Most geeky thing you've done recently.

[Sagekilla]

I found the most anagramable word. Turns out it's "spear" and its 11 other anagrams.

Uh?

Code: Select all

In[2] := DictionaryLookup /@
    StringJoin /@ Permutations[{"s", "p", "e", "a", "r"}] //
   Flatten // {Identity, Length} // Through

Out[2]= {{"spear", "spare", "pears", "parse", "pares", "reaps",
  "rapes"}, 7}

I only count 7.

[The EGE]

Internet Anagram Server adds 8 more, with spaces:

A Reps
Pa Res
Pas Re
Spa Re
Sap Re
Asp Re
As Rep
As Per

[Sagekilla]

Okay I was about to say the geekiest thing I did today was that Mathematica one-liner I wrote two posts ago.

But that's changed. Most geeky thing I did was read an ~86 page paper on Numerical Relativity. I still haven't
grokked it fully, but now I know why it's so damn difficult. You gotta create surfaces of constant time, and
there are infinitely many ways to choose this. Plus, the initial value problem + boundary conditions that
Physicists are used to becomes a hell of a lot more complicated.

You need to satisfy a set of (IIRC) non-linear partial differential equations to generate the boundary
conditions and initial conditions. Once you have this, then you can start solving for a solution at some time.
But that itself is complicated as all hell.

[PhDFluff]

Today I stole liberated two flat screen monitors from ex-grad students so that I could finally throw out the enormous CRT-screens for our megaexpensive fluorescence meter and microliter-spectrometer. I felt extra nerdy when no one seemed to be joyous about this except me.

[The Geoff]

A couple of years back I searched Pi for valid ISBN numbers. There's three (that have been assigned to books) in the first 5 million digits. Since you ask:

At the four hundred and nine thousand, seven hundred and eighty third decimal place, we have:

Licentiate seminar on environmental engineering and biotechnology
by the Tampereen teknillinen korkeakoulu. Bio- ja ympäristötekniikka.
(Tampere University of Technology. Department of Biological and Environmental Engineering.)

In second place, at the two million, one hundred and twenty thousand, two hundred and fourth place, is:

Sneeuwwitje en Rozerood
by Jacob Grimm

Or as it's known in English, Snow White and Rose Red. Not hugely geeky, but there is something wonderfully gothic and conspiratorial about one of Grimm's turning up.

And taking the bronze, at the three million, six hundred and thirty thousand and thirty third decimal place:

The healing knife
by George Sava

A curious sounding book written under a nom de plume by a "noted Harley Street Surgeon".



More recently I revisited the project to find out if the first place was likely to be usurped in the near future. Sadly, there are no valid ISBN numbers in the first four hundred and nine thousand, seven hundred and eighty digits which have been assigned to a publisher, so the record is safe for now.

[nehpest]

Geekiest thing I've done recently? Volunteer for a science project with people I've never met (and may never meet) doing image processing on pictures taken by a robot millions of miles away of geological formations no human has ever laid eyes on.

That, or sign up for a graduate class on digital image processing with my university's local guru, so I can actually be useful.

[lemmings]

Over the past few month's I've been constructing a bike. When I realized that the 6 mile ride to campus is a little too far for me to ride, I added a motor to it. I dislike using throttles so I decided to add a microprocessor and sensors to read the pedal speed... then I dislike buttons so I'm adding support for bluetooth and writing an android app to control the micro.

A pedal assist electric bike which uses smartphones as the key, yeah, that just about ranks as the geekiest thing I've done in my life.

[Jakell]

I just used the feather in my Fedora as a diffraction grating (the barbs in the feathers are small enough and regularly spaced enough to work very well for this) to show that the yellow lights at a show were mercury fluorescent, and not sodium lamps.

Spiffily enough, there is almost no yellow light from mercury fluorescents, it mostly comes from the green mercury line and the red of the phosphor coating on the lightbulb.

[eSOANEM]

That reminds me a little bit of when we'd been given some stupid practical to do in school like measure the diameter of a hair with a screw-gauge micrometer and I decided to use it as a diffraction slit for the fluorescent lights. You couldn't see anything good though, all of the lights were on in the room so there was a lot of ambient light so all you really got was a spectrum.

[Sockmonkey]

I noticed that the most efficient way to stir in my coffee creamer is to move the spoon in straight lines with the bowl perpendicular to the line of motion for maximum turbulence, as opposed to stirring in circles.

[Copper Bezel]

This.



The first trig I've used in quite some time. (To get the ratio of the height of the pentagon to the base length - I know I probably could have found a ratio already made up, but it was more fun to figure it out - and then, since the shape is made of two large pentagons cut up, to get the ratio of the big pentagon to the little ones.)

I tried one where the bars are closed shapes on the inside instead of flat surfaces, but my pitiable hold on geometry was still better than my papercrafting.

[PM 2Ring]

Nice. The mathematics of the dodecahedron is fun to play with. I assume you noticed the interesting value of the tangent of the angle between adjacent faces.

Now make a stellated one.

Or a rhombic one, like in my avatar.

[Copper Bezel]

Ooh, damn. Well, no, now that you mention it, I'm now actually trying to suss what the angle between two faces is. I can see the irregular hexagon loop in the cross section of the solid, and I know the ratio of the two lengths of sides, but I'm still missing something.

I thought it was crazy and convenient that half the cotangent of 18° is also (sin 36 + sin 72), the ratio of the height of the pentagon to an edge, which means that when extending two non-adjacent edges from a pentagon, they cross at the height in the opposite direction. Anyway, it made all the little pentagons easier to draw. = )

I think I'll take you up on the rhombic dodecahedron. That's a fun shape. = )

[ConMan]

I gave my girlfriend Portal earrings for an anniversary (they would have been for Valentine's, but I ordered them late and they got stuck in the postal system).

[PM 2Ring]

Ooh, damn. Well, no, now that you mention it, I'm now actually trying to suss what the angle between two faces is. I can see the irregular hexagon loop in the cross section of the solid, and I know the ratio of the two lengths of sides, but I'm still missing something.

Here's a hint. The angle between two faces is the angle between the face normals. The icosahedron is the dual of the dodecahedron, so the angle between two adjacent faces of the dodecahedron is equal to the angle between two adjacent vertices of the icosahedron.

I thought it was crazy and convenient that half the cotangent of 18° is also (sin 36 + sin 72), the ratio of the height of the pentagon to an edge, which means that when extending two non-adjacent edges from a pentagon, they cross at the height in the opposite direction. Anyway, it made all the little pentagons easier to draw. = )

Also, once you have analytic expressions for the trig ratios of 36°, you can use the trig difference formulas to get expressions for the trig ratios of 6°, and then you can use the half-angle formulas to find the trig ratios of 3°, and even 1.5°, although they do get rather messy.

I think I'll take you up on the rhombic dodecahedron. That's a fun shape. = )

It is. It makes good 12-sided dice, since it doesn't roll as much as a regular dodecahedron. But it's mostly important because of its connection with sphere packing, as illustrated here:

Also, http://i2.photobucket.com/albums/y43/PM2Ring/RhombicKB1S_85.jpg

This image shows the relationship between the cube, octahedron and rhombic dodecahedron.
http://i2.photobucket.com/albums/y43/PM2Ring/RhombicH3S.jpg

A large version of my avatar image.
http://i2.photobucket.com/albums/y43/PM2Ring/RhombicKC4S.jpg

[Copper Bezel]

Wow, cool. Thanks for the visual aids. The image of the relationships is especially cool. (And now I see why the ratio between the lengths of the diagonals on the rhombus face is what it is.)

Here's a hint. The angle between two faces is the angle between the face normals. The icosahedron is the dual of the dodecahedron, so the angle between two adjacent faces of the dodecahedron is equal to the angle between two adjacent vertices of the icosahedron.

Okay, I think I see how that's going to work, then. Thanks - I'll have to post back when I get a chance to work it out. = )

Also, once you have analytic expressions for the trig ratios of 36°, you can use the trig difference formulas to get expressions for the trig ratios of 6°, and then you can use the half-angle formulas to find the trig ratios of 3°, and even 1.5°, although they do get rather messy.

Yeah, that sounds cool, but a bit beyond my reading level thus far. That's leading me to read up on some of the bits I've forgotten and some of the bits I never got to, as it goes, which gives me more fun things to play with. = )

It is. It makes good 12-sided dice, since it doesn't roll as much as a regular dodecahedron. But it's mostly important because of its connection with sphere packing, as illustrated here:

Oh, and that makes sense, since it tessellates in 3D. Cool.

Edit: It was eating my brain. Like, I figured it out while the nice man was replacing the battery in my car that had stopped working this afternoon. The tangent of the angle between two faces is 2. It is interesting - surprisingly integral. = )

[PM 2Ring]

Wow, cool. Thanks for the visual aids. The image of the relationships is especially cool. (And now I see why the ratio between the lengths of the diagonals on the rhombus face is what it is.)

It was a good excuse to show off some of my ray tracings.

Also, once you have analytic expressions for the trig ratios of 36°, you can use the trig difference formulas to get expressions for the trig ratios of 6°, and then you can use the half-angle formulas to find the trig ratios of 3°, and even 1.5°, although they do get rather messy.

Yeah, that sounds cool, but a bit beyond my reading level thus far. That's leading me to read up on some of the bits I've forgotten and some of the bits I never got to, as it goes, which gives me more fun things to play with. = )

Ah, rightio. The trig sum and difference formulas might look a bit cumbersome, but they aren't too hard to derive geometrically. However, there's a better way if you know a bit about complex numbers, since there's a deep connection between rotation in the plane and complex multiplication.

It is. It makes good 12-sided dice, since it doesn't roll as much as a regular dodecahedron. But it's mostly important because of its connection with sphere packing, as illustrated here:

Oh, and that makes sense, since it tessellates in 3D. Cool.

Exactly.

Edit: It was eating my brain. Like, I figured it out while the nice man was replacing the battery in my car that had stopped working this afternoon. The tangent of the angle between two faces is 2. It is interesting - surprisingly integral. = )

Yay!
When I first stumbled on that I was so surprised I had to double-check my geometry and algebra. But in retrospect it shouldn't have been too surprising, because of the connection between sqrt(5) and pentagons.

[Frenetic Pony]

I get excited every year for the Game Developers Conference. A flood of new white papers and etc., yes!

I don't even work or hobby in graphics development, I just like reading white papers about the subject.

[tomandlu]

It was a good excuse to show off some of my ray tracings.

What program is that? I'm guessing povray but purely because it makes doing mathy things so easy...

[PM 2Ring]

It was a good excuse to show off some of my ray tracings.

What program is that? I'm guessing povray but purely because it makes doing mathy things so easy...

You guess correctly. I love povray, and I was a big fan of DKBTrace, its predecessor.

[tomandlu]

It was a good excuse to show off some of my ray tracings.

What program is that? I'm guessing povray but purely because it makes doing mathy things so easy...

You guess correctly. I love povray, and I was a big fan of DKBTrace, its predecessor.

Ah... I used to enter povray images frequently for the IRTC back in the day (well, 1999 - 2002). Good times... tried to pick it up again recently, but embarrassed to find out how much I'd forgotten...

Even picked up an artistic merit award for the January 2002 round...

[Copper Bezel]

Yeah, I'm not cool. The mathiness of povray always scared me off. (The renders are extremely cool, of course.) I couldn't help noticing that all the example nhedra on Wikipedia were rendered with povray, too.

I did play a bit more, with the rhombic dodecahedron and a truncated icosahedron:



I really need to replace this camera. = P

[PM 2Ring]

Ah... I used to enter povray images frequently for the IRTC back in the day (well, 1999 - 2002). Good times... tried to pick it up again recently, but embarrassed to find out how much I'd forgotten...

Even picked up an artistic merit award for the January 2002 round...

Cool, tomandlu!

Povray is a big program with lots of features. It's easy to forget stuff when you're not using it frequently. OTOH, the docs and the povray forum are pretty good and it doesn't take too long to re-learn stuff, IME. Still, it can get scary when I look at the source code of some of my older povray creations. I sometimes find myself wondering what genius wrote this and what idiot didn't write adequate comments.

Yeah, I'm not cool. The mathiness of povray always scared me off. (The renders are extremely cool, of course.) I couldn't help noticing that all the example nhedra on Wikipedia were rendered with povray, too.

I did play a bit more, with the rhombic dodecahedron and a truncated icosahedron:

Spoiler:

I really need to replace this camera. = P

Nice work, Copper Bezel.

Povray can be mathy, but really, you don't need a lot of maths to use povray, apart from basic 3D coordinate geometry, although a little bit of trigonometry is helpful, too. But I guess it depends on what you're trying to do. Sometimes I use trig-based calculations to position & orient stuff, sometimes I use trial & error. To help build complex scenes there's a handy povray function called trace() that tells you the coordinates of the point where a vector intersects an object; trace() can also return the surface normal at the intersection point.

[tomandlu]

I sometimes find myself wondering what genius wrote this and what idiot didn't write adequate comments.

Just about sums it up for me....

trace() can also return the surface normal at the intersection point.

Ah, trace... very useful - especially when combined with a macro and the rand function... I used it to create the mayo in this first image, the snow and dry-stone wall rocks in the second and to randomly position the trees in the third...







A very powerful language - I seem to remember one user on the forums even wrote a basic ray-tracer using it, which would probably win this thread...

I wanted to model a spaceship the other day for an ebook cover, but I just found it impossible to pick up again fast enough, so I used Bryce. Easy to learn, but amazing how quickly you hit short-comings if you've come from povray. Layered textures? Err... no... Macros? Err... no... Hand-editing image source files? Err... no..

[PM 2Ring]

I used trace() to create this

from this.

And to place the trees in this stereo image.

[The Geoff]

Pff! What you want for 3D modelling, rendering and animation is a program called "Excel".

No, seriously....

http://www.gamasutra.com/view/feature/1 ... ry_3d_.php

I've not been brave enough to even contemplate trying it...

[Meem1029]

Hmm, does sitting at home on a Saturday night and reading the linux Filesystem Hierarchy Standards for fun count?

[KrO2]

How recently is "recently"?

This afternoon: Empirically estimate the volume of a Hershey's kiss by melting down 75 of them and dumping it into a measuring cup. The result of ~3ml per kiss is probably low, due to chocolate sticking to the sides of every utensil involved. (I then used the melted chocolate to coat a meatball sub sandwich, since that has probably never been done before, ever, and put the rest in one cell of an ice cube tray to mess with my roommates, but that's not really relevant to this topic.)

This week: Attempt to calculate the total mass of the air in a hurricane. By using the wind speed and the radius, I can get the acceleration based on change in direction (assuming approximately constant speed on average) by using the time it would take for a given bit of air to make a full circle (if it were actually doing that).By multiplying m*a, the result would give the value of the force of a great typhoon. I failed because I was unable to find data on the exact height/pressure/wind speed as a function of radius. But I'm listing this anyway because...because.

This month: Use Lorentz factors to find out how fast I would have to go before my mass increases and length decreases enough for me to become a black hole from the point of view of earth. (Answer: really freaking stupidly fast. I don't have it in front of me, but if I remember correctly, it was somewhere in the vicinity of c-10^-30. Rechecking may show this to be completely wrong, but I remember thinking that the difference between my answer and c was many orders of magnitude smaller than the speed of continental drift. So I'm not likely to become a black hole any time soon. Except from a certain point of view that is already going that freaking stupidly fast.)

This year: Possibly the fireballs with methane balloons. I'm really not sure how that ranks relative to the other stuff.

[algorerhythms]

I've been making various things in POV-Ray, including making a POV-Ray version of the lab I work in, as well as part of the university campus the lab is located in, and macros for optics lab equipment and road signs. Oh, and the picture I'm attaching to this message, which shows a view off of Dan's Rock in Allegany County, Maryland (there I imported the USGS terrain file for the region into POV-Ray as a height field, then used trace() to place trees, houses, and roads). I don't claim to be anywhere near as good at this as the people who make the images like the ones on the POV-Ray site, but still, it's pretty fun to do.

edit - I should note I didn't write the tree macro for that, I used maketree by Gilles Tran.

[HoodedPianist]

I got mad at a friend, and we were having an argument about division by 0 (he thought it was possible), and almost everyone in the neighborhood could hear us. He's not the brightest, as you can tell.

[PM 2Ring]

Empirically estimate the volume of a Hershey's kiss by melting down 75 of them and dumping it into a measuring cup. The result of ~3ml per kiss is probably low, due to chocolate sticking to the sides of every utensil involved.

Why not just use displacement?

I've been making various things in POV-Ray, including making a POV-Ray version of the lab I work in, as well as part of the university campus the lab is located in, and macros for optics lab equipment and road signs. Oh, and the picture I'm attaching to this message, which shows a view off of Dan's Rock in Allegany County, Maryland (there I imported the USGS terrain file for the region into POV-Ray as a height field, then used trace() to place trees, houses, and roads). I don't claim to be anywhere near as good at this as the people who make the images like the ones on the POV-Ray site, but still, it's pretty fun to do.

edit - I should note I didn't write the tree macro for that, I used maketree by Gilles Tran.

Cool!
Here are a couple of my optical POV-Ray images featuring a fresnel lens I designed.
Looking down on the lens, with some fog to show the convergence.

A side-on view to show that the lens's focal length is 3 units.

I generated those images ages ago, on a slow machine, so the quality of the fog media is fairly low.

Your landscape image is fine. Sure, it's not exactly up to the standard of Christoph Hormann, but that guy is a genius. And so is Gilles Tran. I haven't used his tree generation stuff, I used tomtree by Tom Aust to make the trees in the stereo image I posted earlier.

Are you using a plane for the sky? A big sphere works better.

The lighting in your scene is good, although the detail in the trees is not easy to see. Are you using Lightsys? IMHO, it's ok for outdoor scenes, but for indoor scenes it can be spectacular.

[algorerhythms]

Thanks for the comments. For the sky in the landscape image, I'm using a sky_sphere with a bozo color_map. I didn't know about Lightsys. I'll have to take a look at it. I have only one light_source declared in the source file, 6 miles away in the direction of the sun, but I'm pretty sure I had radiosity turned on for that rendering. I've since changed the main .pov file, so I can't be certain, though.

I'm attaching one of the images I generated from the optics lab equipment macros. The only scattering medium in the image is in the vapor cell. I tried to put a scattering medium in the entire room, but POV-Ray had problems rendering it. I let it run for a week, then calculated based on its progress that it would take ten years to complete the image...

[KrO2]

Empirically estimate the volume of a Hershey's kiss by melting down 75 of them and dumping it into a measuring cup. The result of ~3ml per kiss is probably low, due to chocolate sticking to the sides of every utensil involved.

Why not just use displacement?

Because that wouldn't have been as funny.

[flickering_candle]

On moving to a new apartment, and knowing the bedroom was noticably smaller than where I had been I made a scale floorplan on my computer with (scaled) images for the furniture to a)make sure everything I wanted in the bedroom could fit and b) find an aesthetically pleasing arrangement with a minimum of moving stuff around.

[Xanthir]

I did the same thing with Sims 3. It worked out remarkably well, since most of my rooms were very close to having all their dimensions in multiples of 3ft (which is approximately the size of a "square" in the Sims).

[PM 2Ring]

Thanks for the comments. For the sky in the landscape image, I'm using a sky_sphere with a bozo color_map. I didn't know about Lightsys. I'll have to take a look at it. I have only one light_source declared in the source file, 6 miles away in the direction of the sun, but I'm pretty sure I had radiosity turned on for that rendering. I've since changed the main .pov file, so I can't be certain, though.

I'm attaching one of the images I generated from the optics lab equipment macros. The only scattering medium in the image is in the vapor cell. I tried to put a scattering medium in the entire room, but POV-Ray had problems rendering it. I let it run for a week, then calculated based on its progress that it would take ten years to complete the image...

The sky_sphere is ok, but you may get better results with a real sphere, although that does have the disadvantage that it can get shadows cast on it, unless its finish is purely ambient with 0 diffuse. However, that doesn't work very well with radiosity, since it tends to be too bright, but it's worth experimenting with.

Using media effectively isn't easy, especially with multiple overlapping media of different types. However, there are some media experts on the POV-Ray forum that can offer guidance in setting media parameters.

A common technique when using radiosity is to do it in two passes. On the first pass you do the render at half the final scale, with all the slow stuff (like media & reflections) turned off and save the radiosity data to a file. On the second pass you do it at full scale, with all the slow stuff turned on, and with these radiosity settings
pretrace_start 1
pretrace_end 1
and for extra speed
always_sample off

There are some examples at Jaime Vives Piqueres' site, eg here, and I'm sure you'll find more online with a quick search.

Your optical equipment looks very good, but I noticed that you didn't use antialiasing in that render, or if you did the settings are very mild. I can guess why - antialiasing can really slow things down, especially when you're rendering transparent and translucent things like glass or media. One trick I often use is to substitute a little bit of post-process blurring for high resolution POV-Ray antialiasing.
First render your scene at around 20% larger than the size wanted for the final image, with antialiasing settings ranging from
+A0.4 +AM1 +R2
to
+A0.5 +AM1 +R1
When rendering glass, etc, you may need to drop that to +A0.3 or +A0.2, and you might even like to go up to +R3, although the increased quality may not be worth the increased rendering time.
Once the image is rendered, scale it up by 1.25 to 2.75. Duplicate the enlarged image, blur one copy using a selective Gaussian blur (or similar), mix the blurred and unblurred versions together, and then scale down to the desired size. I usually do the scaling, blurring & mixing using a script of NetPBM commands, or sometimes with ImageMagick, but for more control I do it in The GIMP.

[algorerhythms]

The optical equipment scene was one of the first I made while trying to figure out POV-Ray, and I hadn't figured out the details of how to do antialiasing yet. I just looked at the .ini file I had used for that, and realized why it took so long to do with antialiasing turned on: Antialias_Depth was set to 9. In my more recent scenes, I've been using Antialias_Depth=3, which renders a lot faster.

[eSOANEM]

My mum decided that we all needed to get rid of the random crap we've been hoarding in our house. Going through my things, I found a pair of glasses I had as a kid with diffraction gratings for lenses. Now that I know how they work and what they actually do, I decided to look at the spectra of the various lights in and around the house.

The sodium street lamps behaved as I expected, giving an almost pure yellow colour (although one down the road also had some mercury lines). The fluorescent tube lights gave a continuous spectrum which surprised me, I was under the impression that they actually give off a handful of different visible wavelengths which sum to approximately white. I then looked at the incandescent light bulbs (which of course gave a continuous spectrum) and some of the energy efficient ones (which gave a discrete spectrum).

What interests me is why the fluorescent tubes gave a continuous one but the energy-efficient bulbs didn't seeing as I was under the impression that the energy efficient light bulbs were coiled up fluorescent tubes.

[neato]

I just got off my spring break, which I entirely spent solving problems from Jackson's Classical Electrodynamics instead of hanging out with the other stupid 17 year olds and getting drunk.

Well, I did get drunk once, but we played drunken Mario Kart, so I think they cancel each other out.

[Meem1029]

At dinner my friend and I somehow got on the topic of Markov chains and using them to generate sentences. so when I got back to my room I wrote one just because I had nothing better to do...