More about the iOptron SkyTracker:
Performance measurements and polar alignment charts

[Updated based on further tests.]

This material will probably go on one of my astronomy pages later, but for now, I wanted to put it somewhere, and that's what the Daily Notebook is for — notes!

Brief performance measurement of the SkyTracker:

Total peak-to-peak periodic error: 60 arc-seconds

Period: 552 seconds of time [corrected 2018]

The measured tracking error pattern is close to a sine wave. This implies a maximum tracking error of 20 arc-seconds per minute of time.

That means that if you want no more than 1 pixel of tracking error on a Canon 60Da, you can expose up to 60 seconds with a 50-mm lens, or 30 seconds with a 100-mm lens, or 15 seconds with a 200-mm lens. In practice, the likelihood of getting the maximum error is very low, and you can safely expose twice as long if you are willing to examine your pictures and throw out any that are badly tracked (or use software that does this automatically). Also, if you plan to downsample your pictures to perhaps 1/4 their original size, you can obviously expose 4 times as long, or more. The limits that I've calculated apply to images viewed at maximum magnification.

Bottom line: The SkyTracker is at its best when you're doing sub-exposures in the 30- to 60-second range and stacking them. The longest lens I plan to use with mine is 200 mm. I've bought a vintage Olympus Zuiko 200-mm f/5 lens for the purpose (to use on the Canon with an adapter); I chose it for its light weight.

Polar alignment: Extremely good polar alignment is easy to achieve. The Polar Scope reticle has numbers on it like a clock face, and to find out where to put Polaris on the reticle, you're supposed to run an iPad app. (Go to the iTunes store and search for "iOptron.") For those of us who don't want to take an iPad into the field, I offer two alternatives. One is a chart that you rotate to line up with the visible position of Cassiopeia and/or Ursa Major in the sky; the other is a chart of numbers by date and time of night. Other iOptron mounts use the mount's own computer to show you where Polaris should go, but the SkyTracker doesn't have a display and keypad.

Click for PDF

Click for PDF

Note (in response to a correspondent's question): These charts are valid for the whole Northern Hemisphere, not just for Athens, Georgia.

Permanent link to this entry

First light with the iOptron SkyTracker

Caution! The purpose of these pictures is to detect imperfections, not to produce good pictures. I am deliberately pushing this instrument to its limits so I can measure them.

I've just bought myself an iOptron SkyTracker in order to do wide-field astrophotography without lifting heavy equipment.

I'll say more later about how it's used. Brief impression: It is solidly built and in fact has a 3-inch-diameter spur gear, probably the same one as in other iOptron mounts. Polar alignment is very easy with the "polar scope." It is apparently built with Bogen/Manfrotto tripods in mind and almost seems to be styled to look like a component of one; I used it with a Bogen tripod and a massive Manfrotto ball head.

Hint: Drape the camera strap around the tripod. The camera can fall off if two thumbscrews are loosened.

The picture above is a 3-minute exposure of southern Orion with a Canon 60Da and an Olympus 100-mm f/2.8 lens at f/11. The lens was chosen for light weight and stopped down to f/11 so I could go for 3 minutes in bright moonlight.

The upper picture is cropped and reduced, and the lower one is cropped but reproduced pixel-for-pixel (corresponding to a gigantic enlargement of the whole frame). You can see a total of about 3 pixels of tracking error in 3 minutes. That is 25 arc-seconds, not bad for a mount that provides no way to make guiding corrections. Normally I will use 1-minute sub-exposures for less than 10 arc-seconds of error. Indeed, tracking in a companion 1-minute exposure was practically perfect.

I am in the process of making a chart that shows where to place Polaris in the iOptron reticle, relative to the positions of the constellations in the sky. iOptron says you're supposed to get an iPad app for this, but a map on a piece of paper works just as well.

Permanent link to this entry

Unburdening the bookshelves

I spent part of the afternoon weeding out the bookshelves in my office and making donations to a fund-raising book sale that will be held by the linguistics graduate students.

My heuristic was that if I hadn't used a book in five or ten years, and it was not a keepsake or an irreplaceable item, I didn't need to keep it. In a fast-moving field like computational linguistics, old books really do become obsolete.

The other thing going on is that most scholarly periodicals are now archived online; I can read them from a Web browser, anywhere, any time. That means I don't need to keep my own copies.

I decided to keep my very first ACL conference proceedings volume, and in it I wrote on its title page a checklist of the other ACL meetings I had been to. The next one is in Atlanta this June.

Permanent link to this entry

Jupiter with the 5-inch

Still unable to set up my 8-inch telescope due to a rib injury in October that will take months to finish healing, I decided to take a shot at Jupiter with my vintage Celestron 5, using a Meade 3× Barlow lens and a Canon 60Da in movie mode as the image acquisition device.

This is a stack of the best 5000 of about 7300 video frames, stacked with RegiStax 6.

RegiStax at first misaligned some of the frames, producing a ghost image mostly overlapping the main one, but slightly offset. The cure was to set the alignbox size to 130 instead of 30 and work with just one alignment point, thus making RegiStax 6 work like RegiStax 5. Varying the number of alignment points in other ways (such as using a lot more than the default) is also said to help with this problem.

Permanent link to this entry

Linguist in action

This is what I look like in action, in the classroom. I'm diagramming the structure of an English sentence as it might be analyzed by a computer.

Thanks to Shu Zhang, graduate student, for taking the picture.

Permanent link to this entry

Should rich people pay any taxes at all?

[Revised.]

In what follows I'm going to perform a reductio ad absurdum. I'm going to take an oversimplified idea and run with it so far that everybody can see that it's oversimplified.

Readers who are not easily bored may have read my long series about economic fairness a couple of years ago. Today I want to address something we're being told often: taxes on the rich must remain low so that the rich will have an incentive to do business activities that create jobs for the rest of us.

Obviously, if taxes on the rich were too high, the rest of us would be in trouble. Put a 100% tax on millionaires, and you would simply abolish most medium-sized businesses. It was JFK who pointed out that when marginal income tax rates are too high (and at the time our highest was about 90%), lowering them can raise government revenue because then there's more business to tax.

Fast-forward to 2013, and many people are advocating a flat tax, which means that the income tax, as percentage of income above a threshold, should be exactly the same for the very rich as for the middle class. (A threshold can be included — such as not taxing the first $10,000 — in order to give some relief to the poor. A flat tax that does not have a threshold is called a proportional tax.)
[I thank Douglas Downing for a terminological correction here.]

Let me throw another idea into the pot. If the very rich are the ones who create jobs — and if we want them to have as much incentive to do this as possible — then why shouldn't their taxes be lower than ours?

And if you really want as much incentive as possible, why shouldn't their taxes be zero?

That would maximize their incentive, all right. All income taxes reduce people's incentive to work. If we think the very rich need the most incentive to work, because they are our most productive people, then why not exempt them from taxes altogether?

What is wrong with this picture? Clearly, we don't actually want to exempt the rich from all taxes. Why not?

The answer, I think, is that we believe that citizens have a civic duty to pay taxes. We don't want any citizen to be exempted from taxes, nor to pay taxes at an abnormally low rate, unless paying taxes would cause them undue hardship. That's why we advocate tax relief for the poor, not the rich.

So what is fair? Tax everyone by (1) the same amount of money (like an old-fashioned poll tax)? Or (2) the same percentage of their income? Or (3) at a level that causes the same amount of (regrettable, but unavoidable) hardship on citizens of different economic classes? That would imply taxing the rich at a higher rate, simply because they can spare a larger fraction of their income.

I think we can rule out (1) as simply unworkable. How to judge between (2) and (3) is not so clear. Not to mention other factors, such as the fact that the rich are better at sheltering their income from taxation, so even if a flat tax is what you want, a higher nominal tax rate might be what you need in order to achieve it.

How heavily you can tax them without significantly reducing their incentive is an empirical question. Only actual studies of the economy can answer it.

Permanent link to this entry

The research paper that upset the Tea Party

Fox News reported on the 15th:

The story goes on to give a fairly detailed and balanced account. Researchers at the University of California at San Francisco, funded by the National Cancer Institute, set out to analyze the tobacco policy of the Tea Party and found surprising connections to Big Tobacco.

It is important to note that although supported by federal research funds, the researchers were not government agents and were not told to investigate the Tea Party. Rather, they are public health researchers, and they make it clear that the topic of their paper was not assigned to them.

The paper is detailed and scholarly. It cites evidence from numerous sources. It was published in a peer-reviewed journal. I got it from the University of Georgia Libraries and took a look. For copyright reasons, I can't post it on the Web, but I can e-mail copies to a few interested individuals; e-mail me if you want one. You can read the PubMed abstract here.

Now there are several questions we might ask.

Are the paper's conclusions correct? I have no way to know. But in the arena of peer-reviewed scientific research, you can't wave your hands if you lack evidence, and you can't pass off opinions as facts. This isn't talk radio. Playing fast and loose with evidence will ruin, not make, a scientist's career.

Is it proper to spend public health money mapping out the Tea Party's tobacco policy? For many people, this is the sticking point. Aren't those people supposed to be investigating cancer, not politics?

But think about that for a moment. Smoking and secondhand smoke are public health hazards. That is uncontroversial. Public health research includes public health policy research — helping to make better policies to protect people — and that, in turn, includes studying the thinking of groups that advocate policies.

As I understand it, the researchers didn't set out to discredit the Tea Party; they just wanted to understand its position in their area of expertise — tobacco policy — and they stumbled on some connections that struck them as scandalous.

Bear in mind that governments, including ours, tend to protect the tobacco industry from the consequences of its own actions. If you're a tobacco company, it's legal for you to harm people or encourage them to harm themselves. We let tobacco companies do things we would never tolerate from a patent medicine quack or an environmental polluter. Admittedly, we don't tolerate everything — there have even been criminal convictions of tobacco companies for deceiving the public — but it seems still to be politically incorrect to act directly on the knowledge that tobacco kills.

Besides, a scientific study is not like an FBI investigation. This whole thing is farther from "the government" than most people realize. Government-funded research is commonly done by faculty members at universities. A lot of it concerns society's problems and is politically touchy. For example, almost all economics research either supports or discredits some government policy. Then there are criminology, political science, and international relations. (Do you yank funding from a criminologist who says a certain political party's law enforcement plan is all wrong?) Government-funded researchers are not forbidden to study anything that's out there. If they were, essential research would not get done. Thank goodness for freedom.

(One last thing. As a scientist who has administered grants, I can confidently say that although it may have been part of a $7 million grant, this single paper didn't cost anywhere near $7 million, in itself. It looks to me like something on the order $30,000 worth of effort, about as small as a research project can be. I don't know the researchers and don't know how accurately I've guessed.)

A third question is, regardless of whether the study should have been done, now that we have the results, what should be done with them? Should scientists be penalized when they uncover something apparently scandalous? Of course not.

Tea Party, be very careful here. It looks as if you're objecting to funding of research that might uncover unsavory facts about you. And that is a quick way to claim the moral low ground.

Personal political note: Although I label myself as a (moderate) conservative, the Tea Party struck me as shallow and artificial from the beginning, and I never supported it.

Permanent link to this entry

Facebook safety made really simple

There has apparently been another minor change in Facebook's privacy policy. There are also reports that Facebook has been (or can be) broken into by some kind of malefactors. Panic and misguided advice are spreading. So let me give you my own advice about Facebook safety.

It's simple. Never put anything on Facebook, or anywhere else on the Web, that would be seriously damaging if the wrong people saw it.

The World Wide Web is public. Forums like Facebook normally give you the ability to control who sees what, but they are not perfect. Anything you transmit could get into unexpected hands.

(Actually, the most common way privacy is breached is that the recipient of your message forwards it to someone you didn't expect. That's easy to do.)

All this is closely related to the advice I saw in an 1890s business handbook: Anything that is written down may one day find its way into print. Same principle, different technology.

Permanent link to this entry

A point on which all modern medical research may be wrong

I can't remember whom I heard pointing this out in some public forum recently; it's not my original idea. But it certainly resonated with me and I saw immediately that it is mathematically sound.

Modern medical research pretends there's no variation between individuals.

Medical research invariably relies on randomized controlled trials, which means you divide your patients up randomly into two matched groups and give them different treatments. Then you measure how many people improve, or what the average amount of improvement is, and compare groups.

The logical problem here is that it only tells you what treatment is best for the majority. If there is a sub-population that needs a different treatment, a randomized controlled trial will never tell you.

Yet we know there is a lot of genetic variation between human beings. We don't all have the same body chemistry. But the effects of variation are only discovered clinically: every doctor knows (for instance) that not everyone needs the same statin, or the same antidepressant, or the same diet.

It would be good if every randomized controlled trial were followed by attempting to find out what is different about the patients who didn't get the same results as the majority. [I hasten to add that there is starting to be considerable interest in this, in connection with genetics and mechanisms of gene expression.]

And of course individualization is a major part of "alternative health care" (chiropractic, acupuncture, homeopathy). About different alternative treatments I have different feelings. I don't believe all the claims of chiropractic, but in many cases it is simply very skilled physical therapy. I don't trust homeopathy because it seems to deny obvious facts about chemistry. I do believe in acupuncture, partly because of the way it was rediscovered, as "dry needling" — a researcher was trying to find the minimum amount of cortisone to inject into painful muscle knots, and he found that it was zero — in many cases a needle with no actual injection would do the job.

But my point is, alternative health care practitioners do a great deal of individualization, and even if their actual theories are wrong, they probably help patients find things about their lifestyle, diet, or medication regime that are disagreeing with them as individuals.

Food for thought.

Permanent link to this entry

A car problem I'm glad I didn't have

Today, while stopped at a traffic light, I saw a Jeep SUV having an automotive problem I've never seen before. One of the rear wheels was still turning (forward, if memory serves me right) and, literally, burning rubber.

What was it? Combination of stuck throttle and lack of brakes on one wheel? Or what?

Permanent link to this entry

How men and women can communicate better

Today, for St. Valentine's Day, I tackle another of the world's great problems: male-female communication.

This blog entry stems from a recent Facebook conversation with several young people, mostly women, one of them an extended family member. I politely offered my services as an old man who happens to have a Ph.D. in linguistics, and who has been studying human communication professionally for a long time. They accepted my offer. I've been happily married for 30 years and remain on good terms with two daughters and a granddaughter, so I must be tolerably good at communicating with human females.

Let me first state my belief that, contrary to a notorious book title, men aren't from Mars and women aren't from Venus. Men and women are from Earth and ought to be able to communicate. I have no time for prissy games in which people pretend that communication is impossible. Nor for people who deliberately misunderstand things so they can make a show of being offended.

I do not think there is an inborn barrier to communication between men and women. I do think there are some differences, and they're mostly cultural. I really don't know how to tell to what extent they may be inborn, and I know that cultural influences last for centuries past the time you think they died out. I'll just concentrate on what the differences are and how to overcome them.

One difference is that, through centuries of culture, men are more attuned to communicating with strangers, and women are more accustomed to communicating with the members of their own household. That means women have higher expectations. They expect other people's minds to work like their own and expect communication to succeed. Men are more prepared for failures of communication and are not upset by them; they just correct them (gracefully or clumsily) and truck on. Women are more likely not to notice that a failure has happened.

A second difference is that women, more than men, use speech to show and receive attention, not just to exchange information. To women, more than men, silence is insulting and the quantity and occasion of speech can matter as much as the words.

But the third, biggest, difference is that women are more sensitive to indirect speech acts — to messages that the hearer is supposed to infer (from how something was said or whether something was said, not just from the information being expressed).

This cuts two ways. Women send signals that men don't recognize. And women sometimes recognize signals that aren't there, by reading too much into the way something is said. In the Facebook conversation that I mentioned, one of the women asked, "Why do men send mixed signals?" and as she explained the situation, I had to reply, "How do you know he was sending any signals at all?" One of the other participants had an insight: "I just realized I need to make more of a distinction between his actual words and the story I'm making up in my mind." In particular, if you're guessing, remember that guesses can be wrong, and be prepared to back up and change your guesses.

In fact, that's my main recommendation for both men and women: Always distinguish what's being said from what's going on in your mind. Don't imagine you've expressed more than you really have, and don't imagine the other person has expressed more than he or she really has. And remember that it might have meant something other than what you thought.

Melody, with whom I discussed this, has further advice, especially for women: Don't act like characters in TV sitcoms or soap operas. Much of the comedy or tragedy comes from miscommunication, and you don't want to be one of the people who miscommunicate!

I would add one more thing: Don't feed your mind garbage from "women's" or "men's" magazines. Neither "Ten reasons to imagine your husband is cheating" or "When she says no, she always means yes" is going to be edifying reading. Both are quick ways to get out of touch with reality.

Permanent link to this entry

Is postmodernism one of the causes of our political strife?

I've repeatedly pointed out my dismay about how American political discourse has gotten more and more polarized and strident. Instead of seeking facts and wisdom, people just rant at each other and splinter into fringe groups.

Today I had an idea why that might be happening. It might be because too many people no longer believe in objective facts. When you don't believe in objective facts, everything is just taste and opinion, and all debates become power struggles. Instead of seeking information, you just root for one side or the other, as if competing ideas were rival sports teams.

Why don't people believe in objective facts? To answer that, I need to describe, as briefly as possible, two philosophical movements that have greatly affected our culture during the past century.

The first one, modernism, is basically the notion that traditional wisdom is worthless and everything needs to be reworked scientifically. Inspired by the great advances of science since 1850 or so, modernists called for throwing out everything old-fashioned and reinventing everything: art, music, religion, morals...

Modernism was very much in the air in the 1960s, when people were talking about the "new morality" and thought it must be good because it's new, even if nobody quite knew what it was. The modernist takes for granted that you cannot believe what people believed in 1700; they were all wrong because they lived in "prescientific" times.

The arrogance of modernism provoked a reaction, which is called postmodernism. Where the modernist says, "Now we know it all," the postmodernist says, "No, we don't. We're not sure of anything." The key doctrine of postmodernism is that we see everything through the preconceptions of our own minds and our own culture.

That implies that there are no objective facts. All we have are our own individually prejudiced viewpoints, none of which is really any better than any other.

To this I would reply: That's a good insight, but not to be swallowed whole. Yes, we have our own viewpoints, but objective facts can overcome them. That's why we call them objective facts. If that were not so, we would never have been able to learn anything at all.

Postmodernism is the source of the notion that it's unfair to say that anything is objectively better than anything else, that if you do, you're just "showing your cultural prejudices" and that you shouldn't "impose your beliefs" on others.

You see the logical problem, of course. How can it be objectively true that there is no objective truth?

If you think postmodernism leads to tolerance, think again. I believe in tolerance because I believe in objective facts; the truth is out there, and we can get closer to it by debating it. Postmodernists don't see it that way. To them, all debates are just power struggles, and there is no objective reason why anybody should tolerate any dissent, much less actually look for facts that might conflict with their opinions. Just retreat into a fringe group that supports whatever you feel like believing, and try to shout down all your opponents.

Permanent link to this entry

1999 Isuzu Rodeo and Honda Passport:
Speaker replacement

Click here for instructions with pictures.

Use 4-inch speakers in the rear. Remove the plastic lens of the dome light that is at the top of the tailgate. Then remove the Phillips-head screw. Then simply pull the panel off (it is held in by snap connectors) for access to both speakers.

[Corrected Feb. 15] The front speakers are more problematic. There are tweeters high up in the doors, accessible only by removing the door panel. The woofers are down low and easy to get to. They have the hole positions of 6 1/2-inch speakers but are not full circles (instead they are a "bulging square with ears" shape like 5 1/4-inch speakers). Standard 6 1/2-inch speakers do not fit and 5 1/4-inch speakers are too small. You can install 5 1/4-inch speakers by making new screw holes in the plastic, at the top, bottom, left, and right. However, you will then have poor baffling from a ring-spaced open space around the speaker, and this will reduce bass. (You could make the circle smaller and solve the baffle problem by adding a strip foam rubber like what is used around window air conditioners.) We opted not to replace these speakers at this time. Should you do so, the cover in front of each speaker (on the front doors) simply pops off; there is a slot for a screwdriver at one bottom corner.

The alert reader will have gathered that we're fixing up Jim Mauldin's Isuzu, which went a couple of years without being driven much. One of the casualties of time was that the rubber surround of the rear speaker cones had completely crumbled away. The original speakers themselves are cheap ones with single cones.

Permanent link to this entry

Very short notes

Recommended: UltraEdit, an editor for text files of all types, including extremely large ones. It is also a hex editor for binary files. Unlike Notepad++, it can be told to expect incoming text files to be UTF-8 (rather than Windows-1252). In Notepad++, you have to supply that information after the file is opened, each time.

Minor surprise: In the University of Georgia Libraries online periodicals catalog, if you look for JAMA (Journal of the American Medical Association), you also get every periodical, past and present, in all of Jamaica. It assumes you want it to complete the name...

Permanent link to this entry

Unicode follies
The dreaded EncoderFallbackException, or,
Why some C# character strings can never be output to files

[Revised for clarity.]

Internally, C# and Java use Unicode to represent characters. Unicode is a replacement for the older ASCII code, and it includes the writing systems of all the world's languages.

Unicode was originally designed to be a 16-bit system with 65,536 character codes ranging from hex 0000 to FFFF. Internally, the char type in C# and Java is still 16 bits. And thereby hangs a tale.

The designers of Unicode decided that 65,536 characters are not enough — even though that's enough for all of Chinese and Japanese, as well as all the other living languages.

So they added on a "high range" that allows a total of more than a million characters. In 16-bit (UTF-16) Unicode files, and internally in C#, these are represented by pairs of 16-bit codes in the range D800-DBFF (for the first one of the pair) and DC00-DFFF (for the second one). These are called surrogates. There are 1024 possible values for each of the two surrogates, giving more than a million combinations that can represent high-range characters.

Now consider what happens if, in a 16-bit Unicode file or a C# character string, you get an invalid use of a surrogate, such as DC00 not preceded by something in the DB00-DBFF range.

If you output it as UTF-16, the computer may never notice the problem. But if you output it as UTF-8, or possibly if you convert it to something other than Unicode, things go wildly wrong. In C#, you get an EncoderFallbackException. What's worse, if you're writing to a file, this exception is thrown when the buffer is flushed, not when the invalid code is sent to the file, so you may never know exactly where the invalid code was!

(I should explain UTF-8. It is a compact format for Unicode files that consist mostly of ASCII characters. It's the usual kind of text file in Windows in English and other European languages. Characters in the range 0 to 8F, the ordinary digits and letters of the alphabet, occupy single bytes; higher codes are represented by sequences of bytes. C# automatically interprets a UTF-8 file as a sequence of 16-bit Unicode characters on input and maps it back on output.)

Anyhow, here is a simple example of how to crash a C# program by creating invalid Unicode:

string s = "\uDC00 is invalid here";
System.IO.File.WriteAllText("invalid.txt", s);

Can your program actually do this? One of mine did. It was reading a file of characters and manipulating them, without paying any attention to the possibility that there might be a high-range character (a two-character character). The two parts of a two-character character were getting split apart, so that later, they appeared in invalid contexts.

Solution? When reading a file, before processing the characters, look to see if any of them have

Char.GetUnicodeType(...) == System.Globalization.UnicodeCategory.Surrogate

and if so, convert them to question marks or something. Then you've lost support for the high-range characters, but unless you're processing Egyptian hieroglyphics or something, that's no loss. If you actually want high-range characters to enter your program as characters, you'll have to do some more programming, and store them in a data type other than char.

Final note: This whole problem applies only when writing out to a Unicode text file. Typically, invalid surrogates can be written to the screen or to non-Unicode files with no problem; they are replaced by symbols that stand for unknown characters.

Postscript: I am thinking of starting a Campaign for Pure Unicode in which codes go only from 0000 to FFFF, and codes in the range D800-DFFF are simply treated as invalid. Who will join me?

Permanent link to this entry

Secondhand smoke kills

A recurring theme in recent medical research is that breathing secondhand tobacco smoke is much more harmful than we thought, and this wasn't known until enough places went smoke-free that we could compare them to the ones that hadn't done so.

Look for example at this. A smoke-free work environment cut the incidence of heart attacks by 33% in this example.

With statistics like this, I expect that all workplaces will soon be smoke-free because no insurance company would allow it any other way. An employer who makes people breathe secondhand smoke is exposing them to a high risk of death. To put it another way, secondhand smoke makes a workplace intolerably hazardous.

Note that I'm not talking about new legislation, just responding rationally to facts. I think the tobacco industry will be killed off by the operation of existing liability laws and common sense, not new restrictions. In fact, new restrictions may give it an undesirable safe harbor ("you can't say we didn't warn you, we were displaying the government-required warning placard"). Let's hold people accountable for the known effects of what they choose to do.

A related issue is basic decency to other human beings. It may have taken this long to prove that secondhand smoke kills, but smoke has always smelled bad, and a civilized person does not force other people to endure bad smells. Unfortunately, the etiquette of about 1955-1980 was topsy-turvy — smokers were considered so superior that they claimed the right to make the rest of us breathe smoke, and we had no right to complain. I consider smoke-free indoor air to be one of the major human rights victories of the 1990s.

Permanent link to this entry

How to get previous versions of your files under Windows

One of the handiest features of Windows 7 and later is that, by default, it automatically saves previous versions of your files about once a day. (Actually, to save space, it only saves the difference between the past version and the current one.)

So if you accidentally change or delete a file, just go to the folder that contained the file you want back, right-click on it, choose Restore previous versions, and follow the instructions.

These are not backups because they won't survive failure of the disk drive. They are called "shadow copies." For more about how to use them, click here.

You still need to back up your disk to something (such as a removable disk drive) that you can store in another building. My backups are in a safe deposit box in a bank vault, and I swap them out weekly.

Permanent link to this entry

Sick of PC-versus-Macintosh harangues

[Updated.]

An afterthought that sums it all up:

The biggest misconception about Macs and PCs is that they are radically different. They aren't. There is amazingly little that you can only do on one or the other. There isn't even much that is substantially easier on one or the other. If you need software that is only available for one of them, then use whichever one it is. Otherwise, don't be needlessly partisan.

I'm sitting at a laptop that dual-boots Windows and Linux. If MacOS were available for it, I'd have that too, but it only runs on Apple machines, and right now I don't have enough need of it to justify buying another computer. But, in general, I'll use whatever tool is best for the job.

Plenty of people seem to think Windows and MacOS still differ as much as they did in 1990. The truth is far from it. It's not true that you "can't do art on a PC" or "can't do office work on a Mac." There isn't even any difference in what kinds of art or office work you can do, or what it's like to do it, as far as I can tell.

Above all, don't choose your computer based on fear. I regularly hear from people who use one type of computer and are afraid that everything else will be impossibly hard to use. That's like thinking that because you drive a Ford, you can't drive a Chevrolet.

[Original entry follows:]

I'm tired of being harangued about how Macintoshes are perfect and PCs are worthless — or the other way around. I really don't want to know whether you feel an emotional loyalty to one or the other. I am a professional, and I use whatever tool is best for the job. I know that a Windows PC and a Macintosh are much more alike than most people realize.

But I will state a few facts that may help to clear up misconceptions.

(1) Apple is a computer company but Microsoft is a software company. Windows runs on machines made by thousands of manufacturers, some much better than others. It's useful to clarify whether you're talking about the quality of the machine or the quality of the software.

(2) MacOS gains some reliability because it runs only on Apple machines. Windows sometimes malfunctions because the underlying machine does not follow specifications, rather than because of something wrong with Windows.

(3) Apple MacOS and Microsoft Windows aren't the only two choices. Linux, which is high-quality freeware, is the third option.

(4) One of Apple's most important goals is to hide technical details from the user. Some people like this, while to others, it feels like driving a car whose hood you can't open. Windows gives you more control over technical details, and Linux, more yet. (Since Apple switched to a UNIX-based MacOS, they of course expose lots of technical details through the console window; you can do everything you can do under UNIX. But most Macintosh users never see this.)

(5) Macintosh hardware is very well-made, and priced accordingly. Equally high-quality PC hardware exists but does not dominate the market. More often, PC manufacturers aim for a middle or low price point. (From here on, by "PC" I mean machines designed to run Microsoft Windows.) The lower cost of PCs is an important consideration.

(6) Both architectures have evolved over the years. In fact, "Macintosh" has been the name of several essentially unrelated computers with different CPUs but similar screen appearance.

(7) The hardware architecture of the two is nowadays the same (Intel Pentium-based). You can boot Windows on a Macintosh if you want. If the Intel Pentium architecture were worthless, Apple would not have adopted it.

(8) Historically, the PC originated as a small computer that was scaled up (an 8086 replacement for early 8080 micros). The Mac originated as a larger computer that was scaled down (based on Xerox workstations). This is why the Mac had a better graphics system at the beginning.

(9) Apple did not invent the mouse and graphical user interface; they got the idea mainly from the Xerox Star, although mice and windowing were also used on UNIX workstations from several manufacturers, especially Sun.

(10) Until the early 1990s, PCs were not as good at graphics as the Mac. That was changed by several architectural changes, followed by the advent of Windows 95. PCs have been fully capable of advanced graphics ever since — witness their popularity with gamers.

(11) Adobe Photoshop and related products have run well on PCs since the 1990s and are not limited in any way compared to the Macintosh versions. (There are still people saying Photoshop doesn't run on a PC!)

(12) It has always been easier to develop software for Windows than for MacOS. This is largely because Microsoft has done more than Apple to make software development easy. As a result, there has always been more software available for PCs than for Macs.

(13) There are considerably more viruses for Windows than for any other platform (MacOS or Linux). If you run MacOS or Linux, you may be able to get away without using virus protection. With Windows, you should run Microsoft Security Essentials (a free component for Windows) or a third-party antivirus package. Microsoft Security Essentials has little or no visible effect on performance.

(14) The difference in the virus risk is not due to architectural differences — it is just that MacOS and Linux are so much less common than Windows that virus authors don't target them much.

Actually, Windows Vista and later are quite resistant to unauthorized changes. Almost every virus infection requires the cooperation of the human victim. ("Here, install this, it will make your cursor look like Halley's Comet! Be sure to click 'Yes' when the computer asks if I can make destructive changes!") If you install software that came from nobody-knows-where and promises to do nobody-knows-what, you'll always be a ready victim for malware, no matter what your operating system is.

(15) At least where I hang out, computer viruses are lot less common than they used to be. I haven't encountered one in several years. I haven't actually had one infect my computer in perhaps as much as fifteen years. I haven't had more than about five virus infections in my life, and over that time I've been in charge of dozens of computers.

One last note. I don't quite understand why people criticize Microsoft for being a "monopoly" (versus Apple) when in fact Microsoft supports computers made by thousands of manufacturers, while MacOS computers are made only by Apple. Maybe just that Apple is the underdog?

Permanent link to this entry