But How Do It Know? - the Basic Principles of Computers for Everyone (27 page)

In a computer, there are a great many bits. As we have seen, a lot of bits have to be set in certain ways in order to make the computer do something useful. The bits in the computer are always there. If you want the computer to do a certain thing, you set those bits on or off according to the pattern that will make the computer do what you want it to do. This pattern is called software. It is not a physical thing, it is just the pattern in which the bits are set.

So the difference between hardware and software isn’t like metal versus rubber. Both metal and rubber are hardware as far as the computer definition is concerned. Hardware is something you can pick up, see, handle. Software is the way the hardware is set. When you buy software, it is recorded on something, usually some kind of disk. The disk is hardware, the specific pattern recorded on that disk is software. Another disk may look just like it, but have completely different software written on it.

Another way to see the difference between hardware and software is how easy it is to send it across a distance. If you have a vase that you want to send to your aunt Millie for her birthday, you have to pack the vase in a box and have a truck take it from your house to her house. But if you want to give her the present of music, you might go to the store, buy her a disk and mail it, but you might also buy her a gift certificate on the Internet, send her an e-mail, and have her download the music. In that case, the music will get to her house without a truck having to go there. The music will be transported solely by the pattern of electricity that comes over the Internet connection to her house.

Another way to see the difference between hardware and software is how easy it is to make a copy of the item. If you have a lawnmower, and want a second lawnmower, there is no machine that will copy the lawnmower. You could photograph the lawnmower, but you’d only have a flat photograph of a lawnmower. You couldn’t mow any lawns with the photo. To get a real second lawnmower, you’d have to go back to the lawnmower factory and build another one out of iron and plastic and rope and whatever else lawnmowers are made out of. This is hardware.

Software can be copied easily by machine. All you need is something that can read the disk or whatever it is recorded on, and something else to write it onto a new disk. The new one will be just like the original, it will do all the same things. If the original is your favorite movie, the copy will also be your favorite movie. If the original is a program that will prepare your tax papers, so will the copy.

Software is not a physical thing, it is just how the physical things are set.

By far the most commonly used definition of ‘software’ is to refer to a package of computer instruction code. I think that the way it got this name is that once you have built a device as versatile as a computer, there are many different things that it can be made to do. But when there are no instructions in it, it can’t do anything. So the software is an absolutely necessary part of a computer that is doing some task. It is a vital part of the total machine, yet it isn’t like any other part in the machine. You can’t weigh it or measure it or pick it up with a pair of pliers. So it is part of the ‘ware,’ but it isn’t hardware. The only thing left to call it is ‘software.’

Programs

As mentioned earlier, a series of instructions in RAM are called a program.

Programs come in many sizes. Generally, a program is a piece of software that has everything needed to do a specific task. A system would be something larger, made up of several programs. A program might be made up of several smaller parts known as ‘routines.’ Routines in turn may be made up of sub-routines.

There are no hard and fast definitions that differentiate between system, program, routine and sub-routine. Program is the general term for all of them, the only difference is their size and the way they are used.

There is another distinction between two types of programs that is not related to their size. Most home and business computers have a number of programs installed on them. Most of these programs are used to do something that the owner wants to do. These are called application programs because they are written to apply the computer to a problem that needs to be solved. There is one program on most computers that is not an application. Its job is to deal with the computer itself and to assist the application programs. This one program that is not an application is called the Operating System.

The Operating System

An “Operating System,” or “OS” for short, is a large program that has many parts and several objectives.

Its first job is to get the computer up and running when you first turn the computer on.

Another one of its jobs is to start and end application programs and give each one time to run. It is the ‘boss’ of every other program on that computer. When more than one program is in RAM, it is the operating system that switches between them. It lets one program run for a small fraction of a second, then another program, then another program. If there are ten programs in RAM, and each one gets to run for one hundredth of a second at a time, each program would be able to execute millions of instructions in that time, several times per second. It would appear that all ten programs were running simultaneously because each one gets to do something, faster than the eye can see.

An Operating system also provides services to application programs. When an application program needs to read from, or write to the disk, or draw letters on the screen, it does not have to do all of the complicated I/O instructions necessary to accomplish the task. The OS has a number of small routines that it keeps in RAM at all times for such purposes.

All an application needs to do to use one of these routines is to load up some information in the registers, and then jump to the address of the proper OS routine. Here’s an example of how it might be done. Lets say you want to draw a character on the screen. First, put the ASCII code of the desired character into R0. Then put row and column numbers of where you want it to appear on the screen into R1 and R2. And here’s the tricky part: You put the address of the next instruction of your application program, into R3. Now just jump to the OS routine. The routine will take care of all of the details of drawing the character on the screen, and then its last instruction will be JMPR R3. Thus, these routines can be ‘called’ from any application, and when done, the routine will jump back to the next instruction in the application that called it.

There are several reasons for having the OS do all of the I/O functions. One is that it makes it easier to write application programs, the programmer does not even need to know how the peripherals actually work. Another reason is that it would waste a lot of RAM if every application had its own copy of all of the I/O routines. One of the most important reasons is that the OS can check to see whether the program should be allowed to do what it is asking to do. This is part of the OS’s other job of being the boss.

The heart of the OS is basically a loop of instructions that asks the following questions: Do I need to input anything? Do I need to output anything? Do I need to let any program run?  Then it starts over again. If the answers to all of these questions is no, the CPU just executes the instructions in this loop over and over, millions of times per second. When there is something to do, it jumps to the beginning of the program that takes care of it, and when that is done, it jumps back to this loop where the OS ‘waits’ for something else to do.

Here is a diagram of our larger RAM version, showing what parts of RAM might be occupied by an Operating System and several other programs.

Within each program’s RAM, there is all of the instruction code that makes the program work. Each program may be divided up into its own main loop, and many routines that are used for the various tasks that it needs to do. As mentioned, the OS also has routines that can be called by other programs.

Each program also uses part of its ‘address space’ for the data that it is working on. The calculator, for example, needs to have a few bytes where it stores the numbers that the user enters into it. Solitaire needs some bytes that specify which cards are in which positions. The word processor needs some RAM for all of the ASCII codes that make up the document you are working on. The OS also needs bytes where it can store fonts, keep track of where application programs have been loaded, receive the data that it reads from the disk, and for many other purposes.

And so this is what goes on inside your average computer. There are many different programs and data areas in RAM. The OS jumps to a program, the program jumps to a routine, the routine jumps to a sub-routine. Each program works on its data or calculates something or does an I/O operation. As each one finishes, it jumps back to where it came from. The CPU executes one instruction from one program at a time, and if they are written intelligently, each program will get its job done piece by piece, without interfering with the rest.

If our computer had included an ‘interrupt system’ like we described a few chapters back, every time someone pressed a key on the keyboard or moved the mouse, there would be an interrupt that would call a part of the OS that determines which I/O device caused the interrupt, and then calls the proper routine to take care of whatever it was. When that was done, the CPU would continue on with the next instruction of whatever program had been running when the interrupt happened.

This can all seem very complex, with so many millions and billions of instructions being executed in the blink of an eye. There are ways of organizing programs and good programming practices that can make it much more understandable. A study of these would simplify software in the same manner that I hope this book has simplified the hardware. But that would be the subject for another entire book.

Languages

Writing programs is very hard to do when you’re just writing ones and zeros, but that is the only code that the CPU ‘understands.’

What is a language? A spoken language, such as English, is a way to represent objects, actions and ideas with sounds. A written language is a way to represent the sounds of a spoken language with symbols on paper. Sounds like another code, and a code representing a code. We just can’t get away from these things!

Do you remember that shorthand we used when we were looking at the CPU instruction code and the wiring in the Control Section? Well, that is actually something more than just a handy tool that was invented for this book. It is a computer language. Here are a few lines of it:

A computer language is a way to represent the instruction code. Its purpose is to make it easier to write computer programs.

In order to use this language, you write the program you want with ASCII characters, and save it into a file. Then you load a special program called a ‘compiler’ into RAM and jump to its first instruction. The compiler will read the ASCII file, translate each line into the Instruction Code that it represents, and write all of the Instruction Code bytes into a second file. The second file may then be loaded into RAM, and when the CPU jumps to its first instruction, the program you wrote in ASCII will hopefully do what you intended it to do.

Of course, when computers were first invented, all programs had to be written directly in ones and zeros. Then somebody got tired of the tedium of programming that way, and decided to write the first compiler. Then ever after, programs were written in this easier language, and then translated into Instruction Code by the compiler. With the original compiler, you could even write a better compiler.

So in order for a computer language to exist, you need two things, a set of words that make up the language (another code,) and a compiler that compiles the written language into computer instruction code.

The language that we have seen in this book has only about 20 words in it. Each word correlates directly to one of the instructions of which this computer is capable. Each line you write results in one computer instruction. When you write an 87 line program in this language, the instruction code file that the compiler generates will have 87 instructions in it.

Then someone invented a “higher level” language where one line of the language could result in multiple computer instructions. For example, our computer does not have an instruction that does subtraction. But the compiler could be designed so that it would recognize a new word in the language like ‘SUB RA,RB’ and then generate however many machine instructions were necessary to make the subtraction happen. If you can figure out how to do something fancy with 47 instructions, you can have a word in your language that means that fancy thing.