternative_to_godaddy_damn_bastards_donated_4/
Nearly free speech is always up there so do check them out https://www.nearlyfreespeech.net/ great pricing and great company philosophy

Filed under: Uncategorized

List of linux apps

jimmyhchan — Sat, 29 Aug 2009 20:42:05 +0000

http://www.dedoimedo.com/computers/new-cool-list-linux.html

Posted in Uncategorized

Escript / Finley releases version 3.0

jimmyhchan — Tue, 11 Aug 2009 13:12:56 +0000

Escript and Finely just released version 3.0.

According to the pypi site(http://pypi.python.org/pypi/escript/):

escript is a python based environment for implementing
mathematical models in particular those based on coupled, non-linear,
time-dependent partial differential equation. Parallelization with
both OpenMP and MPI is supported.
http://www.opensource.org/licenses/osl-3.0.php OSL license.

The 3.0 user’s guide describes it as:

It consists of four major components

* esys.escript core library
* finite element solver esys.finley (which uses fast
vendor-supplied solvers or our paso linear solver library)
* the meshing interface esys.pycad
* a model library.

There is a Finite Element code implemented with escript called Finley.
There looks to be support for gmsh (mesher) and output visualization
using matplotlib (matlab-like plotter) and VTK (visualization
toolkit).

Code site: https://launchpad.net/escript-finley
Documentation and examples : http://esys.esscc.uq.edu.au/docs.html
Posted in Uncategorized

List or Misc Compsci papers via Reddit

jimmyhchan — Mon, 20 Jul 2009 00:44:22 +0000

http://www.reddit.com/r/programming/comments/9220o/ask_proggit_recommender_a_compsci_paper_for_me_to/

Someone posted a request to the programming subreddit for compsci
papers to read over a weekend. There is a big list of papers from all
sorts of places.

Some that I found interesting:

Steele and Sussman’s The Art of the Interpreter
http://repository.readscheme.org/ftp/papers/ai-lab-pubs/AIM-453.pdf
Cook, Carpenter and Catmull’s The Reyes Rendering
Architecture http://graphics.pixar.com/library/Reyes/paper.pdf
Courant’s Variational Methods for the Solution of Problems of
Equilibrium and Vibrations
http://home.agh.edu.pl/%7Emilenin/programs/doc/Courant_1943.pdf
Baraff’s Linear-Time Dynamics using Lagrange Multipliers
http://www.cs.cmu.edu/~baraff/papers/sig96.pdf
The implementation of Lua 5.0
http://www.tecgraf.puc-rio.br/~lhf/ftp/doc/jucs05.pdf
Efficient implementation of the smalltalk-80 system
http://citeseer.ist.psu.edu/old/540956.html

It is interesting that there are so many seminal papers in computer
science. One day I’ll read some old school Timoshenko
http://en.wikipedia.org/wiki/Stephen_Timoshenko.
Posted in Uncategorized