trystan.org

“A language that doesn’t affect the way you think about programming, is not worth knowing” – Alan Perlis

Washington State loves booze and gambling and the taxable revenue that follows. In the latest effort to encourage the former business, the state government has recently legalized the legendary beverage absinthe and reduced a distillery license from $2000 per year to $100 annually. A similar move in policy many years ago launched the state’s successful micro-brewery and boutique winery businesses.

Now, I have little against such businesses. In fact, I’m a connoisseur of artisan wines and beer. What gets my goat, whatever that means, is policy makers seem content with further relaxation of regulations around a known dangerous drug while continuing a nonsensical “zero tolerance” policy regarding less harmful substances. In the case of absinthe, there are two dangerous substances, alcohol and thujone. The harmful effects of alcohol are well documented and include interesting side effects such as brain atrophy if abused heavily. It is one of the few drugs of abuse known to cause brain damage in humans. Thujone is a psychoactive chemical found in wormwood which is used as an additive in absinthe. Seattle based KOMO 5 evening news misreported the effects of this ingredient as being “largely mythical and non-hallucinogenic.” Not surprising given the utter lack of technical knowledge and research ineptitude of local journalists. However, research published in Neuropharmachology reports inhibitory action on GABA receptors and the serotonin system [1]. This may explain the subjective reports of stimulant like effects at low doses and hallucinogenic effects at high doses. Historically, absinthe is thought to cause a form of drug induced delirium after prolonged and habitual use. However, it is difficult to separate the effects of thujone and alcohol, which is known to cause persistent delirium after years of abuse.

Meanwhile, lawmakers and journalists continue to demonize other substances such as LSD and cannabis which have not been shown to be neurotoxic, or physically toxic for that matter even at quite high doses (note, as with everything, a lethal dose does exist). In fact, THC and other oils found in street marijuana, have been found to be neuroprotective [2]. Why do we continue to reduce restrictions on dangerous drugs and put people in jail for years for possession of less harmful drugs? Certainly, the effects pf a drug play a part. Alcohol makes one gregarious and potentially agitated or violent, qualities that ‘gel’ with the American culture. On the other hand, LSD is subjectively spiritual and therefore taboo. Only trained authorities, the priesthood, are to commune with extra-physical forces. Our fear of such substances hearkens from our puritan heritage and the control over society and individuals it desires. This will take some effort to change.

For now, I’d be content for reporters to just begin getting the facts strait on this topic.

[1]: Deiml T, Haseneder R, Zieglgaensberger W, Rammes G, Eisensamer B, Rupprecht R, Hapfelmeier G. “Alpha-thujone reduces 5-HT3 receptor activity by an effect on the agonist-reduced desensitization”. Neuropharmacology. 2004;46(2):192-201

[2]: Cannabidiol and (?)?9-tetrahydrocannabinol are neuroprotective antioxidants, A. J. Hampson, M. Grimaldi, J. Axelrod, and D. Wink, Proc Natl Acad Sci U S A. 1998 July 7; 95(14): 8268–8273.

Interestingly, both kinds can make you loopy. The wine we drink and the wine we use to play Windows games on Linux. I recently got a couple older games to work that required a little tweaking. First, Rise of Nations was failing to install upon entry of the CD key. This was remedied by copying mfc42.dll to wine’s windows\system32 directory. I’m guessing there’s probably some legal issue regarding redistributing this dll with wine. Secondly, Heroes 3 was failing to play under my default wine settings. Under the Graphics tab in wine config, I enabled the virtual desktop and then it worked fine. Appearently v1 of Heroes 3 requires a 800×600 video mode which I do not have configured under X.

Random maze generation algorithms are something I dove into in the course of developing an environment for AI agents such as neural networks. I settled on a recursive space subdivision algorithm due the conceptual simplicity. In terms of implementation, I made it a it more exciting by writing such an algorithm capable of maze generation in an arbitrary number of dimensions. A particularly interesting, and very general, sub-algorithm that fell out of the implementation is an algorithm deriving all possible in-place combinations of a set of lists. Please forgive me if my mathematical notation is a little wonky in the explanation below; discrete math was a long time ago.

In the scope of spatial subdivision, consider the case where you have a 2D integer place with boundaries defined by the points (0,0) and (9,9). You pick a random point to subdivide the space, say (2,5). Now the space is split into four sub-spaces defined by s = {[(0,0), (2,5)], [(2,5), (9,9)], [(0,5), (2,9)] and [(2,0), (9,5)]}. Lets define a function ‘c’ that returns every in-place combination of a set of points. Then, s = c( {(0,0), (2,5)} ) U c( {(2,5), (9,9)} ). In general, s = c( S ) U c( T ) where S is the set containing the lower point and the subdivision point and T is the set containing the subdivision point and the upper point. I have verified the correctness of this by exhaustion for examples of up to 4D space. That is, where the points passed to c contain 4 integers (a, b, c, d). I believe it is valid for arbitrarily large positive integer dimensions.

Since the algorithm was implemented in Lisp, the sets passed to c can contain anything. There is no reason the elements need to be integers. Therefore, the algorithm is likeky usable with little modification for any problem requiring the analysis of in-place list combinations. In addition, the algorithm is not limited to a set of two lists. There can be any number of lists with any number of elements. The algorithm is about 20 lines of Lisp code published at http://code.google.com/p/compu-psy-research/, see the function ‘combine-interchanges’.