resolution, part 1

(This is a four-part series on picture/image, screen, and printer resolution and how they all work together, with the hopes that it will clear up Dots Per Inch (dpi), Pixels Per inch (ppi) and why things look different on your computer monitor than they do when you print them out. Taken from a list I’m on.)

On Nov 28, 2009, at 11:24 PM, Mark wrote:


While you are at it and if you don’t mind, an ordered, complete explanation, like you so wondrously can give, of printing, viewing and other designations, especially LPI and PPI and how they all fit together in the whole picture would be a super help.

Thanks for this fascinating thread.


well…. 🙂

There are whole chapters in books written on the subject, so it’s much easier to just answer individual questions than it is to spend a day trying to compose a lesson in it!

How about an outline? (With significant technical oversimplification…)

Digital images are really long lists of numbers. You cannot hold up a file to the light and see an image. The numbers come from an interpretation of a light-sensor’s value. Those are called photo sites, and it’s an incredibly tiny bit of electronics that emits a current when it’s struck by photons (light). The more photons that hit it, the more electricity it gives off. (OK: strictly speaking this isn’t really what’s going on, (it’s more like a photon receptor/bucket) – but let’s live with this version anyway.)

The light sensitive part of your camera (what used to be film) is a flat plate of millions of these things, arranged in a rectangle. So, for example, if your camera is a 12-megapixel one, then (let’s say) the checkerboard of that plate is 4000 x 3000 (since 4000 x 3000 = 12 million).

So, you’ve got 12M little sites each giving off some amount of electricity. Notice that this doesn’t have anything whatsoever to do with colors yet… just 12M different levels of electrical current.

To get color out of that, the little photo sites each have a colored filter on top of them, either red, green or blue. (There are twice as many greens as the other two, but don’t worry about that.) They are arranged in a nice pattern (the Bayer pattern) like a checkerboard. A red filter only allows the red wavelengths thru; the blue only allows blue, and green allows green. (Why RGB? Because that’s what our eyes see, and the combination of all three make up all the other colors.)

So, the net result is that you get a current which represents the amount of red, green or blue light (depending on the filter) for any given cell.

When you click the shutter, the light exposes all 12M photosites, and the current is captured by an Analog to Digital (A/D) converter, which has the fairly simple task of assigning a number to the amount of current captured at each site. In little cameras, that number can be between (say) 0 and 1000, or in more expensive cameras it can be between 0 and 16000. (What we’re really talking about in doing that is how small a difference can be accurately recorded: do you have 1000 shades or tones or do you have 16,000 of them?)

So what you end up with then, inside the camera, is a file which is a huge table of these 12 million numbers, each one representing one photo site.

Now, one of two things happens: either that file is left alone and just sits there on the memory card, without any further interpretation. In that case, it’s in its ‘raw’ form – unchanged… or… the software built into the camera starts massaging it. It will bump up some numbers based on the other numbers that are near by, (“sharpening” the “image”); adjust the numbers based on the overall “temperature” (color temperature) and all kinds of other wonderful things you can do with a computer and numbers. (Yes: inside your camera -is- a computer.) All that is well beyond the scope of this little ditty… except this one thing: if your camera is going to give you a “jpg” file, then it also compresses that whole range of numbers (0-1000, or 0-16,000) down into 0-255.

Yes, you’re right: it throws away a huge amount of data.

And one way or the other (raw or jpg) you eventually download that list of numbers to your computer.

I’ll break there for tonight. It’s well past my bed-time, and I’ll resume some time tomorrow, and cover PPI (Pixels per inch; Lines per inch; LCD and CRT screens, and output for the web, photos and books.

Meanwhile, here’s a definition: A pixel is a picture element; the smallest “dot” of information. One of those numbers from that table of numbers we just downloaded from the camera. (Again, that’s not quite the proper definition, but it’s close enough for this discussion.)



resolution, part 2

OK… continuing (this too will have to be modest, as I’ve just discovered an issue with one of my websites that needs fixing today.)

If you were really paying attention last night, you might see something wrong with the definition I signed off with:

Meanwhile, here’s a definition: A pixel is a picture element; the smallest “dot” of information. One of those numbers from that table of numbers we just downloaded from the camera. (Again, that’s not quite the proper definition, but it’s close enough for this discussion.)

The problem? When we get that file of number loaded into a program on our computer, each “pixel” isn’t red or green or blue: we see it as orange, or yellow, or tan, or some other color, one of perhaps millions or billions of shades or hues.

In fact, we have to separate out a true “picture element” from it’s components. Remember that I said you combine red, green and blue to make a color? (To save typing, it’s RGB from now on…) But each of those RGB numbers from the camera file are individual, single numbers, from different individual photo sites.

So, what’s going on? Well, for each one of those different individual photo sites (which are either R or G or B) the computer looks at the sites around it, and decides from those surrounding sites what the likely amount of the other two colors will be. It’s called “de-mosiac.”

Whoa! What? OK: think of a checkerboard again. Each square is alternating one of three colors: RGB. Mentally impose the whole image over the whole checkerboard. You end up with an image made up of only three colors, and they are resting alongside each other. Not remotely realistic.

Now, look at just one of those sites. Say it’s blue. Next to it is a green one, and a red one, on the sides and top. What the computer (either yours or the camera’s) does is look at those surrounding sites and extrapolates how much G and how much R must have been on the B site (based on how much is actually on the sites that surround the B site.)

Yep: in a sense, it “makes up” the missing colors for each site.

So now, when you open one of those files of numbers in Photoshop, you can click on the “Channels” window, and see, or any Pixel on the screen, the Red channel, the Green channel and the Blue channel, and at the top, the combined RGB channel, in full living color (the combination of the three RGB channels).

When you get a jpg file from your camera, the calculation is done for you, so that each pixel comes ready with three numbers, each representing one of the three primary colors. If you use raw files, however, your desktop computer will do that computing, under your control. Either way, you end up with a point (one of those 3000 x 4000 points) that has three numbers attached to it.

As you can see, we’re still dealing with numbers, however. In fact each of those channels (RGB) is just a list of numbers, which your computer can -interpret- as being red or green or blue. They are just numbers, however. You can tell the computer to interpret them however you like. In fact, when one “converts” a color image to Black and White, all that’s really going on is that you’re telling Photoshop “don’t interpret the channels as colors; just interpret it as bright / dark (called “luminosity”).

What do I mean by “interpret?” Pretty simple: if you are looking at say the Blue channel, and the number in the file for some pixel happens to be 30, the computer will turn on the Blue on your monitor to a level of 30. If the number is 255 instead, it will tell the monitor to blast out that blue pixel with full brightness; if the level is Zero, it will tell the monitor to turn off the blue. Just like turning up or down a dimmer lamp. The higher the number, the brighter.

Zero is off – black. 255 is fully on (either R, G, or B).

Colors besides RGB however, are a combination of the RGB primaries at various individual levels. For example 103, 50, 117 are the RGB values for Violet; 255, 255, 0 are the numbers for Yellow (red + green = yellow.) 128, 128, 0 is also yellow, but it isn’t as bright. And any place where all three numbers are the same, is a shade of gray.

So how does the computer monitor make those colors? Just like I said: if you look with a magnifying glass at your computer monitor, you’ll see that it’s pretty much the same thing as the sensor in your camera: there are little squares (or circles or ribbons) which are either red, or green, or blue, all packed tightly together next to each other. So, if you have a photo which as a big red wall, then the pixels where it’s Red have the RGB numbers 255,0,0 (full red brightness; no green and no blue) and the monitor will go thru and turn on all the red squares, and off all the green and blue. Why does it look solid red to you when in fact 2/3 of the little squares are turned off? Because the square are so little that you eye blends them together.

And that’s why if you have a red square at 103, next to a green square at 50, next to a blue square at 117, your eye sees a violet dot.

In short, if you skip over all the discussion, then it’s really like this: a red photo site on the sensor puts out a number value based on how bright the light was that hit it. That number is then use to turn on a red square on your computer screen to the same level of brightness.

Red(camera) -> 128 -> Red (monitor)

end of part 2

back later.


resolution, part 3

Bruce asked some questions (below) that can easily be answered now, and not really as a detour, either. (The answer to part two of his question, about paper, will come separately.)

But, as to how the eye/brain reacts to dpi/ppi…

First Dots per Inch is strictly a printing term, and should not be (but often is) used when discussing computer monitors. Frankly, it’s hardly a major issue, since in casual conversation we generally know if we’re discussing a computer screen or a book… and in those kinds of practical terms, they pretty much mean the same thing. (I’t sjust kinda nice to use the correct words…)

Generally speaking, however, pixels refer to transmitted light, and dots refer to reflected light. So, you’ll have pixels with a digital camera and a scanner and a monitor, but your photo prints, books and magazines (which you can’t see in the dark because there is no light to reflect off the page) are “dots.”

First, let’s look at your monitor. It’s probably right around 100 PPI. (I’m going to limit this to LEDs for convenience.) That number is fixed. (The iPod Touch is 160 PPI, which is why that screen is so easy on the eyes despite it’s small size.)

So, what it really comes down to is the native resolution of your monitor, and it has almost nothing to do with the “resolution” of the picture. (The other thing is how far away you are from an image when you see it, but I’ll get into that later.)

OK: new word – ‘resolution.’ Strictly speaking a digital image has no resolution in and of itself. It simply has dimensions: 640 x 480; 3000 x 4000 and so on. “Resolution” is “resolving power” and a photo can take it on only in comparison to something else, such as the size of the subject of the photo, or the photo when shown on a screen or printed in a book.

Let me get the first one out of the way, er.. first. If you take a photograph of a penny with a macro lens, so that the penny fills the entire image, the a sensor that has 12,000,000 “pixels” (photo sites) will resolve more detail than a sensor with 30,000. You’ve divided the image up into many more discreet points of information. Thus the 12MP camera can be said to have greater resolution than a camera that takes a 640 x 480 (1/3MP) image.

But the photos -themselves- in either case, do not have “resolution” – just dimensions.

Now you can start talking about resolution in the second sense when you start specifying size (in much the same sense that we specified the size of the penny object).

And that’s where the PER INCH as in DPI or PPI comes in.

Let’s work with the 3000 x 4000 12MP image. That’s a fixed size. It never changes in this example.

If you decide to print that image out at 300 dpi, then the width of the printed image will be 3000/300 or 10 inches. Each inch of the resulting print will have 300 of the original 3000 pixels allotted to it.

Thus the resolution of that -print- is 300 dpi.

If you decided to print that very same file at 150 dpi, then the resulting print would be 20 inches wide; and if printed at 75 dpi, it would be 40 inches wide.

But the resolution at 75 dpi is 1/4 what it is at 300 dpi. Bigger print; lower resolution.

And now we get to viewing distance, and the human eye. Take that first 10″ print an put it on the wall 10 feet away from you. Does all that exquisite detail do you any good? Nope: it’s lost because you are too far away to see it in that little print. But, put that 75 dpi, 40″ wide print on the same wall, and suddenly you can see things you never saw before. It will look wonderfully detailed…. yet viewed up close it will look terrible, and only the 10″ one will look good held in your hands.

If you’ve ever seen one of those JumboTron stadium-sized monitors up close, you’ll find that each dot/pixel/point is about 1/2″ in size. Yep: 2 dpi. But that’s fine: no one ever looks at it from 2 feet away; 200′ is more like it.

How does that work? Hold your thumb and forefinger about an inch apart, at arms length.Put a ruler just touching them. How many 1/4″ marks can you see between your fingers? Should be four. That’s an effective “marks per inch” of four. Now move give that ruler to a friend and have them stand 10 feet away, and look at the ruler between your fingers again. How many 1/4″ marks can you see now? Should be about 48. That’s an effective “marks per inch” of 48.

(Thus the apparent resolution of an image has to do with how far away you are from it. The ruler didn’t change; the one-inch space didn’t change. To see the same resolution (that is 4 marks per inch) your ruler at 10 feet would need marks every three inches, not every 1/4 inch.)

So, when someone specifies a photo as “…your file should be 1000 x 2000 at 300 dpi” they have not got a clue what they are talking about. A digital image file is always specified by it’s dimensions, and if they specify a DPI, it -only- makes sense if they specify the “I” – the inches – as well. “5 x 7 inches @ 300 dpi” is an image that is 1500 x 2100 pixels.

So: armed with that info, let’s look at your question about why a nice, small jpg looks fine on your screen, but terrible when printed out.

A very common size on the web is 320 x 240, which on your monitor will make an image about 3 inches wide and just under 2.5 inches high. Now part of the reason that looks fine is that you’re seeing projected light: it is just like shining a flashlight into your eyes. That obscures/blurs detail. And part is, as you suggested, because your mind fills in a lot. Part can be skillful trickery (read: psychology) by the image creator (sharpening does not add any detail; it just makes you think it’s there.) (Your TV probably has a resolution of about 1/2 of that of your monitor; around 50 ppi, and that’s for HDTV!)

On the other hand, in reflected light an image printed at 100 dpi just doesn’t cut the mustard. You want something at least 150 and likely higher (288 – 360 for a photograph.) You can get that out of your 320 x 240 image, of course… just tell your printer to print it at 300 dpi. And the resulting photo will be about 1 inch wide, and 2/3 of an inch high, and will look plenty detailed… if you like postage-stamp sized photos.

So what if you love that 320 x 240 picture and want to print it out at 7 inches wide? Well, you can if you set the dpi to 45. And it will look pretty terrible… (Unless, of course you look at it from about 10 feet away…)

Or, you can tell your software to “enlarge it” but the simple fact is that no software can make up details from something that is not there in the first place. To print that 320 x 240 at 300 dpi and 7 inches wide, those 320 existing pixels have to magically become 2100 pixels. Where do they extra 1780 pixels come from? Thin air, is the answer. The software can try various tricks (some pretty sophisticated, frankly) but really it only has one original pixel for every 6.5 it’s trying to make up.

The result is an image that looks, er, odd, to say the least. (Unless, as I said, you look at it from across the room.)

In sum then, a one-word answer to your question as to why one looks good and the other doesn’t, then, is: “resolution.”

In fact, if you want to see it in action, take a 640 x 480 photo and put it up on your computer screen, and also on an iPod Touch. You’ll instantly see that the iPod image looks much better, because the screen resolution is so much higher. If you make them each the same physical size (use a ruler) the detail will be lost on your computer monitor because it can resolve only 100 ppi, while the iPod can resolve to 160 ppi.

(Which again proves that aspect of it which is, as you noted, in the venue of the mind: what you get used to seeing. Another way to see, side by side, the difference is to compare a photographic print with a reproduction in a book. The book is likely to be about 150 dpi, while the photo may be as high as 1440 or 2880. You can see the ‘dots’ in the book, but you cannot see them in the photo (with an unaided eye.)

Perhaps that will help clear up the confusion I hear in your statement “…even if it has only been reduced to say 150 or 300 dpi or ppi…” I hope you now see that you’ve not ‘reduced’ the image at all. The image is still the same size it always was, and in fact, if you want it printed at the same size you seen on the screen you will have to use some fancy software to -enlarge- the image (not reduce) so that it can print at 150 or 300 dpi. If you don’t do that, you’re just taking a tiny square pixel and blowing it up into a large square block.

Out of energy… off to spend some time with my wife.


On Nov 29, 2009, at 6:26 PM, Bruce wrote:

Your comment below reminds me to ask you, when you get to it (in a few days?), to please clarify an underlying aspect that I have never been clear about. That is the difference between print standards and screen viewing standards. (By the way, let’s continue to discuss this at the not-TOO-technical level we have been using so far here.)

I think what I am asking about is mostly a matter of human visual perception (or if you prefer, of brain integration of visual information). In short, why does an image look just fine to me on a monitor at a “display” resolution reduced down to 72 or 96 dpi or ppi or whatever? Yet even a total amateur can be dissatisfied by printing out the same image, even if it has only been reduced to say 150 or 300 dpi or ppi. Why should the paper printing process be so closely examined by our eyes and brains, while a monitor image can get away with such less information and still be pleasing?

Or am I perhaps even fooling myself with this observation?

But it seems to me that this needs to be addressed before much can be said about the needed ppi, dpi, lpi, or whatever on printing.

resolution, part 4

Part 4, and we’ll begin with Mark’s question. Mark profoundly confuses physical print size with image size. (Which is likely why he asked me to start this missive in the first place…)

As pointed out in Part 3 of this Encyclopedic treatise, a graphic image has a fixed size, say 3000 x 4000. The size on printing that image has to do with how many of those 3000 “dots” are used in a given inch.

If you print that image at 3000 dpi, the printed image will be one inch wide. If you print it at 300 dpi, the print will be 10 inches wide. If you print it at 30 dpi, the print will be 100 inches wide.

The DPI is an OUTPUT specification, and has nothing to do with the image itself.

Let’s go into Photoshop, and make a new document. When you do that, you’ll get a dialog box asking you to specify the size. Make sure the pop up menu says you’re specifying pixels (the list is pixels, inches, mm, cm etc).

Enter 320 for width and 240 for height, and as you do so, look in the lower right of the box. You’ll see the “Image size” change. That’s the size of the file. If you increase the size to (say) 600 wide, the file size will grow larger. Reduce either dimension, and the file size will grow smaller.

Now stop making changes, and look at the file size. Remember it, and go to the “resolution” field. Enter 9000. The file size does not change. Enter 72. The file size does not change.

That’s because the DPI is a specification for how the file will be printed- how many of those pixels of width will be used in a single inch when you print it out.

Now let’s do it the opposite way.

First, put 72 back into the resolution field.

Change the popup menu to inches instead of pixels. (Now you’re specifying the output size – the inches of printed width.)

Put 5 in as width (inches) and 7 as height in inches. If you have color: rgb and 8 bit in the other menus, then your image size will be 531.6K.

Briefly, switch the “inches” back to pixels. You’ll get 360 as the width because 5 x 72 = 360. Switch back to inches again.

Change the resolution to 300 ppi and watch the image size: it will balloon up to 9 Megabytes. Switch the width popup to pixels, and you’ll see that now the file is 1500 pixels wide, because 300 ppi x 5 inches = 1500.

So, at this point, I hope I’ve made clear how the dpi/ppi thing works. A given image file is some absolute, fixed size (whatever that may be, such as 3000 x 4000) and that can be expressed either simply (3000 x 4000) or as output-size/dpi (10″ x 13.3″ at 300 dpi or 5″ x 6.6″ at 600 dpi). Each of those describes the exact same file, just at different output resolutions.

If this isn’t complete clear, the rest of what follows will only be confusing, so I’d suggest you go back and re-read until it is, before continuing.

Your monitor, which is an output device, has a fixed display resolution. You can figure out what it is. First, go to the Displays preference panel and find the width in pixels of your monitor at its native setting (or look it up in the book that came with it. My 23″ Cinema Display is 1920 x 1200, for example.) Next, grab a ruler and simply measure the width of the display area (no plastic, no black, non-image parts… just the illuminated width). Finally divide the resolution from step one by the measured width. That’s your screen’s ppi. Mine is 98.14.

(Actually, if you do this, you’l likely find that the horizontal and vertical resolution are not the same! (Mine is 98.14 H and 96.0 Vertical. That’s apparently a manufacturing necessity. But I’d not worry about it: we’re talking 2/100’s of an inch here, so we’re close enough for govt. work.) I used the 98.14 width.

In photoshop, you can actually enter that amount in its preferences so that when you set an image to 100%, it will really be displaying one pixel per pixel… or darned close, eh?

To test this, set your ppi as determined above into the Photoshop/preferences/units&rulers/screen resolution.

Then make a new document, and use inches as the setting. Try 8 inches wide, with a resolution of your ppi – the same thing you just set in preferences. 98.14 in my case.

When the document appears in the window, make sure you’re viewing at 100% (see lower left corner, or just double-click on the zoom tool) and grab your ruler again. Measure the width of the document, and you see that it is exactly 8″ wide.

OK… now, let’s go back to that 12Mp photo you have; about 3000 x 4000. If you load that into PS (Photoshop) and view it at 100%, you will not be able to see everything in the image because your screen is not 3000 pixels wide and 4000 pixels tall (or larger) which you’d need to see an image that big at a 1:1 pixel ratio (100%).

To see that whole picture, you’ll probably have to view it at about 25% (750 x 1000) which will fit your (say) 1920 x 1200 screen.

What follows is a bit involved, but not really difficult, so bear with me. It’s only a long-winded explanation of the ramifications of what has gone before… with the practical result of ending up with the correct DPI to use when printing for a given target size.

First, however, there is this one important difference between printers and monitors: monitors have a fixed pixel resolution. (You can change the size, but to get a 1:1 ratio, there’s only one fixed resolution on a monitor, as determined by the number of pixels it actually has.)

A printer, on the other hand, can print at various resolutions. You can tell it to use 300 dots of ink for each inch, or you can tell it to use 10. 300 – 360 dpi is common for photo printers. (Epson printers have a resolution up to 2880 dpi, and that confuses people at first. I can go into it if you like, but basically it’s because they can print a variable size dot of ink.)

The point however is that printers can use a variable output resolution, while your file size and screen size are fixed.

Back to the monitor –

Here we go: at 100%, you’r seeing 1 pixel in the image for each one pixel on the screen. If you change the view size to 50%, you’re effectively (not -really-, but effectively) seeing 2 pixels from the image for each single pixel on the screen. (You can’t physically do that, of course, since one screen pixel is just ONE screen pixel, so you’re really viewing every other pixel from the image.)

And at 25%, you’d be looking at every 4th pixel from the image.

That is, at 25%, instead of 3000 pixels wide -on the screen- you’re seeing 750 pixels wide. And since your screen shows 98.14 pixels per inch, then on your monitor screen, that image will be (750 divided by 98.14 or) 7.642 inches wide.

And to print that image at the same exact size, 7.642 inches, you’d set the printer resolution to (3000 divided by 7.642) = 392 dpi. (Or you could work it from the other side and because you’ve reduced to 1/4th, you’d multiply by four, and you’d get the same thing : 98.14 x 4 = 392.)

Don’t hurt your brain: The 3000 divided by output-width-you-want is certainly easier to remember, and much less convoluted.

So, if you have a photo that is 3000 pixels wide, and you want it to print at 10 inches wide, you’d set the dpi to 300, since 3000 div 300 equal 10.

Again, you can easily see this in PS. Open up your image and choose Image/image size…

Notice in the resulting dialog box that PS keeps the actual image size completely separate, in its own box, at the top (Pixel dimensions), from the output, or “document” size, in a separate box. As we’ve seen, that’s because the document size is dependent on the Resolution.

Before we do anything else, however, look near the bottom of that box, for the checkbox that is next to “Resample Image:” and make sure the the box is NOT checked.

Now, to set your output size of your 3000 pixel wide document to 10 inches, we know that we have to set the Resolution to 300. Put 300 in the Resolution box and bingo, the output width changes to 10″.

Photoshop even makes this easier, and saves you the (admittedly simple) calculations. You can just put 10 inches into the Document Size: Width: and it will instantly change the dpi to 300 for you.

In fact, that’s the fastest way to approach it: but MAKE SURE that the Resample Image checkbox is NOT checked when you do this.

Then when you enter a desired size, you’ll see what the effective dpi is that the printer will use.

So, in answer to Mark and Bruce then, here’s how you can tell what will happen to an image when it’s printed out.

Try this: create an image in PS, and set the size to 320 x 240. Next, and visit the image size box, it will show you what the actual size is in pixels in the upper box – Pixel dimensions: 320 x 240.

The lower box, Document size, will show you what happens when you print it out.(But again: MAKE SURE that the Resample Image checkbox is NOT checked when you do this!)

Printers want at least 150 dpi; photo printers want about 300 dpi… so that’s what you’re looking for in the Resolution field. If your image then is 320 x 240, and you set the Document box width to 5 inches, you’ll see that the print resolution is only 64. You can pretty much bet that will look terrible.

If you set the resolution to what the printer wants (say 300) then you’ll see that the print will be just about 1 inch wide. At 150 dpi, the image will be just over 2″ wide.


That all said, you -can- check the Resample Image check box. The first thing you’ll notice is that the Pixel Dimensions box on the top becomes active. That’s because you are about to try to change the -actual dimensions- of the file! You are going to try to make something out of nothing! Remember, the actual data, as captured was only 320 x 240.

Put in 300 dpi and a size of 5 inches… and you’ll see that the pixel dimension have changed to 1500 x 1125. If you now click the OK button and go back and look at your image, you’ll see it’s all blurry… and that’s because you told PS to make up information to fill in the blanks in the change from 320 pixels wide to 1500 pixels wide.

And you’re about to print out that nice, blurry image.

And I hope that all helps to explain what is going on with images, monitors and printers.


“Open With” in Snow Leopard

If you’ve updated to 10.6 (Snow Leopard) then you’ve likely noticed that the binding between documents and the applications that created them is frequently broken.

That is, create a document in BBEdit and save it to the desktop. Now double-click that document. It may open in Safari, in Textedit, in Word, or, if you’re lucky in BBEdit.

This is intentional on Apple’s part; it is not a mistake. (In the pre-10.6 days, each document had a hidden type code and a hidden creator code, and the Finder used these to link between the document and its creator. That’s not being done anymore… and needless to say, that’s inconvenient.)

That said, all that’s left is the document’s (sometimes) hidden suffix, such as “.txt” or “.doc” or “.html” and so on.

Now you can make one program open ALL documents that end with “.txt” by highlighting the document, and pressing command-i. In the resulting GetInfo box, select the “Open with:” section; pick the application you’d like to use to open the file; and then click the “Change All…” button.

But that suffers from making -all- documents that end with that suffix open in just one program… which may or may not be the application that actually created it.

So, here’s what I did.

First, I did go ahead and set up my “default” applications, as described above.

Then, using Automator (which is the point of this article) I easily made services to open documents with a specific application.

That is: I control click on the document, and from the resulting menu, choose “Services/Preview” or “Services/Textedit” or whatever app I want to use.

For example, I generally prefer to use Acrobat to read my PDF files. However, Acrobat takes forever to print them, so if I’m printing, I want to open them in Preview (which for reasons I don’t understand, seems to print PDFs much faster.)

So, using the GetInfo technique, I’ve set PDFs to open in Acrobat.

Then, to make an item that will appear in the contextual menu when I control-click on a document, I ran Automator.

When Automator comes up, I chose to make a new Service. (If you don’t see the choices, choose “New” from the File menu, and click on the Service icon.)

Next, make sure that “Library” is highlighted in the left-most column. Then, in the next column over, find the item “Open Finder Items.”

Click and hold and drag “Open Finder Items” to the large area on the right, and it will open, revealing a popup menu, which initially says “Default Application.”

Using that popup menu, select the application you want to use. (If you can’t find it, there’s an “Other…” at the bottom of the menu list.)

That’s it. Now just choose Save from the main file menu, and give your new service a name. (I choose “ow Acrobat” or “ow TextEdit” and so on where the “ow” simply means to me “Open With”. Pick whatever is meaningful to you.)

Once you’ve named it, hit the “save” button, and you’re done. You can quit automator, or make some more “Open With…” services if you like.

To use your new Service, control-click on the document in the finder, choose “Services…” and your new service.

This whole process takes much longer to describe than it does to do, and once you’ve done one, you’ll find yourself doing many.

Automator is a very handy tool, and very easy to use. We’ll cover more of it later.

How to format a Mac Hard Drive

When you get a new blank drive, it’s not likely to be properly formatted for a Macintosh; most drives come formatted for PCs.

The Mac will run with those improperly formatted drives, but there may be issues later.

Here’s the whole story, with step-by-step instructions on how to properly erase and format a drive for the Mac.

Note that it’s different for older PPC (G4, G5) Macs than it is for the newer Intel-based Macs, and the difference is important.

On using Disk Utility (DU) :

First: the obvious – changing the formatting type will totally erase the target disk, causing you to permanently lose everything on it (unless you have a backup, of course.)

There is a difference between “drive” and “volume.” The “drive” is the physical thing; the “volume” is one of its partitions – a drive may have several partitions, just as a book can have several chapters.

So: there are drives (the physical hard disk) upon which reside one or more volumes. When you see an icon on your desktop, you’re looking at the volume. The finder does not have a way of working directly with drives. That’s what Disk Utility (DU) is for.



You’ll see something like:

465.8 GB ST3500630AS MacintoshHD

The uppermost one begins with the size of the hard disk, and its model number. This one is the drive itself.

Underneath that, and indented slightly, is usually one or more volume names. If you do not see anything underneath a disk, then it is not formatted yet (and you need to format it.)

You select a drive or a volume to work on in the larger right hand panel, by highlighting it in the left hand panel.


If you select a volume, then above the right hand panel you’ll see: First Aid / Erase / RAID / Restore


If you select a drive, then above the right hand panel you’ll see: First Aid / Erase / Partition/ RAID / Restore (note the addition of “partition”)

To select the formatting of a drive, first select a drivein the left hand column.


Next, click the “partition” button above the right hand column

Now in the right hand column, you’ll see

Volume Scheme: Volume Information:

Beneath Volume scheme, you’ll see a popup menu, probably saying “current” and beneath that, a diagrammatic representation of all the space on the drive.

Below that, you’ll see + – Options…

These are likely grayed out.



Click on the popup menu directly underneath the words “Volume Scheme”, and select the number of partitions, (volumes) you want on the drive. “1 Partition” is the most common, although you may choose more.

Once you do that (choosing anything other than “Current”) the “Options…” button at the bottom of the diagram will become active.


Click on it…


…and select the partition scheme you want for the drive: GUID Partition Table; Apple Partition Map; Master Boot Record. (If you are following along, don’t do this right now. See below first, for how to choose the correct type.)

If you are doing this for a Mac, you want GUID as the partition scheme, and you want “Mac OS Extended (journaled) for the format (in the long rectangle, below.)

Then Click OK.

Screen Shot 2014 05 10 at 2 12 22 PM


Click “Apply” to confirm that you want to go ahead, and the drive will be reformatted in the selection method.


So, which type of partition scheme do you want? Read on:

Here’s a bit about GUID (aka GPT) vs APM.

[GPT = GUID Partition Table = Globally Unique IDentifier Partition Table] [APM = Apple Partition Map]

When you buy a bare drive, it’s formatted as… none of the above: it’s MBR – Master Boot Record – which is what PC’s use, and why you need to partition it with one of the others. (MBR is recognized by the Mac OS, but its structure imposes limits on what can be stored on the disk.)

You need APM to boot a PPC Mac, or use on a PPC Mac that isn’t running at least 10.4.8.

If you’re on an Intel mac, there is no reason at all to use APM (unless, of course, you’re planning on using the drive on a machine that meets the above criteria.)

And… if you’re on a PPC machine, and are running 10.4.8 or later, then the only drive you need partitioned as APM is the boot drive (&, of course, any drives you’re cloning to with the intent of doing a drop in replacement.)

The rest of your drives can be GUID, which offers some advantages.

Entirely anecdotally, they seem a smidgen faster to me… although it seems it’s somewhere in the 8-10% range (and I could simply be delirious, too.)

That said, they employ checksums on the partition map and header, as well as duplicates of each of those and the partition maps are larger. There are other tidbits which, like those I just mentioned, are not anything you’d ever notice in use… but make geeks like me really happy.

(OK… it makes your drive a bit more resistant to some kinds of corruption.)


Intel: GUID all the way. PPC: OS less than 10.4.8 – APM all the way PPC: OS 10.4.8 > : APM for boot; GUID for everything else.

Finally: “Install Mac OS 9 Disk Driver” – check this only if you are going to use the drive with a machine that has been booted from OS 9. If you are only going to use the drive with OSX, you can leave it unchecked.

You should leave “journaling” checked for Macintosh OSX use.

Are You a Power-User of Apple’s Mail Program?

If so, then you should take a look here:

One of the most frustrating issues for those of us who “live in Mail” is that its “rules” (I prefer “filters”) are so minimal. In fact, most email software these days suffers from that condition.

Filters can be immensely helpful, but Mail doesn’t even have a way to filter outgoing email, much less anything remotely sophisticated.

Well, that’s solved with the add-ons from indev. Now, as usual, I only add my software thoughts about software I actually use, and that’s the case here as well.

Let me get my one less-than-positive thought out of the way first: the three plug-ins are too expensive. They should be sold as a bundle for $49 (US.) Am I griping because I paid full tariff? Not really. I’ve just been in the software game (as a programmer and a marketeer) for 30+ years, and that’s my opinion.

Now, what are they? In short, three plugins: Act-On, MailTags, and MiniMail.

MiniMail is a convenience: it’s a small window you can leave up and running that will show you incoming mail. Unobtrusive is its main quality, and I’ve grown accustomed to a quick glance to see if the mail is urgent, or something I can put off until later. Nice, but hardly a “big deal.”

What is a big deal, however, are the other two… and they work together so well that they really should be purchased as a pair.

MailTags is just what you’d expect from the name: you can easily add tags to mail, either incoming or outgoing. Identify each mail as belonging to a certain project; tagged with one or more keywords; add notes (!) or colors; add to iCal projects; create To_do’s and more.

Mail Act-On are the enhanced filters. Almost worth the price alone IMHO is the addition of outgoing filters. When I reply to my photo agent, my reply goes to the appropriate business folder, along with his emails in to me.

Further, greatly enhancing this is that the MailTags are available as filtering criteria as well.

Finally, the Act-On plug-in has a “public face” – a control panel that will let you do many things “by hand” but very quickly. For example, with one or more emails highlighted, I can just tap-tap (hit two keys, which takes, what? 1/2 second?) and move then directly into a sub-mailbox. Or I could add a keyword or set the project.

If you’ve never experienced the power of “real” mail filters (rules) then you’re in for a surprise. (However, I should note that using filters requires a logical mind-set, and some getting used to the logic, especially if you’re a “filter newbie.”)

OTOH, if you’ve used (and missed) strong filtering before, this software is a real no-brainer.

You should visit the site, of course, for details, and to discover the features I didn’t mention here.

For the right audience, this software is highly recommended.

Renewing MobileMe

Apple will charge you $99 to renew your MobileMe account. However, you can shop around (Amazon, JR, etc) and find it for as little as $57. It looks like a new signup (“MobileMe Individual” or “MobileMe Family”) but what’s really in the box is a slip of paper with a number on it.

All you need to do to renew (and you can do it up to a year in advance) is visit

Click the button that says “Renew an existing account” and enter the number from the slip of paper in the box.


To verify it or to just see if you have a renewal pending, just sign into your MobileMe account and click on your account settings.

You’ll see something like:

Your subscription will renew on (or expire on) (date)

Then click the options button.
On the resulting Account Options page, if you have a renewal already entered, you’ll see:
MobileMe Subscription renew:(checkbox) price: $99
Activation Key -$99

What that says is that you have a renewal pending (that’s the (minus sign) $99) which will kick in on (date), and you’ll be good for another year.

Why it’s called Micro$oft

Today I updated VMware Fusion to 3.0 on my MacBook. My installation of Vista Ultimate didn’t like that, and promptly announced that this copy wasn’t “genuine” and likely pirated.

Now, I’ve owned, and paid for:
Windows 95
Windows 98
Windows ME
Windows XP
Windows XP Professional
Windows 2000 Professional
Windows Vista Ultimate.

A small fortune in generally crappy software (which I needed to have since I wrote cross-platform software, and needed to test it.)

I have each and every original disk, and each Product Key.

So I wasn’t particularly worried that it complained, as there is a built-in mechanism for verifying and activating… except this time, it didn’t work.

Verification told me to put in my Product Key (instead of verifying) and then told me that the product key was for an update, and couldn’t be used.
Activation told me to put in my Product Key (instead of verifying) and then told me that the product key was for an update, and couldn’t be used.
Choosing enter product key, told me that the product key was for an update, and couldn’t be used.
Visiting the website for activation… (follow me on this) gave me explicit instructions which (you’re ready?) to enter the product key, and that the product key was for an update, and couldn’t be used.

At no point did I get a phone number, or the “activation key” sequence I’d seen before.

You see, apparently with Vista, M$ has decided not to tell you the customer activation number, and instead offers to sell you a new key.

Well, I found the phone number via Google, (no, not Bing) and called.

“Dave” (with a strong Indian accent) kept interrupting me, to tell me things I’d already done, until I finally sat silent for 20 seconds and he wondered if I was still there on the line. This time I asked if he’d care to hear the full description of the problem. When he finally listened, he told me that he was in product activation, and that sounded like a technical problem to him. Would I mind if he transferred me?

Nope, not at all. Thank you. (I”m always overwhelmingly polite with these poor folks, who spend most of the day talking to irate computer-illiterates, as I’ve discovered that always works best.)

Off to “Jim” (with a strong Indian accent) to repeat the entire story again… and be told that was a question for the activation center, and that he couldn’t help. Would I mind if he transferred me?

Nope, not at all. Thank you. (Figuring that maybe I would not get “Dave” again.) I didn’t. I got “Tom” (with a strong Indian accent) and I repeated the problem yet again.

Well, obviously that was a problem for customer service, not activation! (Sure! Silly me!) Would I mind if he transferred me?

Nope, not at all. Thank you.

This time I got “Bill” (with a strong Indian accent, although I thought by now that surely I’d reached Gates himself… but the accent was wrong…) where I explained the entire situation for the 4th time.

And was told “it’s too bad that you’ve had the software longer than 90 days, so your free support has run out. You can visit our forums; buy a new activation key; or continue talking to me for an incident fee of $59.”

As I hung up, I’ll admit my thoughts were less than cordial.

And that, my friends, is why it’s spelled “Micro$oft.”


PS. I finally did get it installed, but the way I did it was never even suggested by M$. It was to simply re-install XP and then update it to Vista. I had a backup copy of my Fusion XP file, and ran that, and installed the upgrade from there. On a Macintosh, the total upgrade time, from scratch, with all online updates: 90 minutes.

The same for a PC: 6.5 hours. Why? Because on a Mac, there is a cumulative update package, that will take whatever your current software version number is, and update it. One download. One update. PCs on the other hand have to go thru every single upgrade individually, in the order it was released, one at a time. There were 95 individual updates to be downloaded and installed. Total time for just the updates (on fast cable) was 5 hours (+ 1.5 hours for the software install in the first place.)

No. Really. I love my PC. I mean what’s not to like about the Registry?

Wait. Gotta stop. My tongue is bleeding.

How to make a password

Issues with forgotten passwords come up all the time.

First: know that generally speaking, if you forgot your password, you’re screwed.

Second: do not choose a word for a password. For example: Ford, Billy, password, Iloveyou and secret are all passwords that can be cracked (guessed) in about 1/1000 of a second.

Billy1234 is also terrible.

No password at all is the first thing tried by the bad guys.

password is number 2.

pwd is number 3.

Bad guys use dictionaries of words and passwords to simply look up your password.

Here is a very strong password: *t,hRj9GYwpg9ny

Wonderful. How are you supposed to remember that?


Here is an equally strong password: Our1Family2Black3Cat4Is5Named6Doofus


Yes, it can be as simple as using a phrase you’ll never forget, and a simple rule or two you’ll always remember.

In this case, the phrase is “our family black cat is named doofus”

The rules: 1) each word is capitalized; 2) spaces are replaced with numbers, starting at number 1.

That’s just a sample.

Start with a phrase. Capitalize part of it. First letter? Last letter? 2nd letter? you decide.
replace the spaces with numbers and/or symbols (ie instead of 1,2,3,4 use !@#$ (which is shift 1,2,3,4)

What if the place won’t let you use a password that is 36 characters long? (length is good, by the way).

Then make a new rule: O1f@B3c$I5n^D is rated just as high as the others listed above, but is only 13 characters long. (This one is based on the phrase and suggestions above and the rules are rememberable. I’ll leave it to you to do the detective work and figure out the rules applied…)