But How Do It Know? - the Basic Principles of Computers for Everyone (28 page)

Then someone invented an even higher level language where the words that make up the language don’t even resemble the CPU’s actual instructions. The compiler has a lot more work to do, but still generates instruction code that does the things that the words in that language mean. A few lines from a higher level language might look like this:

Balance = 2,000

Interest Rate = .034

Print “Hello Joe, your interest this year is: $”

Print Balance X Interest Rate

The compiler for this language would read this four-line program, and generate a file that could easily contain hundreds of bytes of instruction code. When that instruction code was loaded into RAM and run, it would print:

Hello Joe, your interest this year is: $68

Writing software in higher level languages can result in getting a lot more done in a shorter amount of time, and the programmer no longer needs to know exactly how the computer actually works.

There are many computer languages. Some languages are designed to do scientific work, some are designed for business purposes, others are more general purpose. Lower level languages are still the best for certain purposes.

The File System

As we saw earlier, the way a disk actually works is pretty foreign to most people who use a computer.

To make things easier, someone invented an idea called a “file.” A file is supposed to be similar to the kind of paper files that people use in offices. A paper file is a sheet of cardboard folded in half and placed in a file cabinet. This folder has a tab on it where you can write some sort of name for the folder, and then you can put one or many pieces of paper in the folder.

A computer file is a string of bytes that can be any length, from one byte up to all of the bytes available on the disk. A file also has a name. A disk may have many files on it, each with its own name.

Of course, these files are just an idea. To make a file system work, the operating system provides a bunch of software that makes the disk appear to be like a filing cabinet instead of having heads, tracks, sectors and blocks of bytes.

This file system gives application programs an easy way of using the disk. Applications can ask the OS to create, read, write or erase something called a file. All the application needs to know is the name of the file. You open it, request bytes, send it bytes, make it bigger or smaller, close the file.

The OS uses part of the disk to maintain a list of file names, along with the length of each file and the disk address (head, track, sector) of the first sector of the data. If the file is smaller than a disk sector, that’s all you need, but if the file is larger than one sector, then there is also a list which contains as many disk-type addresses as needed to hold the file.

The application program says create a file with the name “letter to Jane.” Then the user types the letter to Jane and saves it. The program tells the OS where the letter is in RAM and how long it is, and the OS writes it to disk in the proper sector or sectors and updates the file length and any necessary lists of disk-type addresses.

To use the file system, there will be some sort of rules that the application program needs to follow. If you want to write some bytes to the disk, you would need to tell the OS the name of the file, the RAM address of the bytes that you want to write, and how many bytes to write. Typically, you would put all of this information in a series of bytes somewhere in RAM, and then put the RAM address of the first byte of this information in one of the registers, and then execute a Jump instruction that jumps to a routine within the Operating System that writes files to the disk. All of the details are taken care of by this routine, which is part of the OS.

If you ask the OS to look at your disk, it will show you a list of all the file names, and usually their sizes and the date and time when they were last written to.

You can store all sorts of things in files. Files usually have names that are made up of two parts separated by a period like “xxxx.yyy.” The part before the period is some sort of a name like “letter to Jane,” and the part after the dot is some sort of a type like “doc” which is short for “document.” The part before the period tells you something about what is in the file. The part after the dot tells you what type of data is contained in this file, in other words, what code it uses.

The type of the file tells both you and the OS what code the data in the file uses. In one popular operating system “.txt” means text, which means that the file contains ASCII. A “.bmp” means BitMaP, which is a picture. A “.exe” means executable, which means it is a program and therefore contains Instruction Code.

If you ask the OS what programs are available to execute, it will show you a list of the files that end with “.exe”. If you ask for a list of pictures that you can look at, it will show you a list of files that end with “.bmp”.

There are many possible file types, any program can invent its own type, and use any code or combination of codes.

Errors

The computer is a fairly complex machine that does a series of simple things one after another very quickly. What sorts of things could go wrong here?

In the early days of computing, when each gate in the computer was relatively expensive to build, sometimes there were components that actually had moving parts to make electrical connections. Two pieces of metal had to touch to make the electricity go to where the builders wanted it to go. Sometimes when the machine stopped working correctly, the fixit guy would look inside to find out what was wrong, and he would find that a spider had crawled inside the machine and had gotten itself wedged in between two of these pieces of metal that were supposed to touch each other. Then when one piece of metal moved to touch the other, the spider was in the way and they wouldn’t touch. So the electricity wouldn’t get to where it needed to go, and the machine would not operate correctly anymore. The fixit guy would remove the bug, clean up the contacts, and report “There was a bug in the computer.” And he literally meant a bug.

Over time, whenever a computer appeared to be operating incorrectly, people would say that the computer had a bug. There are two main classes of computer bugs: hardware and software.

A hardware bug actually means that the computer is broken. This could be as serious as you turn the computer on, and it catches fire, to there is one byte in the RAM where one bit is always off.

Now one bit in RAM that refuses to change may be a problem or it may not. If the byte where that bit is located somehow never gets used, then the computer will work just fine. If that byte is part of a place where a name is stored, then the name may get changed from “Joe” to “Jod.” If that byte has some program instructions in it, you may get an XOR instruction changed to a JMP instruction. Then when the program gets to that instruction, it will not do the XOR like it is supposed to, but rather it will jump somewhere else and start executing whatever is at the new location as though it was a series of instructions. The contents of those bytes will determine what happens next, but it will almost certainly be as wrong as a train falling off its track.

If a gate is broken in the stepper, for instance, so that step 4 never comes on, then the computer will not really be able to operate at all. It would still be able to fetch instructions in steps 1, 2 and 3, but every instruction would execute incorrectly. Certainly the program would make a mess of things after ‘executing’ just a few instructions.

Software bugs can take many forms, but they are all ultimately programmer mistakes. There are probably many more ways to write a program incorrectly than correctly. Some errors just create some kind of incorrect results, and other errors cause the computer to “crash.”

One of my favorite stupid programmer stories is this: Someone bought a car on credit. He got a coupon book with the loan, one coupon to be sent in with each payment. But when he made his first payment, he accidentally used the last coupon in the book instead of the first one. A few weeks later, he received a computer-generated letter from the loan company saying, “Thank you for paying off your loan in full, next time you need a loan please use us again.” Obviously, the program just checked the coupon number and if it was equal to the highest number coupon in the book, jump to the routine for a paid-in-full loan. It should have at least checked the balance remaining on the loan before deciding that it was paid off. This is a subtle error, it might not be caught by the loan company until they audited their books months later. The computer did exactly what it was told to do, and most of the time it was adequate, but the program was not written to anticipate all of the situations that sometimes occur in the real world.

One of the worst software bugs is getting stuck in a loop. The program executes a series of instructions, and then jumps back to the beginning of the series and executes it over and over again. Of course, loops are used all the time in programming, but they are used to do something that has a finite number of similar steps. It may repeat until 50 bytes have been moved somewhere, or keep checking for the user to press a key on the keyboard. But the computer will exit the loop at some point and continue on to its next task. But if there is some sort of programming error where there is a loop that has no way out, the computer will appear to be completely stuck. This is sometimes called being ‘hung,’ the whole computer may need to be turned off and restarted to get out of the loop and back into useful operation.

There are all sorts of errors that end up with the CPU trying to execute something other than instruction code. Lets say you have your program residing at address 10 through 150, and you have some ASCII data such as names and phone numbers at addresses 151 through 210. If the program is written incorrectly so that under certain conditions it will jump to address 180, it will just continue fetching and executing the bytes starting at address 180. If 180-189 was filled with the ASCII for “Jane Smith,” the “program” will now be executing complete garbage, a series of bytes that were not designed to be Instruction Code. It may put itself into a loop, or jump back somewhere into the program, or issue the command to erase the disk drive. And it will be doing garbage at its usual high speed. If you looked at the patterns in the bytes, you could see what it would do, but it could be just about anything. If the name at address 180 was “Bill Jones”, it would do something completely different. Since it is not designed to be useful, most likely it will just keep making a bigger mess out of what is in memory until the computer will have to be powered off to get it to stop.

Another type of error could occur if a program accidentally wrote “John Smith” into the place where a font was stored. In that case, every letter “E” that got drawn on the screen thereafter would look like this: ‘
.’

The computer executes hundreds of millions of instructions every second, and it only takes one wrong instruction to bring the whole thing to a screeching halt. Therefore, the subject of programming computers in a manner that will be completely ‘bug free’ is something that gets a lot of attention. Almost all programming is done with languages, and the compilers for these languages are designed to generate Instruction Code that avoids the most serious types of errors, and to warn the programmer if certain good programming practices are violated. Still, compilers can have errors, and they will never be able to spot an error like the one above with the car loan.

As you can see, the computer and its software are pretty fragile things. Every gate has to work every time, and every instruction that gets executed has to be correct. When you consider all of the things that could go wrong, the high percentage of things that normally go right is actually quite impressive.

Computer Diseases?

Another place where human characteristics get assigned to computers is something called a computer virus. This implies that computers can come down with a disease and get sick. Are they going to start coughing and sneezing? Will they catch a cold or the chicken pox? What exactly is a computer virus?

A computer virus is a program written by someone who wants to do something bad to you and your computer. It is a program that will do some sort of mischief to your computer when it runs. The motivation of people who write virus programs ranges from the simple technical challenge of seeing whether one is capable of doing it, to a desire to bring down the economy of the whole world. In any case, the people who do such things do not have your best interests in mind.

How does a computer ‘catch’ a virus? A virus program has to be placed in your RAM, and your computer has to jump to the virus program and run it. When it runs, it locates a file that is already on your hard disk, that contains a program that gets run on a regular basis by your computer, like some part of the operating system. After the virus program locates this file, it copies the virus program to the end of this file, and inserts a jump instruction at the beginning of the file that causes a jump to where the virus program is. Now your computer has a virus.