2009-2025 Arcalon columns

SOFTWARE SUCKS: old school PC rant



  
Evil is the most underrated concept. New types are invented daily, and there's no rock bottom. Worse futures than we can imagine will eventually be praised by the best propaganda. You don't have to believe in the devil. The world is evil because people agree with it.

Computer programs and web interfaces are staggeringly, incomprehensibly, insanely, monstrously, demonically evil.
They take too long to learn, are hard to use, and above all can't be relied on. The intense frustration and sense of malevolence that comes from using Windows for a few decades are like being endlessly stabbed by Bill Gates. Software just blatantly doesn't work so often that one must assume the programmers are evil. To put it in the nicest terms possible, the people working for Microsoft are all hypersatanic shitpeople. They don't want non-experts to have useful tools.

Forget about programs being more reliable or accurate than humans, when they can be less reliable than a drunk in a coma. Humans don't randomly destroy things, dispose of the evidence, and then forget all about it. They may do two out of three. At least vandalism has some nominal reason.
Windows introduced evolving levels of evil in inscrutable interface layers. Since the 1990s, my previously owned copies of Windows have crashed thousands of times. It may start with a crunching sound. If suddenly nothing happens on the screen, nothing might happen ever again. As I write these words on Windows 10, I'm waiting for Firefox to unfreeze in another tab.

PCs destroy data mendaciously. Windows can randomly and surreptitiously delete or destroy files. It can report it saved a file without doing so, or only saved a small part. Files backed up to other media don't show up when the backup is inserted in another computer. 'Irreversible' commands ruin the interface layout. Jiggle the mouse wrong and it highlights and deletes your life's work like it never existed. Windows will secretly scramble hard drives at the command of automatically downloaded malware.

Microsoft has the philosophy of an avalanche, which may be tolerable if you don't mind money rolling out of your pockets. There should be a criminal prosecution of Microsoft's chief "architect", but the legal system seems to make him untouchable, even though Microsoft has basically stolen money from millions of angry users.

It's not all Microsoft's fault. Others' browsers also won't allow users to read downloaded web pages offline. Video sliders don't work right. Highlighting a folder item causes the screen to freeze and jerk violently. Control panels shift with no way to restore them. Websites generate unstoppable floating boxes.

Worst of all is 'free' software. Open source programs (you know binaries don't you) are richly seeded with traps. New users are expected to learn through osmosis, watching experts from a distance.

At its best, software embodies miraculous stupidity. The simplest things are the hardest, full of occult rules: the current dialect, how to input data, how to change settings and activate hidden tools. A tiny mistake will permanently cripple functions.

The greatest insult is the 'Help' button. Missing is a list of all possible functions. Ideally, every program would only have a few functions when first installed.
Brutal startup screens are cluttered with inscrutable hieroglyphics and intimidating interfaces inspired by 1990s hacker movies. Stylish graphics only raise the stress levels.
Those in the know don't want to simplify things.

Make it so IT JUST WORKS. If necessary make the program worse until it does work. Make it so simple it does exactly what it should.

Most software becomes bloatware, a pain universe with hidden laws. Like touch typing, learning how to find features is agony for those without inborn talents. Users have to figure out everything for themselves. Programmers rely on the aftermarket and users to create 'tutorials' written for credentialed experts.

The most backward field is software training. Education has barely progressed since the dark ages. The greatest challenge in the world would be to find an easier way to transfer knowledge. Software only creates additional mysteries.
Society is stagnating while technology is elaborating.
Human brains are random accumulations of improvised solutions that are hard to keep together. Thinking gets harder when there are more choices, which explains the resistance to learning. People are so specialized that everyone is stupid in many ways.

Inefficiency generates disproportionate stress. When something doesn't work, it creates physical pain. People scream in agony, and the stress lasts for hours.
Strangely enough, this also explains why programs are so bad. The programmers give up as soon as they get the software to work at minimal reliability. Patches lock in underlying design flaws.

So many bad habits have accumulated that programmers will have to hit rock bottom. The most infuriated howls of hate never trouble Bill Gates. Like ending war, it would require a break in human nature. All current software should be abandoned in disgust.


1/6/09-3/12-3/14-9/15-8/18-12/22

the interface project: how to make computers suck slightly less



  
The most infuriating proof of the perverse nature of evolution may be the Interface Problem. It happens to be responsible for most things wrong with the world.
After bureaucracy, the clearest example is software (just one random evil is how you can't right-click on Yahoo Search results to copy text). As time goes by, complex computer programs seem to be getting more evil, at least for new users. Large programs are full of secret traps, though not all are deliberate. And they're not getting better.
This is completely separate from my endless experience of how computers keep freezing and crashing. The way webpages won't load because of all the adware is also demonic.
One clear example is called the 'Gimp', which is just about unusable. Cloaked in non-overlapping labyrinths of inscrutable obfuscation, gnomic obscurantists perfected a closed macrocosmos of subconscious barriers to make it impossible to edit images. It's easier to visualize the intersection of an eight dimensional tesseract than to figure out the hidden tool buttons.

Earlier versions of popular programs weren't completely impenetrable yet. Once, they were simple enough to figure out to some degree.
Those who have painfully mastered the secret art of programming may not believe in usable software. They had to suffer to make it this far, and now it's the new users' turn.

It's an educational problem. Why are most persons poor? Motivation, aptitude, and IQ are all genetic, but poor folks also lack information they can understand.
Currently the political left is demanding public education spending in the tera-dollar range, but ignores the fact that schools suck. The solution is what causes the problem. Schools don't just ignore controversial truths (like average racial and social differences in preferred learning styles and other things), but also all the undefined knowledge that students must learn indirectly. Learning barriers multiply as a subject expands. Few students enjoy traditional roundabout, contextual teaching methods. Perhaps achievement requires feeling superior to the other students. For some it takes pain to form permanent memories. Those who can't handle it fall behind.

Could there be a way to increase human IQ without changing human genetics? A method to perfectly describe subjects so that anyone could learn anything in a series of precise steps.
It would have to be like injecting vocabulary, databases, rule-sets and ways of thinking directly into the student's mind, whether they want it or not.
There is virtually no research, nothing to upset the status quo. Most high-skill or at least high-status professions require a decade of expensive, painful training.

The first test of a new education method would be a way to teach touch typing, a skill requiring the student to tie their hands into magic knots. Keyboards were deliberately designed to slow down typing.

There are only a few signs education can be improved, fewer ways to alter memory and perception without drugs:

  • The 'spacing effect' forces students to remember facts by repeating lessons at measured intervals but does not make learning easier.
  • It's possible to change long-term memories by re-imagining them.
  • Recordings played while the student is asleep may slightly help them remember word lists.
  • Hypnosis is a form of intense concentration which might help students pay attention.

    The mind remains unknown teritory, but it may be possible to manipulate how high-level information is integrated. By looking in a double mirror while performing motions, it's possible to cure 'phantom pains' in amputated limbs.
    For now, the most useful improvement may be to write the easiest possible explanation of every bit of knowledge, the ultimate metaphor for any subject.
    Software should be broken down into component sub-programs, all tools clearly separated for new users. Simple to understand interfaces would make programs easier, but with long lists of functions.
    Every program needs descriptions written by outsiders. The hard part is knowing what to leave out. The really hard part is to include what's too obvious to mention.

    There is a vast difference between discovering a skill, understanding it, and practicing it.
    The most skilled chefs are not the best food experimenters. Being able to understand science (or music), and actually doing them are very different. The educational-industrial complex provides only the second option.

    * Popular science documentaries are the most infuriating. Whenever they get interesting, they change the subject like Sarah Silverman. From 'saddle shaped' hyperbolic geometry to superstrings to DNA spirals, Discovery/PBS use the same few tired metaphors.

    The simplest possible description at any level must assume many things about the student. Some of the most neglected ideas could be introduced in comics and movies (like the public service campaign suggested in "The natural depravity of mankind", Lundberg). More detailed versions would have fewer readers at each level.

    * The opposite of science may be art, which also seems to be about making things more confusing. Most artistic insights are wasted. Readers might prefer articles condensing the knowledge of many books, but authors would outlaw such summaries.

    A Proposal:
    Someday, someone may invent a workable and above all usable human mind extension. The first version won't do more than record and edit short lists, but this simple tool alone could double productivity. Unlike a notebook, it would be impossible to lose, and the data would automatically be sorted and analyzed.
    What works for individuals may work for groups. Human skills are split among many specialists with little in common. The future may not belong to individuals, but to new types of groups.
    If better interface software can be invented, individuals may then be connected into large 'virtual minds'. No single member would fully understand what was going on, but the collective would.
    The danger is that once society becomes self-aware at some level, it may start making decisions none of its members understand. The ultimate outcome might be something like communism.


  • 1/6/09-3/12-4/14-9/15-8/18-

    The explanation project



      
    How do we know that physics isn't a practical joke? It's a genuine philosophical question, an unsolvable thought experiment.
    No one can explain relativity or quantum theory to a total outsider in a way that proves these things aren't made up. I may be such an outsider, though I've read tons of "popular" books and articles on the subject. If you can't understand the way these things are currently communicated, you have to take it on faith that light speed is 'constant', or that waves and particles or wavicles are 'interchangeable'.
    And it's not just physics.
    What precise terms replaced GOTO in software? There is an answer but it's impossible to find out. How can you write programs without something like a line command? Maybe software is divided into modules, but what exact instruction code tells the program where to go? As of 2018, it's not possible to get an answer to this question from Quora. Those I asked online replied with insults. Could this be a glitch in the Matrix?
    What does Python do? Why does hot air come out the back of a jet engine but not the front? There is a good reason. The worst place to ask is a Wikipedia talk page though Yahoo Questions are good for a laugh. AskReddit depends on getting to the front page.
    The system is set to handle approved questions only. Popular non-fiction books about software development are actually about entrepreneurs and their companies.

    It may be a reality failure. No one is supposed to think outside their specialty. Humanity seems crazy that way.
    Science is so elaborate that it has become as incomprehensible as nature. Instead of solving the mystery, it became part of the mystery. Instead of humans mastering nature, nature mastered humans.

    It's time for a new type - a higher or maybe lower level - of science that treats existing science as just another mystery. A research effort to reorganize all knowledge, reverse-engineer important results, and prove them in simpler terms.
    The next step would be a method to explain concepts in the simplest way possible, ideally in one sentence.

    Students should be 'reprogrammed' as painlessly as possible, with tens of thousands of data atoms and factoids combined any number of ways.
    Limited understanding of any subject is surprisingly easy, and would change education. The current purpose of schools is mainly to teach skills.
    This is all a very long way off.
    The problem may be overspecialization. Smart people can't explain their knowledge to outsiders. Sometimes they won't answer questions, or they may even tell outright lies. I'm not criticizing the infuriating Sokal Hoax - where a physicist made up a complex but fake paper about physics (incorporating concepts from the humanities) to prove that social scientists are easily fooled - only what it reveals about both sides of the Sciences/Humanities divide. It's easier to design an atomic bomb than to explain it.

    The core problem is that knowledge also makes people dumber. Overconfident experts absorb the maximum number of symbols they can handle. The more they know, the harder it is to see the big picture. It's hard to see hidden similarities from the inside.

    Society got it wrong by requiring immense dedication to learn anything. People or search engines or symbol predictors with superficial knowledge might recognize distant similarities better.

    Understanding is only part of the Interface Problem. Why don't you get job offers in the mail? Organizations should be competing to recruit members. If the problem of forming weak or strong social bonds could be solved, every other problem could be solved.
    The modular solution
    In many endeavors, the first step should be to remove less efficient group members, automatically making the remaining ones smarter. That would also make them cooperate better.
    Human knowledge might eventually be arranged in modules. Every insight from factoids to facts, metaphors, theories, textbooks, up to expert systems.
    The goal would be to create and combine enough modules. Higher modules could keep track of lower ones.
    This could even extend to the material world. Small hospitals or power plants could be designed and assembled from standard components. Streets and stores around the world would start to look more alike. With improved reliability, supply chains could become longer and more diffuse.
    A modular society would have unintended side effects. Some innovation would be discouraged for the sake of stability.

    Humans disagree about almost everything except certain taboos. The above proposals would take about a human eon of spare time to realize.
    The first step might be to voluntarily coordinate the process, something like a "2030 Automation Initiative".
    Better than a charity, its mission would be to reduce the need for human work as much as possible.
    Unlike for computers, the most meaningful human resource is time. This matters even more when the stakes are too low.
    Like rage and depression, stress is caused by the laziness impulse, the inevitable response to having too much to do for too little reward.


    1/3/09-3/12-4/14-8/18-12/22

    Toward a Grand Unified Social Theory

    The distant goal to combine all social knowledge, observations and trends in a single model. A Top Down diagrammatic approach.


       There has been a big drop in scientific plots in science fiction stories since the 1960s, the Golden Age of the Boomers. There have also been fewer philosophical stories since the death of Stanislaw Lem (except by Eliezer Yudkowsky, Scott Alexander, maybe Greg Egan and a few others online).
    Most fiction (like World War Z) is bound by genre conventions. In the profitable Potter franchise, why don't the magical creatures rule the world?

    Average intelligence seems to be declining while economic growth sputters ons. The most important questions can't be asked. Robotic conformity masquerades as discipline. The education-industrial complex is too self-important to be questioned.
    One reason for the conformity is that the earth's population is too small for interesting networks to form. Not everyone can find a community for their interests. That would take a population with fifteen zeroes.
    On this planet, we must make do with an average population of forty million per country, who speak on average ten languages, mostly obscure. Making them occupy the same space requires many suppressions.

    Most human effort takes place to stop attempted progress. It's safer to stay the same. Change is almost always bad in the short term. There are few resources, and most investments fail. The more precarious the society, the less deviation is tolerated. Limited experiments create new balances after epidemics, wars, famines and the like.

    People resist learning new things because humans who evolved to find learning unpleasant have left more offspring.
    This conservatism also explains Third World bureaucracies. Getting government permission to do anything in Africa is as complicated as adjusting state boundaries. According to the now very unpopular 'alt-right', lower average genetic IQ rates are more than balanced by many other Q factors, like dominance and passive adaptability.
    Life in 'rich' countries is about increasing credentialism. As income lagged, Western civilization became over-regulated.
    Bureaucrats are everywhere. Every job requires permits. You need to produce a birth certificate to listen to the radio - or at least to SiriusXM satellite radio, which only accepts bank payments, which require a government ID, which requires a birth certificate. It's no longer possible to build new nuclear power plants or airports. Benefits and entitlements once granted can never be withdrawn, a ratcheting racket of obligations.

    Complicated systems commandeer expanding support networks. The interchange of highways around a large airport can seem almost as massive as the runways or terminals, and are as expensive to permit and build.

    Perhaps no human can create something genuinely new anymore. Success is rearranging existing resources. Let others spend their lives trying to evolve one good idea. The 'real' genius is exploiting that idea, and it's also more fun.

    Solving big problems seems to take forever, and then the difficulties are hidden. After the false leads are eliminated, the resulting model implicitly incorporates all the things not to do. This is for experts only. The final step of simplification is just too hard. Who wants to make the added effort?

    The worst example is software. Only a tiny minority of users is capable of moving the output of one program to another, or automating common tasks. Millions of frustrated punters keep performing mind-numbing functions manually, wasting thousands of mice-clicks because they can't tell the computer to complete the job in milliseconds.
    Most experienced users only think they are experts, because after many years they've learned to work around the sometimes deliberate barriers. That's just how the real experts like it. As mentioned over nine times in this essay series, these people are evil.

    It would take a new derivative science, almost a natural philosophy, to bridge the gap. Knowledge should be oversimplified in the best way.
    Special Relativity is never properly explained because it seems uncomfortably strange to normals. But even principles first learned in kindergarten are equally misunderstood: for example the difference between kinetic energy and momentum, the ultimate potential of a lever (the weight of a flea could theoretically accelerate a galaxy to near-lightspeed in nanoseconds in a sufficiently strong gravitational field), or even the laws of motion. Scifi movies always get it wrong.
    No one grasps the big picture. Barriers are lowered by making things more complex.

    In the distant past, anyone who tried to measure the earth by walking around it would have been killed one hundreth of the way around. World travel wasn't possible until the sixteenth century. The world remains fractured.
    The reason Tutsis can't connect to Hutus is that they inhabit different realities. At best, they appear boring to each other. But boredom is a form of stress.
    Racism and culturalism are universal because they're easy. An individual merely has to focus their awareness.
    In practice, one group tends to dominate the other if they occupy the same space. Eventually they may negotiate a balance, though some thrive on chaos.

    Simple emotions drive history. War is mostly about retaliation. During World War 2, when Germany was being invaded from all directions, why did the Nazis keep fighting when they knew they would need to rebuild?
    It would have been more painful to surrender and let the victors run wild. Better to weaken them first, so they would have less energy for creative disruption.

    Large bombing raids like on Dresden didn't really help, but they felt better than bombing power plants. Instead, a portion of Germany could have been turned over to its victims after its defeat. Israel could have ended up next to Belgium. This could have been an incentive to surrender sooner.

    Incentives are strange. Human behavior is controlled by pain aversion, the urge to achieve a lower profile or hide in the tall grass. That doesn't motivate anyone to work harder. Those who have already been punished feel they've earned the right to relax.

    At the base of society are families. The most common favor trading involves some coercion. Unexpected pregnancy can radically simplify a sex partner's future.
    There's a genetic motive for offspring to be skeptical of their parents (while closely studying how their parents treat others). Parents are biased and manipulative to maximize their own genetic potential, not their offspring's. Traditionally, they often barter one offspring in marriage so another offspring can gain a better marriage partner.
    Like most limits to freedom, this problem won't be solved with laws.

    The core problem of politics is that any entity with a monopoly on force is likely to use it wrong.
    Could there be a better way? The USA is the only country that has been both open yet resistant to outside influences.
    In some ways it's been a closed world, existing outside history.
    Freedom could be achieved by doing less wrong things, especially by governments. Planned economies have about one tenth the output of owner economies.

    * A political solution is to subdivide power among competing groups or individuals, so no one would have too much power.
    That would require a widespread desire to be left alone. It might happen if enough people could experience steady wealth accumulation.
    It would not necessarily cause a more fragmented society, if people could link up and exchange services in more ways.
    The Net could be a shared reality, hiding cultural barriers and differences. In some imaginable futures, there will be only one website, but no one owns it, and it can do anything.

    Humans have not evolved to live free. If freedom is forced upon them, they will try to reduce it, subsuming their identity within groups. No group can have too much power, so competing groups keep forming. The best possible human future would be like some eternal soap opera.


    2/6/10-2/12-5/14-8/18-12/22

    software: the evil interface


       Hate is a complex emotion. It can destroy the hater while leaving its subject (thief, tyrant, rapist, sadist) completely unperturbed. Hate includes a component of dread not seen outside of illness, famine, or torture.
    In the real world, evil usually wins.
    Often, hate is the only appropriate emotion.

    This article is about solving software. Computers remain so unusable, unreliable and outright defective that many users would sentence the programmers to death if they could.
    This is by design. These people are actually evil. To me computer programmers are the world's top devils.

    * The problem is controlling the program, persuading it do any specific task.
    Programmers are 'bottom-up' experts who care less about 'top-down' interface design. They deliberately make their users dependent by forcing them to learn proprietary interface methods, while making them pay to upgrade often. Agony breeds loyalty.
    Windows 8 & 10 programmers deserve to be shot after a fair trial. They think of their software as a community or a lifestyle instead of a tool. They are tools.

    * The first interface barrier is the keyboard. This pain is also known as 'touch typing'. Keyboards were invented before touch typing but adapted for this dark art, making them impractical for hunt and peck typers. A billion unskilled typists keep hitting the wrong keys, which might not happen if they were bigger or further apart, or had barriers between them, or were angled differently, or any other innovation. Touchscreens are worse though.
    Learning touch-typing can be harder (often impossible) than human skills like quantum interpretations or Basque palindromes.

    * Online user interfaces are arbitrary.
    Like the world, the Net remains shockingly primitive (see PayPal or Bitcoin), defined by what a user can't do. Tracking visitors, web layouts, recording what's onscreen (the DMCA doesn't like that at all), and data backups are deliberately made agonizing for non-experts. Each tool is a trial by fire.

    * The greatest censorship innovation is the network rule that bandwidth must be paid for by the content provider instead of the downloader. Around 2004 came the Web 2.0 of irritating 'rich' media and 'streaming' content to better control information. User options have been stealthily reduced to make users more dependent. Mobile devices are beyond crap if you want to control your own data. Individual websites of the 1990s are an ancient memory. The Net now exists for the benefit of social and corporate media collecting user data.


    * In many ways, the future will suck. Inefficiencies will only increase as society becomes more complex. Users won't even notice as barriers become mandatory standards, like how dictionaries don't include brand names. For some reason, almost no one complains about ever more bloated Javascript and adspam encrusted websites.

    * To become less malevolent, software should become dumber, but with more lists. Dumb is easy: a simple but predictable personality like 'Rain Man'.
    The future will be chaotic. Now is the time to consider standards based on things humans still have in common.
    Ideally, there should be only one program environment, a format to connect all data, simplifying interactions like a nervous system.
    The web could be divided into "data atoms" that could be reassembled as needed (each element text-readable). All files could be described through a universal comment layer. Every link could have independent descriptions. Self-updating, text-based link lists, and so on . . .

    * Simple solutions are better. Software designers shouldn't even talk to each other, but communicate by text to simplify their thoughts. Linear communication makes it harder to squeeze in hidden assumptions.
    Instead of repetitive clicks or swipes, software should work faster with face scanners, muscle interfaces, and eventually brain interfaces.

    * Everyone could evolve their own interface, instead of a standard OS packaged by a secretive corporation.
    Applications should simplify themselves in response to each user, by backing up and sorting old content, arranging bookmarks and file hierarchies, and identifying common tasks and describing them in flowcharts, with the goal of fully automating them.

    * This is a long way off. At the moment, society is splintering faster than software is improving.
    The solution for humans to cooperate may be to lie better - i.e. diplomacy. Manners took most of history to evolve, but can be forgotten much faster.
    Social taboos and delayed gratification are necessary to share space with people even with similar goals and outlooks. Unfortunately, that won't work anymore when diversity increases faster than rules can be learned.

    * This can already be seen online. The Net is fracturing like the real world.
    Japanese and Chinese web services are only available to users with Japanese emails. What's happening over there? Social networks invent their own codes. If we're lucky, the first virtual nation may start as some exclusive online environment. The first virtual war will begin soon after.


    * The penultimate goal of computer science (before human-level AI) should be life management software. It would help people cooperate by identifying traits, norms, and cultural aptitudes held in common by different users.
    It would set the user's schedule, track every activity, test their personality, and try to improve it. Every user would be constantly analyzed and compared to standard personality profiles, based on their interests and habits.

    * Ideally, everything would become functionally transparent and user serviceable. This applies to their physical environment as well.
    Simplifying to the extreme, future geeks may inhabit live-in boxes with exposed utility pipes and stilt-mounted foundations: a simple paradise compared to the unpredictable lives people have endured throughout history. Instead, they would enjoy some sort of information-existence in cyberspace, still a very remote dream.

    * One problem with integrated life control software is that it could be used by the powers-that-be to monitor and enslave their subjects.
    The only way to prevent these records from being successfully subpoenaed would be to encrypt them. This could be done by auto-compilation: making the life-management software totally dependent on each user's unique personality and habits. The files and settings would make no sense to anyone else.

    * Eventually, privacy concerns will become irrelevant. If history is a guide, most people will want to join the Borg Collective.


    2/6/10-4/1/11-5/14-9/15-12/22-07/25