In computer programming and data structure design, a "B-tree" is a "binary tree." It is called this because each node can have at most two nodes connected to it.

B-trees are data structures used in many, many places in computer programming, not just disk directories. A type of parser used in writing compilers called a "recursive descent parser" takes a computer program written in a high-level language, generates a "parse tree" that is a B-tree, and then emits object code as it walks the tree. A program I once wrote that plays the game Othello makes a B-tree of all the possible moves, then walks the tree looking for the best move. Compression programs that use a type of compression called Huffman encoding work by generating a B-tree of the data to be compressed. Some sophisticated sort algorithms construct a B-tree of the information to be sorted.

It is important to understand that the Attribute B-Tree in a Mac disk directory is not the same as the catalog B-tree.

The catalog B-tree is a binary tree that keeps track of all the files and folders on your disk. It also makes use of another data structure, the Extents B-tree, which is a binary tree that keeps track of files that are broken up in multiple pieces on a disk. f either of these things is damaged or corrupt, you're in trouble--you may lose track of files and folders on the disk.

The Attribute B-Tree is new in HFS+ (Mac OS Extended), and is not really used quite yet, except (I believe) in a very limited way by Tiger's Spotlight. The Attribute B-Tree is not used to store information about where the files are located on a disk. It is, in theory, used to store "metadata." In a future version of Mac OS, Apple has talked about allowing you to tag a file with any sort of descriptions, comments, or keywords that you want. For example, you could write a letter to your mother in Microsoft Word,and you could save it as "letter-to-mom.doc" with the tags "family," "mother," "personal letters," and "recipes for pie." You could then, for example, use Spotlight to say something like "Find a list of all files on my hard drive that are personal letters" or "Find a list of all files on my hard drive that have something to do with my family." Spotlight would search the attributes B-tree looking for those attributes, and find the files.

Right now, Spotlight can not do this. It can search for words inside a file, but if the letter you wrote to your mom does not contain the word "family," then Spotlight won't find it. With attributes, you can tag files with names or keywords or any other kind of association and Spotlight will find it. You could even, for example, make a Spotlight "smart folder" with the attribute "family," and whenever you open that folder, all your files that have something to do with your family would be inside it.

But that has not been implemented yet. The Attributes B-tree is not really used for anything of note yet. I don't know if Leopard will make use of file attributes or not.

The simplest, most straightforward difference: The Core Duo chip is a 32-bit processor, like the PowerPC G4. The Core 2 Duo chip is a 64-bit processor, like the PowerPC G5.

Each "Duo" chip means that there are actually two separate processors built into the same package. A computer with a Core Duo or a Core 2 Duo is a multiprocessor computer with two processors. The Mac Pro systems which have two Core 2 Duo Xeon chips, have a total of 4 processors.

The "2" means second version. The Intel Core was the first processor built around this chip design; the Core 2 is the second, just like the Intel Pentium was the first processor built around that design, and the Pentium 2 was the second.

Core 2 processors are faster and more efficient than Core processors, can process more data at the same clock speed, and are 64-bit processors rather than 32-bit processors.

If you are considering a purchase and can afford to go either way, go with Core 2 rather than Core. Buying a computer with an Intel Core processor is like buying an original Pentium after the Pentium 2 came out; you're getting an older processor design that is not as fast.

Resolution is how big each pixel is, nothing more. If an image is 100 pixels per inch, each pixel is a square that is 1/100th of an inch wide. If a picture is 300 pixels per inch, each pixel is a square that is 1/300th of an inch wide.

And absolutely, positively none of that makes any difference whatsoever if you are displaying a picture in a Web browser.

For the purpose of a Web browser, forget about pixels per inch. It does not matter. A Web browser discards that information when it displays a picture. A Web browser does not even bother to read the picture's resolution.

Instead, a Web browser always displays a picture at "one screen pixel equals one image pixel." A Web browser throws away the resolution, and then sets the resolution to be whatever the resolution your screen is.

This is a difficult concept for many people to grasp, but the only thing--absolutely, positively the only thing--that matters to a Web browser is the total number of pixels.

When you are explaining this to someone the only thing that matters is that the picture is 1288x966 pixels. The 230 ppi does not matter. The result would be bit for bit, pixel for pixel identical if it was 72 ppi or 300 ppi or 8,000,000,000 ppi. The ONLY thing that matters to a Web browser is the total number of pixels.

Let's say you have three versions of the same picture. One is 300x200 pixels at 72 pixels per inch. One is 300x200 pixels at 200 pixels per inch. One is 300x200 pixels at 1,200 pixels per inch. A Web browser will display all three pictures absolutely identically--same size, same everything.

If a picture is too large, tell them "Your picture is 1288x966 pixels. Please reduce it in size so that it is no more than 500 pixels wide. You do this by using the Image Size command in your photo editing program.".

Nothing--no technique, no algorithm, and no program--can take a small image and blow it up without degrading quality. It is not even theoretically possible. There is no way to add detail to an image that does not exist in the original.

Genuine Fractals produces results that are moderately better than Photoshop's built-in interpolation for some images--emphasis on "moderately" and "some images." On some images, the results produced by Genuine Fractals are worse.

Of course, you won't see that in Genuine Fractals advertising. Part of advertising a piece of software is in carefully choosing data that makes your software look best. I used to demo software at trade shows, and we would be very, very careful about the samples we used in the demos. I have known cases were people who are giving software demos will pay their own employees to act like audience members and to ask questions designed to give the person running the demo a chance to show off some particular, obscure data set or procedure that really makes the program look good.

Some images benefit from Genuine Fractals; some don't. In any event, if you blow up an image using Photoshop's bicubic interpolation, you won't see pixels either--the image simply looks soft and slightly out of focus, just as it does with Genuine Fractals. Nothing, however, will ever look as good as simply making the image at the right resolution in the first place!

I became disenchanted with Genuine Fractals, and believed that the company is unethical and practices snake-oil salesmanship, when I saw one of their brochures. The brochure had a lot of pairs of pictures. In each pair, one picture was scanned at high resolution, and one was scanned at low resolution and blown up using Genuine Fractals. The brochure challenged viewers to guess which one was high res and which one was enlarged.

I guessed correctly on every image. However, every image enlarged with Genuine Fractals looked better. Why? Because the images scanned at high resolution were not sharpened. The images enlarged with Genuine Fractals were blown up and then sharpened. Everyone who works prepress knows that any picture printed on a press must be sharpened. The company was trying to skew the results. If they would have sharpened the high-res scans the way they sharpened the blown-up pictures, the high-res scans would have looked better.

People keep looking for a quick fix. There is none. if you want a high-quality image, you must start with a high-quality, high-resolution original. There is no shortcut.”

Have you ever seen a tile mosaic? The kind where a picture is formed by using little ceramic tiles of different colors?

That's what a scanned picture is. It is just a tile mosaic. It really is that simple. Each tile is called a "pixel."

Let's say I walk up to you and say, "I am making a tile mosaic. I want it to be three feet wide. How many tiles do I need?"

You can't answer me. You don't know how big the tiles are. If each tile is two inches across, I need eighteen tiles across. If each tile is one inch, my mosaic is thirty-six inches across.

That's what "resolution" is. "Resolution" means "how big the pixels are." A 300 pixel per inch image means the pixels are 1/300th of an inch across. A 72 pixel per inch image means each pixel is 1/72 of an inch apart.

Now, for the tricky part:

THIS ONLY MATTERS FOR PRINTING A PICTURE ON A PRINTER.

When you display a picture on a computer monitor, the computer monitor is divided into pixels already. If your computer monitor is set to 800x600, that means your computer monitor is a grid of tiny squares, 800 across and 600 down. If your computer monitor is set to 1024x768, that means your computer monitor is a grid of squares, 1,024 across and 768 down.

When you look at a scan on your computer monitor, the size of the pixels does not matter. Your computer monitor is already divided into a grid of pixels.

If your computer monitor is 800 by 600 and your picture is 800 pixels across and 600 pixels down, the picture will exactly fit on your computer screen. If your picture is 1,600 pixels across and 1,200 pixels deep, it will be twice as big as your computer screen in each dimension.

THE COMPUTER SCREEN DOES NOT CARE HOW BIG THE PIXELS ARE. It displays the picture so that one image pixel is one computer screen pixel.

If your screen is set to 800 by 600, a picture that is 800 pixels by 600 pixels will exactly fill your computer screen. Always and forever. If it is 800 pixels by 600 pixels at 72 pixels per inch, it will exactly fill your computer screen. If it is 800 pixels by 600 pixels at 1,000,000,000 pixels per inch, it will exactly fill your computer screen.

Your computer screen does not know how many pixels per inch the picture is. it ONLY cares what the TOTAL NUMBER of pixels is. The pixels per inch ONLY matters when you PRINT the picture.

When you make a scan, how many pixels do you end up with? Well, that requires a little math.

Let us say you scan something that is one inch by one inch. How many pixels are there?

If you scan it at 72 pixels per inch, it is 72 pixels by 72 pixels. If you scan it at 300 pixels per inch, then it is 300 pixels by 300 pixels. If you scan it at 600 pixels per inch, the image will be 600 pixels by 600 pixels. If you scan it at 2400 pixels per inch, the image will be 2400 pixels by 2400 pixels.

So, when you look at the picture on your computer screen, the more pixels per inch you scanned it at, the more pixels you have; the more pixels you have, the bigger it is on your computer screen. Why? because the computer screen only cares about the TOTAL NUMBER of pixels. The bigger the number of pixels, the bigger it is on your screen.

The "kernel" of a Unix-based operating system is the core part of the operating system. it can not draw windows on the screen or accept commands from a user or run programs or do any high-level tasks; instead, the kernel is what loads when the computer first starts up, and the kernel does low level stuff like talking to hard disk drives and other devices, allocating memory, assigning CPU time to programs, and that sort of thing.

A "kernel panic" is a system error that has screwed things up on such a low level in such a fundamental way that most of the operating system--the parts that do things like draw windows on the screen--can no longer work. This is an error that is so great, it's destroyed the ability of the operating system to tell you what the error is.

In older versions of Mac OS, the computer would just freeze if this happened. The mouse pointer would suddenly stop, and the computer would just sit there. In Windows computers, this kind of error brings up the "Blue Screen of Death" – a blue screen with some numbers in the middle of it that describe, on a low level, what the processor was doing when things "blew up.”

Viruses are not created by antivirus companies, if that's what you mean. There's no way an antivirus company would risk it--they're not going to risk having a large corporation like Symantec go bankrupt, and having all the executives sent to prison. Besides, there's no need. There are zero Mac viruses, yet people still buy Mac antivirus software!

Computer viruses are not written by random vandals, antivirus companies, or bored kids; they are written by dedicated, organized professionals, often associated with Russian or Eastern European organized crime, for the purpose of profit.

Almost all modern computer viruses are written to take control of the infected computer. Lists of infected computers are then sold to spammers. The spammers seize control of the infected computers and pump spam through them.

These viruses do more than just let the virus writers send spam through them. They also sift through the computer looking for things like bank account information, eBay information, and so on, which is sent back to the virus writer.

Computer viruses have become one of the top moneymakers for Russian organized crime. One person in particular, Leo Kuvayev (head of the so-called "Russian spam gang") works closely with Russian mafia writing computer viruses which infect computers and install mail servers on them. Leo then sends large amounts of spam through these infected servers, advertising everything from fake "penis pills" to so-called "OEM" pirated software. He also sells lists of infected computers to other spammers, and to people who make money via extortion, using networks of virus-infected computers to attack popular Web sites and then demand payment from the owners of those sites.

In some cases, viruses install compact Web server software on the infected computers without the owner's knowledge. When you receive those fake emails trying to con you out of your PayPal password or your bank account number by telling you your account is suspended and sending you to a Web site that looks just like PayPal or a bank, odds are good the Web site you are seeing is hosted on a virus-infected computer.

With so much money inn computer viruses--it is now more profitable for Eastern European organized crime than more traditional ventures like drugs and prostitution--there's no shortage of computer viruses. The antivirus companies don't make the viruses; they don't have to.

"DS Store" stands for "Desktop Store." The .DS_Store file is created and used only by OS X. It serves many of the same functions as the Desktop file did in OS 9--it stores window views, icon positions, and so on.

It is not essential. If you delete this file, you won't harm your system. All that will happen is the next time you open that folder, the folder view will be set to the default view, and the icons will be set to default positions. When you change folder views or move any icons, the .DS_Store file will reappear.

Normally, each folder accessed or used by OS X will contain one .DS_Store file. If you copy or move a folder, its .DS_Store file will go with it; that means if you FTP a folder or stuff a folder, its .DS_Store file will be FTPd or stuffed along with the rest of the folder.

If FTP transfers "time out" when you are doing long file downloads, this indicates that the FTP server was set up incorrectly on the Web host. This is a problem you might want to notify them of.

It's a very common configuration mistake. FTP is a weird protocol; when you use FTP to connect to a server, your computer makes two connections, not one. One connection is called the "data port" and the other is called the "control port." Commands, such as "log in" or "display a list of files" or "give me this file," are sent through the command port, then the actual file is sent through the data port.

Often, when people set up an FTP server, they want to set it up so that if nothing happens for a while, the connection times out and you are disconnected. This is to keep people from accidentally remaining connected all day.

But the configuration mistake they make is that they only check for activity on the "control port." If there is no activity on the control port for a long time, they disconnect.

But when you are uploading or downloading a file, all the information is traveling over the data port, not the control port.

This is such a common mistake that so many ISPs and Web hosts make that Fetch actually has a preference for it; by turning on the option to tickle the server during a long file transfer, you're telling Fetch to occasionally send nonsense commands (actually, commands that do not do anything) over the control port in order to keep the server from timing out. But the real fix is for your ISP to fix the FTP configuration mistake.

For starters, never send a JPEG as an ad. It's an inappropriate format to send ads.

Here's what happened:

You created a JPEG at some resolution. You sent it to someone who looked at it in a program like a Web browser or an email program.

Web browsers and email programs do not pay any attention at all to a JPEG's resolution. All JPEGs are considered to be the same resolution as the monitor. On printout, all JPEGs are considered to be 72 pixels per inch.

Say I send you a JPEG that is 3 inches at 300 pixels per inch. You open it in your Web browser and print it. When you open it in your Web browser, it fills most of your screen, because it is 900 pixels wide. When you print it out, it prints 12.5 inches wide, not 3 inches wide, because the browser thinks it is 72 pixels per inch, not 300 pixels per inch.

Needs updating...

"Bad blocks" and "bad boot blocks" are not the same thing, and are not even remotely similar or related in any way, even though it sounds like they might be the same thing.

A "bad block" is caused by physical damage to the platter inside of a hard drive. It is a physical flaw in the surface of a disk. You can't write information onto it.

"Bad boot blocks" means that the information written in the outer edge of the disk is not recognized by Disk Warrior. When you turn on your computer, it reads the information written in the very first part of the disk drive, and assumes that this information is a computer program. the computer executes whatever it sees.

This tiny computer program is the "boot code." It instructs the computer how to load and run the operating system. The place on the disk where the boot code is stored is the "boot block." When your computer starts up, it loads the information into memory, then runs it, and by running it, it loads and runs the operating system. DiskWarrior looks at the information stored in the boot blocks and compares it to the information on a disk when it is shipped from Apple.

Changes to the operating system may change the boot code. Let's say Apple releases an operating system, then Alsoft releases DiskWarrior. then let's say Apple changes the operating system, and changes the boot code. Software update tells you a new version of the operating system is available. You download it and install it. It changes the information in the boot blocks. DiskWarrior doesn't know that Apple has changed the boot code. DiskWarrior thinks the information in the boot blocks is wrong. It's a harmless "error." Ignore it.

Most consumer DSL is "aDSL"--asynchronous DSL. It gives you high speed downloads--files go from the Internet to your computer quickly--but low-speed uploads. When your computer sends information to the Internet on aDSL, it's no faster than an old-fashioned dialup modem.

If you ever send (by email or Web or FTP or any other means) large files, aDSL is not for you.

There is a form of DSL--sDSL ( "synchronous DSL")--that is fast in both directions. It is usually quite expensive--as much as double the cost of aDSL, or more. If you have to send files, cable is a better choice. Cable modem connections usually do not have really slow uploads.

MIME encoding is not a form of stuffing. All email programs, with no exceptions, always encode all email attachments, using typically one of three encoding methods: MIME (used by 95% of all email programs, including AOL, Outlook, Entourage, Mail.app, and so on), UUencode, or BinHex.

AOL *also* compresses attachments if you attach more than one file, or if you attach a folder. It's based on the number of files, not on the size of the attachment. One file: MIME encoding. Multiple files, or one folder: MIME encoding plus compression.

Now, here is where it gets complicated: Mac files have two "parts," called a "resource fork" and a "data fork." Other computers do not use or understand two-part files.

Apple has invented a type of encoding called AppleDouble to handle two-part files. The file is broken into TWO files: One part of the data fork, one part for the resource fork plus special Mac information (icon, Type code, Creator code, and so forth). Each of those two parts is encoded using MIME. Then both parts are emailed.

Some mail programs do not understand AppleDouble. The person who receives the email sees TWO attachments, not one. One attachment is the file. The other attachment is the resource fork and Mac information. The first attachment is useful, the second attachment can not be used on a PC and should be ignored.

AOL gets very confused by AppleDouble emails. AOL will not show you the attachment. Instead, if you send an AppleDouble attachment to an AOL user, the AOL user will see this: "This email has an attachment encoded using MIME. AOL can not decode the attachment. To decode the attachment, you will need to download the email and then use a decoder program. For more information, visit Keyword: MIME."

The message means exactly what it says. The attachment can't be decoded by AOL. In order to use it, the AOL user will need to download the email and then run a MIME decoder program. (I believe the newer versions of Stuffit Expander can decode MIME.) Unfortunately, there is no way to turn off AppleDouble in Mail.app.

There are two common causes for this problem.

Cause #1: You are not obeying the PC rules when you name the picture. There are specific and very rigid rules you must follow when you name a file being transferred to a PC; the slightest deviation means the file will not work. The file name must contain no "funny" characters like / or ? or \ or *, and it must end in a period and a special three-letter code that describes the file. The code for a JPEG is ".jpg".

Cause #2: The person receiving the file is not decoding it right. All files must always be encoded in order to be sent through email. Many Mac programs, such as Mail.app, use a type of encoding called "appledouble." Appledouble breaks the file into TWO pieces--one that contains the actual data, and one that contains information that only a Mac can use--information like the icon and the Type/Creator codes.

PC programs often can not read AppleDouble files, which means that the PC user sees TWO attachments. They both have the same name. One is tiny, and can not be opened--only a Mac can use it. The other is bigger, and contains the actual file.

"Disc at once" does not refer to making preparations to burn a disc; nor does it refer to copying a disc.

There are three ways to write information onto a CD-R or CD-RW; they are called "disc at once," "track at once," and "packet writing."

In "disc at once" writing, all information, including gaps between tracks (gaps between songs in a music CD, gaps between the directory and the disc data, separate tracks on a multisession or hybrid CD) is recorded all at once. The CD burner never stops burning and the laser never turns off, even when it is writing a blank space.

In "track at once" recording, the laser is shut down briefly between tracks. If you are writing a music CD, for example, the computer sends the first song, then after the first song is written, the laser shuts down, the computer stops burning and pauses for a moment, then begins sending the next track. The laser starts up again and the next track starts to burn. The same thing happens between the disc directory and the actual data.

The first CD recorders on the consumer market only did track at once, and could not do disc at once. There are two disadvantages to track at once writing. First, the laser flickers briefly when it shuts down, recording 2 or 3 sectors of garbage noise between tracks. This makes no difference to an audio CD player or to a computer CD-ROM, but it means that the CD can not be used by a professional duplication machine that makes factory-pressed commercial CDs; the machine which creates the "glass master" can not read a CD written track at once. Second, there is a gap of 2 seconds between songs on a music CD; you can not have two songs that blend right into each other with track at once writing. To make two songs blend straight into each other, you must use disc at once writing.

The third way to write a CD is called "packet writing." It means that data can be written to the CD in small chunks. Programs that pretend a CD is a floppy--programs, like those for the PC which let you add files to a CD, "delete" files from a CD (you can't really delete them, they still take up space on the disc, but the directory is rewritten without the file), and so on, use packet writing. CDs written this way are often created in an oddball nonstandard format that can not be read on any computer without special software that understands the proprietary format

Here is what happened:

You bought the drive. It was formatted for DOS, not Macs. You did not reformat it.

You plugged it into an OS 9 computer and copied files onto it.

Now, a Mac file can not be put on a Windows-formatted disk. Mac files have two "parts," called a "data fork" and a "resource fork," and they also have extra information (the Type and Creator codes, icon, and other information) called "Finder metadata." The PC disk format is incredibly primitive compared to the Mac disk format; there is no way for a PC disk to store a two-part file, and no way for a PC disk to store Finder metadata.

So when you put a Mac file on a PC disk, the Mac has to work some magic. It breaks the file into TWO files, one for the data fork and one for the resource fork, and then puts the Finder metadata into the resource fork part. it then saves two separate files on the disk. You do not see the two separate files, because the Mac automatically reassembles them on the fly when you use the file.

OS 9 and OS X use two different, incompatible ways of breaking the file into two separate files and putting the two separate files back together. You can not use a PC-formatted disk to move files between OS 9 and OS X.

"Safe boot" is a way to start Mac OS X without loading any device drivers except the ones written by Apple. Starting in safe boot does not fix anything in versions prior to 10.4. In Tiger, Apple made Safe Boot automatically delete caches and run fsck (file system check), though you can't see it happening. Generally speaking, starting in safe boot will stop every non-Apple gadget or device attached to your computer from working.

You only start in safe boot as a way to try to figure out why a computer will not start up. If your computer will start in safe boot but will not start in normal boot, you have found your problem: you have a device driver or system extension that was not made by Apple installed on your computer, and it is making your computer crash. From that point, you can uninstall your non-Apple device drivers or system extensions until you find the one causing the problem.

The "terminal window" you are seeing is not a terminal window. It is Apple's Open Firmware; if you read all of the words on your screen, you will see "Welcome to Open Firmware" in there somewhere.

Starting in Safe Boot will not help because Open Firmware is not Mac OS X. If you see an Open Firmware prompt, your computer has not loaded Mac OS X yet--in fact, it has not booted at all yet. So making changes to Mac OS X will not solve the problem

You can upload a Terragen file just fine. In order for you to understand what is happening, it is very important for you to understand that a Terragen file IS NOT AN IMAGE. A Terragen file is text. All Terragen files are text. What you see in your browser is the true contents of the file.

Terragen is a landscape program that lets you make pictures of scenery. When you save a .tgd file, you are not saving a picture of scenery. You are saving a series of commands to Terragen.

Think of it this way. Suppose I told you: "Take out a red pencil. Draw a circle one inch in diameter that is half an inch from the left and two inches from the top of the page. Take out a blue pencil. Fill the circle in blue."

What I just said is not a picture. It is text; my words are not a picture of a red circle filled in blue.

That is what a Terragen file is. It is a text file that contains a list of commands to the computer program Terragen. Terragen follows those commands, and draws a picture. But the TGA file is not a picture; it is a list of commands to follow in order to make a picture.

What do you want to do? Do you want to put up a picture you created in Terragen so that other people can look at it? If so, you need to "render" the picture--that is, you need to tell Terragen, "Open this list of commands, follow the commands, and then save the picture you made by following the commands." When you render a file, it will make a picture, which you can open in Photoshop or some other graphics program and save as a JPEG for placing on the Web.

When you use an SMTP server, you have a choice; you can use authentication, or not use authentication.

Generally speaking, you can set up a zillion email addresses on a zillion different ISPs but only ever use one SMTP server, because the SMTP server has nothing to do with your email address or your incoming email. When you use the SMTP server of whatever ISP you are using to connect to the Internet right this instant, you do not need to authenticate; for example, if you are connected to the Internet right now using Bob nd Fred's Easy Internet Service, you can use Bob and Fred's Easy Internet Service SMTP server with no authentication. It is only if you use an SMTP server belonging to someone other than who you are using to connect to the Internet right now that you must authenticate.

In a hypothetical case, then:

You are connected using Earthlink. You do not need to authenticate to use Earthlink's SMTP server, but you do need to authenticate using GoDaddy's or Comcast's or RoadRunner's or any other SMTP server.

Now you take your computer to your friend's house. He uses Internet Junction. Even though you are not an Internet Junction customer and have no Internet Junction email accounts, while you are there you can use the Internet Junction SMTP server without authentication. However, you must authenticate if you want to use Earthlink's, or GoDaddy's, or Comcast's, or whoever.

Now you and your friend go to the local coffee shop. They use Knology. While you are there, you can use the Knology SMTP server without authentication. However, your friend must authenticate in order to use Internet Junction, and you must authenticate in order to use Earthlink or Comcast or whoever.

Outgoing SMTP servers are always open-access if you are connected directly to the network of whoever owns the SMTP server.

A .dmg file is generally compressed and is usually not stuffed or zipped. If it is zipped on a PC (eg, if a person downloads a .dmg onto a PC, then zips it on the PC), it will be destroyed.

The .zip format does not recognize and does not store resource forks. Mac programs that create .zip files can do all kinds of tricks (MacBinary encoding the file, or using a nonstandard extension to the .zip format) to keep a resource fork, but any .zip created on a PC is absolutely certain to destroy the resource fork in a Mac file. (Remember, the .zip format was invented on PCs, which have never had resource forks.)

Mac-specific files, such as Mac fonts and most Mac applications, have two "parts," called a "data fork" and a "resource fork." Other kinds of files can also have resource forks; for example, graphics files created by Photoshop may have a preview stored in the resource fork.

The .zip file format was invented for PCs. PC files do not have a resource fork. PC files also do not have Mac "metadata," such as Mac icons and Mac type and creator codes.

When you use the Finder to compress a file into a Zip archive, the Finder breaks the file into two pieces. One piece contains the file's data; the other piece contains the file's icon, the files Type and Creator code, and the file's resource fork (if it has one).

The file's data fork is stored in the Zip archive with the name of the file. The Finder metadata and the resource fork are stored as a separate file whose name begins with "._". So for example if you zip a file called "myvacation.jpg," then the Zip archive will have TWO files in it. One will be called "myvacation.jpg" and will have the actual JPEG data. The other will be called "._myvacation.jpg" and will contain the Type code, the Creator code, the custom icon (if it has one), the resource fork preview (if it has one), and so on.

If you look at the Zip archive on a Mac, you will not see the file called ._myvacation.jpg because the Finder will automatically put it back into the file when you extract it. Take the Zip archive to a PC, though, and a PC user will see both items in the archive--"myvacation.jpg" and "._myvacation.jpg". The file called ._myvacation.jpg can not be used on the PC.

There are other programs that can make Zip archives without saving the resource information and Finder metadata; DropZip, for example, can do this. However, if you zip a Mac specific file like a font or an application this way, the file will be destroyed beyond all hope of recovery.

All computers manufactured after (and including) the B&W G3 will show "Network Disk" in the Startup Disk control panel. What does it mean? It means the computer can boot from a network.

Here's how it works:

You put a computer running MacOS X Server on the network. You configure oS X Server as a "boot server." When the networked computer boots, it does not load the system file from its hard disk, it loads the system file from the OS X network server.

The idea behind this is that a person running a large corporate network would not have to keep updating the system folder on all the computers on the network. He would set up OS X server, and all the computers on the network would boot from that. If he wants to upgrade the operating system on all the computers on the network, he just upgrades the OS on the server. Since the computers all boot from the server, that means they all run the upgraded OS.

Also, in certain high-security settings (think government black ops or super secret research projects), you can set up your computers so that they don't even have a hard disk! They boot from the OS X server computer, they save files on the OS X server computer, they don't even have a hard disk or a CD-ROM installed in them, so nobody can copy software onto or off of them without the network owner knowing about it.

In any event, all systems capable of booting from a network server will show the network volume in the Startup Disk control panel. Just ignore it.

For all intents and purposes, "IDE" and "ATA" in this context mean the same thing. Your hard drive can be one of three basic types: SCSI (which all Macs used to use until the first G3); IDE ATA (which all Macs used up until the G5), and serial ATA or SATA (which G5 systems use).

A hard drive that has bad blocks is a hard drive which is beginning to become defective. The only way to deal with this problem is to reformat the hard drive using the option "zero all sectors." This takes a long time--up to a day or more for big hard drives--and may fail if too many sectors are bad.

If you continue to use a drive that you know has bad sectors,sooner or later your computer will try to record information on those sectors--at which time your computer will probably freeze up and that file will be gone beyond all hope of recovery. For that reason, the best solution to a hard drive with bad sectors is to throw it away and buy a new one.

DiskWarrior will not attempt to locate or fix bad sectors. DiskWarrior is a one-trick pony; all it does is repair damaged disk directories, and it makes no effort to fix any other kind of problem.

The old phony marketing-speak about cable users "sharing" their Internet access isn't actually true--or, at least, it's true in a sense that's also true of DSL users.

With a cable modem, the various subscribers are connected to a "local loop," which essentially means "all the people in one neighborhood who are homed to the same router at the cable company." Yes, they do share bandwidth.

DSL users each have their own, separate connection to the DSL provider's router; however, after the router, they're all connected to the same main connection between the service provider and the Internet.

Ultimately, no matter which way you slice it, there is one connection between the service provider and the Internet, and everyone is sharing that connection. With cable modems, the connection is shared before the router; with DSL, it's shared after the router; but it's still shared.

In terms of real-world speed, the theoretical maximum speed of cable is higher. In most places, this is true of real-world speed as well; in fact, here in Tampa, my cable modem speed is very high indeed--about 50% higher than the speed of a dedicated T-1 line, and roughly three times faster than ADSL.

There are two types of DSL: ADSL, and SDSL. ADSL stands for "asynchronous digital subscriber line;" SDSL stands for "synchronous digital subscriber line." The difference is that ADSL has a different upload and download speed. The download speed is fast, but the upload speed is not; in fact, when you send information from your computer to the Internet, the information is sometimes sent at the same speed as a normal telephone modem. What that means is that if you do a lot of uploading--for example, sending files to a Web site, or transferring large files from your computer to another computer--ADSL is not your best choice.

SDSL has the same upload and download speed; the connection is fast no matter which direction the data is going. However, SDSL is not available in ll places, and in most places where it is available, it is more expensive than ADSL.

Many people who make CDs on a PC use a program that lets you click n files, drag them to the CD just like you would to a floppy, and then add files to the same CD later just by clicking on the files and dragging them to the CD.

It's handy and convenient. What these people do not understand is that it creates a nonstandard CD. The CD it creates can only be read in a PC that has that particular software installed.

You read the files on your Dell most likely because your Dell came with a CD recorder, and so it has that CD burning software in it. Put that same CD in a PC with no CD recorder, you would not be able to read it.

It is very, very important that PC users understand their software, because you can tell a PC to make a standard Cd when it records a CD. Problem is, most PC users don't know that. Most PC users don't even understand that there is more than one way to make a CD!

I've had many clients buy nice, shiny new PCs and start making CDs with no conception of what they are doing, and then give me those CDs. Those CDs are useless to me, because I have both macs and Windows, but without the exact software that the client is using, the CD is useless. In fact, if I have two different clients using two different programs to make CDs, they can't even read each other's CDs!

This can really screw up a PC user BIG time--let's say that you back up your important files onto a CD-ROM, then your computer is hit by lightning and destroyed. You buy a new PC. It comes with a CD recorder, but it comes with different recording software. Guess what? You can't read any of your backup CDs!

These programs all make UDF Cds--but not standard UDF CDs. The manufacturers of these programs all write their own proprietary variations on UDF. Their goal is to make it seem like you can treat a CD like a floppy--you can drag files onto it, you can "delete" files from it (the file is not really deleted--a new directory is written that doesn't include the file, that's all). But there's no way to really do these things with a CD.

The manufacturers make it *seem like* you can do these things by doing all sorts of funky, tricky, nonstandard ways of writing the CD directory.

Some programs have a command that let you "finish," "freeze," "finalize," or "standardize" a CD. (The command is called different things by different programs.) That will tell the program to write a standard CD. If you don't do this, there is no guarantee that the CD will be usable in any other computer--Mac or PC.

I sometimes wish that either computer users were better educated, or manufacturers never created proprietary CD formats--because so many people have no idea what they are doing when they make a CD, and so they have no idea that the CDs they make can't be used by anyone else.”

Color management is a field that's extremely complex and challenging to understand.

The first thing to understand is that almost all graphics programs, and programs like Web browsers, do not understand color profiles, do not do color management, and can not use color profiles. Few programs use color profiles. So if you are saving an image for, say, the Web, you should not include a color profile; the profile makes the file bigger, but all browsers either can not use it or by default do not use it.

So what is a color profile?

Well, a pixel in a picture is nothing more than three numbers: one for red, one for green, and one for blue. For example, the numbers 233, 207, 197 represent a light pink.

But here's the thing. Say you have a computer, and I have a computer, and Bob has a computer. Your computer is a G5 with an Apple Cinema Display; mine is an iBook; Bob's is a PC with a $60 Viewsonic monitor. All three of us put the color 233, 207, 197 on our screens. Guess what? We see three different colors! Who is right? Which one of the three different colors is what the numbers "233, 207, 197" should REALLY look like?

That's where a "color profile" comes in. A color profile is a description of a color. For example, the sRGB color standard says "the numbers 233, 207, 197 should look like thus and such, with this luminance and these chroma values; this is the definition of what 233, 207, 197 looks like and any monitor that shows anything different is wrong." So if Photoshop knows the standard for what 233, 207, 197 is supposed to look like, and it ALSO knows the exact brand of monitor you have and I have and Bob has, it will change the color 233, 207, 197 on your screen and on my screen and on Bob's screen so that they all look the same even though we have very different monitors.

There are many color profiles, and for color management to work, Photoshop has to know about your exact monitor, and your monitor has to be adjusted just right. If everything is adjusted and profiled correctly, and you are using the right profiles, then when you and I and Bob open the image, it will look exactly the same on all our computers.

In practice, few people correctly tell Photoshop about their monitors. In practice, few people use the same profile. (sRGB was intended for images form cheap consumer cameras and scanners, and for images displayed on the Web; it's kind of a "worst-case" color profile that specifies colors in a way that's muted and lacks a wide range of tones.) In practice, Web browsers and most other programs do not use color profiles at all.

If you are in Photoshop and you select View->Proof Setup->Monitor RGB, you will see the colors the way other programs, such as GraphicConverter and Web browsers, will display them. For Web work, that is what you want to do, because Web browsers do not do color management.

Unless you understand color profiling and color management, and you know your system is properly profiled, and you are sharing your pictures with people who understand color management and properly profiled systems, I would not suggest embedding profiles in your images. The profile was put there by your camera, and it's there with the assumptions that (a) you want to use sRGB and (b) you have a properly profiled workflow. If either of these isn't true, you aren't gaining anything by preserving the color profile.

Let’s say you have to print a CD label that will use 3 colors. Is this label going to be printed professionally? If so, what will probably happen is that the CD label will use spot color, rather than process color. However, you can still use shades and tones of color.

There are two ways to make something be color when you are talking about professional printing. The first is "process color"--it always uses 4 colors of ink (the four primary colors Cyan, Magenta, Yellow, and Black). All other colors are made by mixing these 4 primary colors. If you want to make something be purple, you mix cyan and magenta ink. To make something green, you mix cyan and yellow ink.

The second way is by using "spot color," which means that you use a specific color of ink for each separate color in the printed piece, no mixing. If you want to make something be purple, you use purple ink. To make something green, you use green ink.

CDs are often printed in spot color by screen printing. that means that you CAN use a grayscale image--the grayscale image is printed in black ink

Everything you see printed on a press, of any kind whatsoever--magazines, newspapers, postcards, makes no difference--is printed using only four solid colors, no shades of color or gray. The shades are simulated by an optical illusion. The process is called "half toning"--the image is broken into dots of varying sizes. The large dots give the illusion of darker shades; the small dots give the illusion of lighter shades. Look at a newspaper with a magnifying glass and you'll see the halftone dots.

Dot gain occurs when the ink hits the material being printed and spreads out; the ink spread makes the dots bigger (hence the name). Screen printing has a LOT of dot gain.

In very light shades, you can have another problem, "dot loss." The dots are so small that the ink can't be forced through the screen; essentially, any area lighter than about 5% printing just plain disappears, because no ink gets through the screen onto whatever's being printed.

If you're burning from iTunes, set your burning prefs to "audio CD." You can get there under the iTunes menu at the top->preferences->advanced->select burning tab. From there, select the radio button to burn as "audio CD."

In general, I'd burn the CDs as "AIFF". That's what you get when you do it as above in iTunes. BTW, because the files in mp3 CDs are compressed, the AIFF files are much larger files, so you won't get nearly as many songs on the CD. It'll be like a store bought CD--about 15 songs or so, give or take based on the length of the song, file size, etc. The mp3 CDs hold many more songs but may not play on all players.

It's a common misperception that the files on an audio CD are AIFF, because many programs that rip uncompressed audio from CDs create AIFF files. But it isn't true. The AIFF file format was created in 1988; the audio CD format, in 1982.

There are no files on an audio CD at all. An audio CD is just a single stream binary information that represents 16-bit, 44.1KHz stereo data, with a table of contents that indicates how many bytes into the sound stream the start of each song is. There is no disk directory, no file structure, and no list of files on an audio CD; it's just a simple stream of binary information.

When you make a CD that you want to use in an audio CD player, you need to make sure you tell your software to make an audio CD, not an MP3 CD or a computer data CD.

What exactly does the crash log log?

The crashed components or what caused the crash?

The crash log logs the program that crashed, the thread within the program which caused the crash, the processor state at the time of the crash, and the specific exception that triggered the crash. So it logs the program that crashed and why it crashed. However, the information it logs is on such a low level that it's difficult to interpret to a non-programmer.

As a way of analogy, imagine that you're driving along in your car and your tire blows out. The car swerves into a ditch. A person would say "The crash was caused by a blown tire."

The crash log would say "The crash was caused by abrupt and anomalous drag on the right-hand side of the car resulting in an unanticipated vector delta." Not really useful unless you know the language (a "vector delta" is a change in the direction you are moving) and a lot about cars (the most likely cause of high drag on one side of a car but not the other is a flat tire, but it can be caused by other things as well, such as driving with two wheels off the road).