I’m a cloud computing skeptic. Let me explain what it is and why I have this skepticism.

“Cloud computing” has, over the past few years, become a buzzword for technology marketers pushing a certain type of client-server computing. Client-server computing describes arrangements like the internet, where you on your machine (the client) send and receive little packets from another machine somewhere in the world (the server) that is processing and serving up data. That’s how you’re viewing this website, and all other websites. Your client machine sends a request to some server, which in turn sends data to your client, and so the loop repeats over and over.

Cloud computing is a term signifying the use of this client-server model to run programs which are normally run on your local machine. Google Docs is one example of this. We’re all familiar with Microsoft Office, where you purchase software, install it on your computer, and then run the program on your computer. This is the traditional model. A cloud computing version is something like Google Docs – where rather than installing and running the program on your machine, the program runs somewhere on a Google server and it sends you the results (the formatted document, etc). There’s still some computing going on on your end to display the data but typically this is minimal. And although the term “cloud computing” refers specifically to the process of computing data, this model usually stores saved data “in the cloud” and not on your local machine.

This certainly represents a significant change in the computing paradigm and many new opportunities, but I see at least two significant problems with it: privacy concerns and technological-economical limitations. The privacy concerns are things that non-technical people like to talk about being a problem with cloud computing – and I agree with them on this. How can you guarantee that your data will be secure from peeping eyes when it is located in a server somewhere (you don’t know where), subject to the laws of the government of the country where the server is physically located and the good will of the company’s employees and their internet security policies? That in and of itself is a huge political and security problem for the concept of cloud computing; however, rightly or wrongly most of us already freely hand over Facebook a lot of our personal data without much hand-wringing.

But to my mind, the technological-economical issues are much more damning, and they are more easily thought of by modeling computing and internet connectivity as commodities rather than technological innovations.

The cloud computing model requires two commodities: the first, computing power, represented by some more or less desireable and more or less powerful configuration of hardware that crunches numbers and runs programs. The second commodity is connectivity between machines. The cost of this connectivity commodity – both in quality (usually measured by speed) and dollar price – varies depending on the machines involved, their location, and the type of the connection (cable, dial-up, DSL, T3, etc.) .

In the traditional computing model, the only commodity you need is the computing commodity – the amount of processing power available to you. As with most commodities, the more you pay for it, the better the thing you have is (in this case, more powerful processing). But as the internet has spread, the communication commodity has become more and more important. Because connectivity was initially very expensive (a very high dollar-to-bandwidth ratio), the internet served chiefly text. Then, as the dollar-to-bandwidth ratio decreased and bandwidth became more available, internet content expanded to include images, and then audio, video, and interactive content. What the cloud computing model is banking on is that bandwidth is now cheap enough and available enough that rather than running a process locally, you can utilize that bandwidth and talk to one of the already existing tens of thousands of collections of servers all around the world and have them more efficiently do most of your computational work for you and send the results back to you locally. It’s important that the “more efficiently” part, for reasons I won’t go into, isn’t much disputed. Everything is computed far away and you get to see and interact with it on your screen and keyboard. Plus, you can then access all your programs from any machine as long as it’s connected to the internet, because all your programs and data are living out there in the cloud somewhere. It’s a miracle!

The problem I see with this is that the local processing commodity is currently (and I suspect always will be) far cheaper than connectivity. Your local machine executes single commands at the order of one every few nanoseconds. The time it takes you to send and receive a packet of information from a server is, if you’re lucky, on the order of a couple milliseconds. Already, talking to a server is far more costly than running a local computation when it comes to speed. This makes intuitive sense, however: in order to compute something locally your processor must shuffle data among its local caches and memory, communicating at the speed of electrons-over-a-wire over a few inches of space (the speed of light minus some significant not-in-a-vacuum costs); whereas communicating with a server somewhere is the same cost over miles and miles of distance plus intermediate routers directing your data where to go. Doing things long-distance always takes more time.

But if we’re talking nano- versus milliseconds, do we really care? The human brain doesn’t notice time on that scale so perhaps it doesn’t matter. Aside from the obvious observation that long-distance communication is slow, long-distance communication is also costly. Monthly internet charges are often $40-$50for the lowest level of broadband connectivity (at least where I live). Computers, on the other hand, run from $200-$300 for the cheapest netbook to $1500-$2000 for your really snazzy MacBook Pro or other high-end device. The point is that the cost of the computer is equal to somewhere between 4 (on the low end) and 30 (on the high-end) months of low-end broadband connectivity.

But people already pay for broadband connectivity, so why not utilize that existing resource to further drive down the cost of the end-user computer, since we are already connected to the internet at these high speeds? I would suggest two reasons: the first is that that difference I illustrated in computational cost versus connectivity cost is going to be paid by someone – there is no free lunch. Either the company providing the cloud computing is going to pay for better connections to their servers for the masses (driving up the cost of their service), or you are going to pay for a better-than-low-end broadband connection to more effectively and quickly communicate with “the cloud” that is running your programs. Because connectivity is expensive, someone will pay for it, and either directly or indirectly costs will be passed on to the consumer. The second reason you cannot simply co-opt existing end-user internet connections for generic cloud computing is the way cable modems work. A cable modem shares its broadband among multiple end-users. This is a great system as long as not too many users are doing too many heavy communication tasks at once. Currently, this almost never happens, and broadband usage follows a continuum from the normal user whose heaviest tasks are watching Hulu and doing the occasional download to the much rarer user who is constantly downloading and uploading gigabytes of data (usually illegally) over peer-to-peer torrents. My point is that if we are all constantly communicating our dara on everyday computing tasks over the internet, we’re going to look less like the modern everyday user and more (although never completely) like the heavy-end torrenting client, taxing the cable broadband infrastructure we’re relying on for connectivity. And increasing that broadband is (again) going to be more expensive than offloading tasks to local machines.

This may seem a bit like an arm-waving exercise since I haven’t quoted direct data and hard numbers. Fair enough: this is partly because I’m lazy and partly because this isn’t meant to be in any way scholarly, just my own mad musings. But where does this leave me in my skepticism? I suspect that data-light computation can (and maybe will be) easily be offloaded onto the cloud: things like word processing, spreadsheets and the like, and viewing videos (the computational power required to store and serve up the data is extremely minimal compared to that required to display it to a screen – this is why Netflix streaming and Hulu are already easily functioning). Streaming data that is then rendered and output on your local machine is pretty easy; this is already happening – it is when you want to in any way modify this data that it becomes a problem. So I don’t think programs like Photoshop or any video editing software could be offloaded into the cloud. Same thing for operating systems: these will always be local, although there is occasionally the starry-eyed optimist thinking this can be done over the internet. Ditto for computer games (although we’ve become pretty committed to a gaming console model): to a limited extent data can be communicated easily over the wire, but you still need fairly beefy hardware to store most of the game’s data locally and compute all of it locally. I suspect the limit of cloud computing is this: because of the high cost of connectivity relative to computing power, simple and common end-user tasks, including office productivity and media consumption, can easily be moved to the cloud, but increasingly popular media-interactive tasks like photo and video editing, and games will always require computation to be performed mostly on local machines.

I may be totally wrong on this, and if you have a different opinion I’d love to hear it. The tasks I’ve listed that I think are amenable to a cloud computing model represent the majority of user tasks on a computer, and so I think there is a legitimate potential space for cloud computing (barring security/constant-connectivity concerns which could potentially be cloud-killers); it is just far from a panacea for all computation. Ultimately, I think the extent to which cloud computing catches on depends on the value people assign to its benefits (data and interactive programs available from any computer) versus the presumably-higher cost of cloud computing.

I’ve been reading an interesting series of posts from Rod Dreher and Ta-Nehisi Coates. Rod starts off with a post about a 602-pound woman who is intentionally trying to get to 1,000 pounds. He brings this up first before going on to rail against an article in the New York Times suggesting that stigmatization against fat people is wrong and unproductive.

TNC, responding to the same article, muses:

I’m not clear on precisely how much shame can actually help. It’s shame that’s created our absurd McWeightLoss culture where Octomom takes to the cover of celebrity magazines to show off her new bikini body, and retired athletes claim to have found the secret to losing five pounds a week. It’s symptomatic of who are, of our abiding belief in short-cuts, and our technological ability to elide truth.

Rod responds here defending shaming the obese, relating his own struggles with being overweight, and that he doesn’t want to slip into accepting or being okay with his vice (of, he says, being mediocre and gluttonous and lazy).

Where to start?

I suppose I can start where Rod started – the six-hundred-pound woman who is trying to get to half a ton, and her encouraging and enabling co-fetishists. Although coming up with a correct and consistent theory of morality is extremely difficult, I am fairly sure I can call this behavior ‘imprudent’ if not out-and-out ‘morally wrong.’* One would assume that this woman (and her boyfriend and other enablers) has some desire to see herself be enormously fat. Fulfilling desires is generally a good thing: however, we are complex creatures with many and often conflicting desires. I would imagine that this woman also has other desires: desires to live and not to die, desires to express and give love, desires to be part of a community of human beings, to be self-determinate in her life and act as an independent agent, and so on. I think it is fairly self-evident that these desires are all thwarted by her immense size. If this fetish is like other fetishes, then I suspect it is consuming enough that she either does not think of or does not realize how negatively the rest of her life is being impacted by being 600 pounds. And as Dan Savage might have advised her, she could almost certainly have satisfied and managed this desire by overeating once in a blue moon and having her boyfriend call her a pig to her heart’s content, and then to stop dwelling on it and go running in the morning and maintain a healthy weight – one-quarter to one-third what she is now. It would certainly be much healthier. Although it is difficult to comment on the quality of life of people you don’t know, I find it a virtual certainty that she is drastically diminishing her own quality of life and shrinking her horizons by expanding her size. ‘Imprudent’ is a rather modest way to describe this behavior.

What I find a bit more telling, however, is that Dreher uses this extreme example to segue into the health costs of obesity, and contrasting this with starving people in Haiti, to make his later point that obesity is a spiritual problem. That’s some fancy slight of hand but it’s not very convincing. It is difficult for me to see how people who are starving are either helped or hurt by someone far away being obese. It may look awfully pious to paint fat people as out-of-control moral devils and the starving as moral angels, but it’s totally unrelated to why either group is the way they are. It is not as though the world has a limited amount of food supply and those who eat too much are preventing others from eating enough. In fact, we overproduce food and feeding the planet is not a matter of insufficient food production – the reasons people go hungry and are starving have to do with economics, infrastructure, and politics. “But there are starving people” is not a legitimate charge against the morbidly obese; “but you are diminishing you capacity to enjoy and experience these valuable things in life” is.

Rod also talks about his own struggles with weight. He says he’s put on some weight and is about 20-25 pounds heavier than he should, and even that this is approaching obesity. I find it very difficult to take this complaint seriously. Here is a picture of Rod from last year:

It is hard to imagine how 20-25 pounds on that slight frame can be reasonably taken as a man who seriously struggles with his weight. Perhaps he does, I don’t know. But knowing people who legitimately are larger and trying to lose weight, I have to ask myself if this is a bit of self-posturing and self-flagellating. If Rod is 20-25 pounds overweight due to bad habits, he certainly could stand to lose some weight. However he is hardly a member of the class of people who are obese, have food-addictions, or have weight-related health problems and are trying to get down to a manageable size. These 20-25 pounds are almost (but not quite) vanity weight. Bolstering my interpretation, Rod later goes on to talk about how lazy he is and how hard he is on himself about his weight in his follow-up post. Although Dreher himself might not be, it is entirely possible to have an extra 20-25 pounds without being either a glutton or a sloth. Perhaps he is exaggerating for the sake of the blogosphere, I don’t know; I just find his caricaturization of a handful of pounds to be a bit unbelievable.

I can’t help but wonder if people like Rod who go to lengths to stigmatize being overweight (and have concerns about their own weight) are being counterproductive. I would agree to some extent with the original NYT article and also with TNC: shaming is a solution to almost nothing. Shame encourages evasion and hiding more than seeking a solution. And because weight affects our perception of beauty, a focus on weight loss is often indistinguishable from a focus on increasing beauty. (A cursory glance at weight-loss advertisement is a pretty good indication of this.) I’d be much happier to see Rod – or anyone for that matter – talking more about health and activity than weight and fatness. For the vast majority of people, it is the altering of our lifestyle around healthy activities, exercise, and healthier and better foods that affect our overall health more than a single-minded focus on weight. I suspect there are deeply-rooted cultural and systematic reasons** why populations become overweight. But I also suspect the more we promote a single-minded focus on weight and attach strong stigmas to fat, the more likely we are to get our piousness and self-righteousness from attempting to attain some standard of beauty rather than promoting our health. I suspect that pursuing healthy activities and a healthy eating pattern first would lead to the net side-effect of smaller waistlines. Weight and health may correlate, but one does not cause the other; and the same for weight and gluttony, and even weight and beauty.

*The reason being I find it difficult to make a moral issue out of what one does to oneself – nevertheless one does not exist in a vacuum and even given that a person cannot morally wrong themselves, this woman’s actions could still be classified as immoral. Regardless of whether it is immoral or imprudent, this line of behavior I feel pretty comfortable condemning as a wrong way to live. The immorality of her boyfriend and her enablers, however, I am much more comfortable in asserting.

**The wide availability of cheap, government-subsidized foods (such as corn, which has given us the ubiquitous high-fructose corn starch) is one example of a systemic cause tending to increase weight. More difficult causes may be things like an increase in cultural activities that require sitting down, long periods of time spent in vehicles, and generally living in an environment radically different from the one we evolved and mostly lived in only ten thousand or so years ago.

As the Academy Awards continue to drone on in the background, I thought it might be appropriate to honor another type of movie: the really bad movie.

Recently I’ve discovered and succumb to Rifftrax. To understand Rifftrax we have to go back a little ways to Mystery Science Theater 3000. If you don’t know what Mystery Science Theater (MST3K to its friends) is, check out its wikipedia page real quick. The basic premise is a guy and some robots trapped on a satellite being tortured by a mad scientist by being forced to watch really bad movies. To survive, they make fun of (“riff”) the movies they’re watching. Well, Rifftrax is the same guys who were in the last seasons of MST3K providing their service for more modern, and often more successful, movies. You just purchase the rifftrax audio file (usually for $4), and sync it up with the movie (and they make it a relatively easy process), and sit back and enjoy.

I have seen four riffed movies so far: The Happening, Paranormal Activity, Twilight, and The Room.

The Happening is such a bad movie that not even Rifftrax was able to improve on it. Then again, I’d already had the displeasure of seeing the movie before, and in the theaters at that, so it may have been residual pain coming through.

Paranormal Activity is an independent movie that was excessively hyped as “one of the scariest movies of all time.” That is a lie. It is more boring than it is scary, and both of the characters are annoying and unsympathetic. Paranormal Activity is one of those movies that we have been cursed with since The Blair Witch Project that thinks it’s a cool idea to integrate the camera into the story. The problem with all these stories is there are too many scenes we should not be able to see because only a sociopath would be concerned with filming instead of attending to his fellow human being’s needs. It’s hardly a “horror” movie if you’re continually annoyed and frustrated at the main characters. Add that onto the fact that you will spend ten minutes at a time watching people sleep. The only action you will have to tide you over during these scenes will be a door moving slightly. Then after the couple wakes up, you will get to watch the douchey guy and his ditzy girlfriend watching the recording of the scene you just finished enduring on a computer, and commenting “look at that, it moved!” Repeat ad nauseum until the entirely predictable ending. Rifftrax made this film bearable.

Twilight was something else, however. Twilight is a very, very bad (and very high-grossing) movie, and the guys at Rifftrax made it hilarious. You probably know about the movie already, but the riffs, which came consistently during the utter inanity of the film, were a delight. It even tops some of my favorite MST3K episodes.

And to the last one: The Room. The Room is hard to describe because it is just so very surreal. It’s not supposed to be surreal, though, it’s supposed to be a drama. The Room was conceived of, starred in, and directed by a man named Tommy Wiseau. It’s unclear who Tommy is, although it’s obvious that his first language is not English, and it’s not clear that he understands human behavior at all. It’s possible that he’s an alien. Maybe an autistic alien. Certainly a very ugly one. And that ugliness will come back to hurt all his viewers. I didn’t know what I was getting into when I rented the movie and downloaded the Rifftrax, but there are three incredibly bad sex scenes. They are not sexy. They are not necessary to the story (although that implies that anything in this movie is necessary to form any “story”). The sex scenes seem mostly to be there to assault you with Tommy Wiseau’s naked backside, and they can actually be medically prescribed to induce celibacy. If you watch this, I recommend either looking away or going into another room and doing the dishes during these scenes.

Tommy also continuously greets people, in his vaguely European accent, with, “Oh hi, <name here>.” Every time; he just doesn’t disappoint. It doesn’t matter if they’re old friends, someone new, someone expected, unexpected. It’s all, “Oh hi, Denny;” “Oh hi, Susan;” “Oh hi, Lisa.” It’s just one of the many ticks in this movie that makes it so surreal. It’s hard to say exactly what it is about this rifftrax that is so hilarious, but it simply is. There was one point where I had to pause the whole ordeal and spend the next several minutes getting the laughs out. Definitely the funniest of the four rifftrax that I’ve seen so far. If you want to check one of these out but are a bit skittish about the sex scenes in The Room, you can definitely check out Twilight, which is second to The Room in my mind, and furthermore is much cleaner – inasmuch as a movie whose target audience is fourteen-year-old girls can be “clean” rather than “a crime against humanity.”

Anyways, that’s all. Just a fun thing for you to check out.