My new server is up and (mostly) running.  After the old Sony laptop became unstable, the wife got me a Christmas present: a new low-power system (still much more computing-powerful than the Sony.)  This one runs on AMD’s dual-core E-350 APU, which supports hardware virtualization, so I installed a base server and added the web server on top of it as a virtual machine.  Seems to run pretty well, and now I can back up the entire web server and do a complete restore on different hardware without any re-configuration.

Back in the 70′s it was popular to predict that the global trend of air pollution was getting so bad that it was about to block out the sun, leading the earth into a civilization-crippling ice age.  The disaster did not arrive on time.  (Paul Ehrlich claimed it would be a done deal by 1985.)  In the 90′s it became popular to predict that the the world was about to overheat, again, quite disastrously, and some people rode this this one well into the 2000′s.  Well, you all know Al Gore.

Today we learn from NASA that, quite unlike a leopard, the sun is changing its spots.  According to their predictions, sometime after 2013, (which, according to the Mayans, is already post-apocalypse) the sun will enter period of coolness, at least with respect to what gets directed towards the Earth.  Interesting stuff, to be sure, and they’re comparing it to another little Ice Age.

It seems to me that two things haven’t changed: First, the apocalypse is still right around the corner.  This has been the case since long before the 1970′s.  In the US, Seventh-day Adventists have William Miller’s failed doomsday in 1844 to thank for their religion, but even that event came relatively late.  Check this out for a brief and entertaining history of the end.

Second, we still seem to be using the motion of a pendulum to determine the state of the Earth.  Foucault would be proud.

I was working on a discussion post for my Operating Systems class on the topic of Deadlock detection.  The discussion turned to the overhead inefficiency of running the deadlock detection algorithm at regular intervals, because, well, running the deadlock detection algorithm too often creates a lot of overhead in a system and can be inefficient.  This question occurred to me: why should deadlock detection have to run at regular intervals when deadlock conditions are not always possible?  A related question occurred, to wit: even when deadlocks are possible, why should detection run at constant intervals when they are not always equally likely?  This led to the idea that it might be possible to combine two different deadlock handling approaches to gain some of the benefits of both.

Still, knowing how many people are sitting down at any given time (and the general appetite of each one) can be a very useful thing, even if you don’t know with 100% certainty what each one will order.  (Ask any restaurant manager.)  Also, the resource-claim matrix algorithm is relatively lightweight, more so from an overhead processing perspective because it only runs in the rare (in machine terms) event that a new process is initiated.  (roughly, whenever the user launches a new program, though there are many exceptions to this.)  Also, while you need precise numbers when you are relying on one method being able to predict worst-case scenarios, estimates or averages are acceptable in cases where you have a backup plan to catch relatively rare occurrences.

My idea is that an operating system could be designed to take advantage of a three-tiered “deadlock probability” state.  In this system, we actually maintain two resource-claim matrices, to be updated at each process initiation or termination, just as in deadlock avoidance.  The first claim, or “safe” matrix would be the classic one, containing the worst-case scenario resource claims of every running process.  The second “probable” matrix contains a best-guess estimate of what resources the process is likely to be using at any given time.  This may be a tweakable or variable value, as a straight average of resource use might not adequately represent the bursty nature of resource usage.  The final matrix would be the total system resource matrix, which would represent the total number of resources of each type available system-wide.  The deadlock probability state would be determined as a comparison between the three matrices.

If a comparison of the safe matrix with the system matrix shows that there is no possible conflict for any resources on the system, then the deadlock probability state can be set to “safe,” and the deadlock detection algorithm does not need to be run at all.

If the comparison of safe and system matrices shows that there is a possibility of deadlock, then the probable matrix is compared with the system matrix.  If this comparison shows that sufficient resources are probably free, then the deadlock algorithm has to run, but it can be run relatively infrequently, so as to keep things efficient.  If, on the other hand, this comparison shows that a deadlock is likely, (and the definition of likely here can be tweaked for performance, by tweaking the values that contribute to the probable matrix) then the deadlock detection algorithm can be run more frequently.

One additional advantage to this that is not immediately apparent is that the values that contribute to the probable matrix for a given program can either be specified by the developer or by the OS maintainer, and can be tweaked for efficiency using either trial and error or detailed advance knowledge of how the program usually runs.  One could even generate these values automatically, on the basis of past performance, though the algorithm to track and record this would be adding extra overhead to the system.  The fact that there is a backup (deadlock detection still runs whenever a deadlock is possible, just less frequently) means that there is a safety net for poorly chosen values and rare coincidences.

A quick search of the internets has not revealed to me any other mention of this structure, but I would not be terribly surprised if it has been thought of before.  The actual utility of this methodology also depends on how much practical overhead is caused by current deadlock detection schemes, as the maximum speedup we can gain here is the total elimination of that overhead.  Modern hardware being as fast as it is, this might not be a significant issue, though perhaps it would be more important on things like phones, smart phones, and other resource-restricted systems.

Even if it’s not practical, I might be able to make some sort paper out of this idea, which is why I’m posting this here.

I have finished my initial review of Jason Silva’s Brazen.  He sacrifices plot suspense in favor of character development in the first few chapters, so it may seem to start off a bit slow, but overall, I think it was a worthwhile trade.  The novel as a whole is excellent, as long as you go easy on him a little on minor things like typos and spellcheck-generated errors.  Just remember when you see those that he doesn’t yet have the benefit of a professional editor.  I think I may have more commentary on it later, as it really is an interesting novel, but for now, here is the review I posted on Smashwords and Amazon:

Just because you are paranoid and have an over-active imagination, doesn’t mean that you won’t be drugged and forced to take part in bizzare, ritualistic sex acts with rodents.  Or maybe it does.  This is the tale of Brennan Frick, an intelligent, high-strung, young man who decides to leave everything behind for a fresh start as a reporter in Biloxi, Mississippi.  Along the way he befriends Les, a man who is larger than life in both the physical and metaphorical sense.  (Though at the same time, he manages to be somehow familiar, like some bombastic favorite uncle.)  Together, they take on the evasive and somewhat mysterious Vietnamese immigrant community, some wicked and depraved young future republicans, and the society of the sickeningly weathly elites at the very top of the Pittsburgh social pyramid.

Silva clearly takes delight in interesting wordcraft, constructing his sentences much like his scenes and images with some of the potency of Lovecraft mixed in with some of the farce and absurdity of Kennedy Toole.  The story itself has many of these elements, but with more of a (for lack of a better term) “literary” slant, in that he actually takes the time at certain points to engage in the lost art of character development, which adds depth to the experience.  It is hard to make a direct comparison of this rather unique style, but you can definitely tell his influences if you’ve previously read “Confederacy of Dunces” or “Fear and Loathing in Las Vegas.”

This novel is silly, insane, exciting, and fun.  It is clearly the product of a diseased mind, and I badly want to read more.  Brazen may be a bit too literary to be subjected to a traditional sequel, (one never knows) but Jason Silva is definitely one to watch.

I awoke one morning from uneasy dreams to find myself transformed into a monstrous vermin.

Here’s the run-down: Franz Kafka writes some really crazy stuff.  If he didn’t actually invent surrealism, (and he may have) he certainly re-defined it, and his definition stands.  It’s been a while since I read it, but I remember the metamorphosis to be shockingly observant and true-to-life, in spite of its complete departure from reality.  It was almost an incredibly biting satire, and yet not really satire at all.  It was in some strange way totally believable in spite of subject matter that was unreservedly fantastic, but then, that is the standard that we know as surrealism.

So Kafka never shows this stuff to anybody, (we’ll say) and when he dies of tuberculosis, he bequeaths his extensive papers which contain this stuff to his friend, Max Brod, saying: “Dearest Max, my last request: Everything I leave behind me … in the way of diaries, manuscripts, letters (my own and others’), sketches, and so on, [is] to be burned unread.”

Brod, of course, ignores his friend’s dying wishes, reads every word, and decides to publish.  The works become famous, (deservedly so) and Brod presumably makes a fair sum.

Brod dies in 1968, leaving the remainder of the papers, some of which are still unpublished, to Esther Hoffe, with instructions  that they be given to an academic institution.  Naturally, Hoffe ignores this request, instead deciding to sell the manuscripts and other documents off piecemeal to the highest bidders.  (Some of which, by accident, happen to be academic institutions.)  She makes a fair sum.

So Hoffe kicks, and of course she leaves the remaining papers to her two daughters.  Presumably, the instructions this time around are to continue on course, hold out for the highest possible price, and make a fair sum.

It wouldn’t be right for a bequest of these documents, -any kind of bequest for these documents- to be honored, so of course the Israeli National Library steps in to put a legal stop to such a fate.  The Library seems to have a pretty good claim, since it would likely have been on the short list of places Brod would have put it himself.  It is also strangely appropriate there is a court case ongoing to decide whether to honor the wishes of a dead person who did not honor the wishes of a dead person, over documents he possessed only because he did not honor the wishes of a dead person.  No word on how much the library expects to make by posessing the material, but hopefully it’s a fair sum.

I personally am in favor of something akin to option two in the link below.

http://reason.com/blog/2010/07/27/department-of-literary-litigat

P.S. Brushing up on my literary skills in preparation to review a new novel.  Stay tuned!

Positioned directy above “the shark”

P.S. Maryland, I still love you, I just wish that you were a free state.

…and with strange aeons, even death may die.

I had to write a post to go with the two links I added today.  You can see them on the sidebar under “Fun Stuff:” the Cthulu Chick Tracts.  I first ran into these several years ago, and quickly spread the word to my wife and co-workers.  My wife even showed it to her mother when she came to visit, and she was really into it.  She even decided to quote that phrase, “That is not dead which can eternal lie, and with strange aeons, even death may die” randomly and without any explanation, during a staff meeting at work.  Last I heard, she had done so more than once.

Yes, my mother-in-law is cooler than most of the people you know.

While the Cthulu Chick Tract is pretty awesome by itself, let me not pass up this opportunity to plug H.P. Lovecraft.  I am most familiar with his work in short story form, and it is a form of which he is a master.  His stories are weird, surreal melodramatic, horror that borders on outright comic absurdity, but for some reason, manages not to feel like outright comic absurdity when you’re actually reading it.

I was first turned on to his work by an old friend of mine, whose blog, Fantasy Heartbreaker, you will find linked to the right.  I’ve kind of lost touch with this friend since college, but I recently dug him up upon joining Facebook, which led me to an odd, “small world” sort of discovery.  You see, my friend actually wrote an awesome “fan fic” style, well, fan fic of an actual Chick Tract.  I had seen a link to this several months ago via my favorite news aggregator/blog, Reason’s Hit and Run.  The link I followed didn’t mention the author’s name, and I didn’t read through the comments to see who it was.  All I knew was that it was awesome and that it had once again confirmed my conviction that the internet rocks.

My friend recently posted a link to his own blog on his facebook wall.  Actually, it was a link to that very same story, which I recognized immediately, and thus did I have my “small world” moment.

Here’s what a trillion dollars (you know, the round figure that’s lately been popular in government circles) looks like in cash.

http://www.pagetutor.com/trillion/index.html

I’m thinking of activating the Addtoany plugin on this site.  That’s the thing that gives you tools to quickly link posts on Facebook, twitter, digg, etc.

Pros and cons thus far:

Pros: I like giving people tools to do what they want, I like information of the kind that will be on this site to be free, and I’m generally in favor of unhindered interconnectedness on the internets.

Cons: I don’t really like the look of it, and it does not fit the color scheme of this site.  It adds an extra external script to load, which I, as a user of a script blocker, generally do not like.  I haven’t fully decided what I’m going to do with this site, but I doubt anything on here will be widely linked, anyway.

I haven’t even given out the link to this site to anyone but my my wife and brother, or really added any actual content to the site at the time of this writing.  I consider it to be in alpha mode.  I’ll hold off on it for now.  If, in the future, anyone wants to weigh in on this, this will be the place to comment.

This is my first attempt at a theme.  This one is modified from one called “Adventure.”  See credits in the footer.  The background photo is of the actual Hindenburg smoking lounge.  I found it at www.airships.net, which is an excellent site for info on the great airships.

Also, the little airship/cigar favicon animation is one I made myself from clip art found online.

Let me know what you think in the comments.  Hopefully, they work.

The Hindenburg Smoking Lounge

• About

A sanctuary for life's little pleasures in a world full of Hydrogen.