And here's the thing: to a sufficiently fuzzy approximation, all the R6RS Schemes out there are interchangeable. They all offer a foreign function interface with roughly the same capabilities, they all offer the ability to launch and control subprocesses, and just about anything else can be built up from there. All of the cool things that can be built from these primitives, there is no fundamental reason they can't be portable.

So in April I started working on a set of portability packages, followed by an install tool, followed by a web-based package upload and browsing interface. And now it's all sufficiently done that I feel like I should release it.

So here you go.

Caveats

It supports Ikarus, Larceny, Mosh, Racket, and Ypsilon, at least on my machine.

The install tool could certainly use more features, like a package upgrade ability, code precompilation for the various Schemes, and complex version constraints on packages. The web interface is horrible code, and I'm just happy it seems to work as well as it does.

But what I am proud of are the simple but extensible package format, with myriad available extension points for new features, the simple repository format which allows anyone to start serving up packages if they don't like my offering, and the libraries which allow FFI and process manipulation to operate identically across Schemes.

Similar Projects

This code doesn't exist in a vacuum, and there are other projects whose scope overlaps with mine.

The most noticeable, to me, is Marco Maggi's Nausicaa, a gargantuan library distribution for R6RS Schemes. I regret to say that I have no experience using it, but the sheer number of libraries is impressive. I would like nothing more than to see some of those libraries available through UCL Install at some point.

The only other R6RS package manager that I have heard of is Dorodango. It does not seem to have been officially released yet, and if I understand the documentation correctly, it only supports Ikarus and Ypsilon at this time. It does have an upgrade command, unlike UCL Install at this point.  I envy the awesome choice of name, and I look forward to seeing what the author does with this one in the future.

Finally we come to Snowfort, an older attempt at a Scheme packaging system. I do not believe it is R6RS compatible, as it defines its own module system, and it appears to have no FFI or other portability layers, so I believe packages are limited to pure Scheme code. I would like to look at either manually or automatically porting some of these packages at some point, if the authors are amenable.

                          

My Prototype Remote Controller

In my junior-senior year (it's a long story) I designed and built an embedded device to control NXT robots for the FTC robotics competition. I still have the original breadboarded prototype lying around.

But this post isn't about that, I was just providing some background information.

This year, IEEE@WashU has decided to build a light-up dancefloor, similar to the old one, but split into a set of modular 2-by-2-foot tiles for easier maintainability, upgradeability, and not-exploding-people-ability. Also it will have pressure sensing.

We were originally planning to run the floor with an ATMega at the center coordinating everything. Unfortunately, our first try at the inter-module communications was prone to noise issues. After much deliberation, we chose a scheme using ordinary UART signalling carried over LVDS between modules, with direct connections between neighboring modules. Unfortunately, that meant we needed a chip with 4 high-speed UARTs.

We ended up settling on the STM32 line of ARM chips, specifically the STM32F105. It has 5 UARTs, of which all are capable of 2.25 Mbaud. Also it has on-board ADC functionality, and runs nearly 4x as fast as an AVR at top speed. So it's a pretty nice chip overall.

The first big problem was packaging. Most AVR chips come in a DIP package, which is nice and friendly for hand-soldering. The STM32F105, on the other hand, comes in a QFP-64 package, which is quite a bit less friendly. For prototyping, I soldered one to a breakout board by hand. It worked, sporadically, for about half a week, but I'm pretty sure my poor soldering is to blame for the failure, either due to bad electrical connections or thermal damage.

But from what I was able to infer about the operation of a non-damaged chip, the STM32 line should be pretty fun to work with.

• Like the AVR, it has extremely minimal requirements for external components. I ran it with power, JTAG, two blinky LEDs, and nothing else connected.
• The STM32F10x Standard Peripheral Library contains a lot of useful functionality already implemented, although the function naming conventions are pretty ugly.
• On-chip debugging is always worth it. AVR chips can do this too, but I never really bothered using it. Now that I've tried it on the STM32, I don't think I could bear blinky-LED debugging any more.
• Lots of fun features:
  • 4x higher clock speed than AVR.
  • Two 16-channel ADCs
  • Two 12-bit DACs
  • Unique 96-bit device signature

Aside from the soldering issue, I believe that the STM32 line of microcontrollers is exceptionally hobbyist friendly, and should be pretty fun to work with once I get some non-damaged chips hooked up.

                          

So I wrote shsel, a handy little tool which allows me to specify multiple shells a user can choose from at login time. It's simple, weighing in at 100 lines exactly, and allows only the execution of very specific commands read from a configuration file. It's nothing fancy, but I was unable to find any combination of existing tools that would achieve the desired result, so it's nice to have a new program to add to my bag of tricks.

                          

We will take our cue from this solution written in SWI Prolog, and encode our puzzle as a set of constraints that our solving code must satisfy.

An 'entry' is a single entity in the solution. In the Zebra puzzle, these are going to be the individual houses. For maximum generality, we're just going to represent all the properties of the entries with strings.

An answer to the puzzle is a collection of several entities which together satisfy all of the rules set forth in the puzzle description.

I was going to explain the rule types at this point, but the explanation ended up being a lot harder to understand than just looking at the rules for this puzzle, so we'll do that instead.

The first four rules simply teach our solver about numbers. After these rules are satisfied, the entries in the solution will be numbered one through five.

Then we specify all of the clues that make up the puzzle itself. These are given in order, so it should be easy to see the correspondence with the clues listed in the wikipedia article

Unfortunately, one drink and one animal are missing from the rules as stated, so here we just inform the solver "someone drinks water" and "someone owns a zebra"

So we have three kinds of rules, for which we'll need a data definition. By now it should be self-evident how each of these work

At this point, we have a nice, declarative specification of what a solution will look like, and we need to write the code to solve for it. The key to solving the puzzle efficiently is to realize that each rule is effectively describing some small portion of a possible answer, with empty strings representing unknown values. What we need next is a way to expand a rule into a list of all those answers that it represents

As we go about solving the puzzle, we will at all times have a collection of answers the solver currently knows to be possible, and a set of answer fragments resulting from expanding the rule we're currently trying to satisfy. What we need is a way to test every possible combination of the old answers with the new answer fragments.

Our solution will make use of the nondeterminism monad, called the "list" monad by the unenlightened. As we iterate over the rules, we will expand each one in turn, test every possible combination with the old answers, and then filter out the impossible ones.

At first, this will cause a combinatorial explosion of possible answers, but as new rules are added, we will eventually reach a point where each additional rule manages only to decrease the number of possible solutions, until there is only one remaining.

From the definition of applyRules, it is clear that our overlay operation needs to have type Answer -> Answer -> Maybe Answer. If any two answers are both defined and different from each other, we return Nothing, and otherwise we return the most defined of the two fields.

Before any constraints are applied, the answer is entirely undefined, so the process of solving consists simply of applying all the rules to an initial empty answer, and seeing what results. In this case, there is only one answer, but removing one or more rules from the list can make other solutions equally valid.

To see how each additional rule changes the set of possible answers, you can try something like "mapM_ print . applyRules emptyAnswer . take ## $ rules" in GHCi, for all the numbers between 1 and 20.

(This post is Literate Haskell, meaning that it can be copied and pasted in its entirety into a *.lhs file, and then run with runhaskell)

                          

Then I was kept horribly busy for a week by schoolwork, so a couple of days ago I started playing around with the problem. The results? Not too shabby!

So first of all, we'll be needing some imports. Since (for some odd reason) Haskell requires all imports up front, and since this blog post is supposed to be Literate Haskell, you'll have to just trust me that we'll need these for now:

Cells can be either On, Dying, or Off:

For convenience, let's define some constants:

Cells progress from On to Dying to Off, and they turn on only when they have exactly two live neighbors.

Since we know from the rules that we'll need the ability to count a cell's live neighbors, let's get that out of the way next.

So now we have all the pieces to progress from one world state to the next. For each position in the array, we need to look up all its neighbors, count the live ones, and then pass that data to the stepCell function. The helper function indexArray creates an array of cell indices. We map over this array to generate new values for each cell.

The `using` parArr rwhnf  is some Haskell magic which causes the array to be evaluated in parallel:

Now we have all we need to run a simulation, but it's not quite enough if you insist on getting some pretty pictures. For my fancy display purposes, I happen to like SDL.

The main function initializes SDL, generates a random initial state, produces an infinite list of future world states from that, and then draws each of the states in turn:

And our world drawing function is positively boring. We map over each combination of X and Y values, draw each one, and then flip the resulting image on-screen:

To take full advantage of the parallelism in this program, you'll need to compile with the threaded runtime and run it on multiple OS threads.

And then just sit back and watch the mesmerising patterns.

This code is available on Hackage and GitHub.

Pretty Colors

                          

Hoogle shows no functions with that type signature, and I couldn't find anything in the documentation that seemed to fit the bill. This seems like a rather fundamental function for getting stuff done, is there something I'm missing?

EDIT: The most code-golfed version I have yet come up with is while p a = p >>= flip when (a >> while p a) can anyone improve on that?

                          

is equivalent to the composition operation. I think

is possibly the coolest function in all of Haskell, so it annoys me when it requires six characters just to use it as an infix operator.

So I present what is, character-for-character, the coolest Haskell trick I have seen so far:

import Prelude hiding ((.))
import Control.Monad.Instances
a . b = a `fmap` b
infixr 9 .

And now, all sorts of fun little tricks are possible, like replacing

with

, and replacing

with

And of course, the same function composition magic still works like always.  Maybe it's just me, but something this simple, elegant, and fun to use just makes me happy inside.

But finally, I have to ask: is there any way of specifically declaring that a module *replaces* the standard Prelude (or even better, individual elements of it)? It would be really nice to just type 

and not have to also bother with adding 

                          

So a while back, I saw a neat little utility called "sponge" which simply buffered data until it ended, and then output it. While useful, this wasn't a standard Unix tool, so I couldn't very well rely on having it available.

Well today, after literally minutes of searching, I have found a command that seems to work just as well:

It outputs all lines in the input, beginning with line zero. Due, I suppose, to the way tail is implemented, it buffers all input before outputting anything.

Not exactly groundbreaking or anything, but a useful little trick to know when playing around in the shell.

                          

To create a nice, short URL abbreviation for git, such as myserver:<project>.git, put the following in ~/.gitconfig

That's all there is to it.

                          

Once you have that, it becomes easy to implement some sane functionality for exception handling, such as my current favorite functions, defaultOnError and errorToMaybe: