Navigation with Browsers

Relative Addresses

URL Addresses | URL Details | Relative Addresses | URL Summary | Methods of Navigation | Going Further

The term "path" is sometimes confusing for new users of the Web. However, it is really a simple concept: A path is the sequence of directories that must be traversed to get from one directory to another. (In Windows directories are often called "folders"; we will use the two terms as synonymns.) In order to construct all but the simplest of Web documents, it is useful to know the basics of how to specify paths between directories and files.

The subsequent material in this section is somewhat more technical than the preceding discussion. It is not essential to casual Web use, but is important if you plan to construct your own Web site.

Directory Hierarchies

Consider the directory structure shown in the figure below.

This typical directory structure has a master directory (usually called the root directory and designated on a PC by the symbol "\") that contains within it all other directories as subdirectories, and each directory so contained may itself contain subdirectories. In this example the root directory contains the directory Dir1, and this directory in turn contains the directories Dir2 and Dir5. Finally, Dir2 contains subdirectories Dir3 and Dir4, while Dir5 contains a subdirectory Dir6.

The files on the system are contained in these directories. In the figure some representative files are illustrated in red. For example, the file index.html (see the green arrow) is contained in the directory Dir2, which is a subdirectory of Dir1, which is finally a subdirectory of the root directory.

Directory Trees

Another useful way to look at this directory structure is in terms of a Directory Tree. The adjacent figure shows a tree diagram corresponding to the directory structure illustrated in the first figure. If we view this as a tree (fallen on its side), the branches of the tree correspond to the subdirectory structure. If you stand the directory tree upright, this analogy also makes clear why the master directory containing all other directories is termed the "root" directory.

Paths

The directory tree diagram illustrates graphically the meaning of a path: it literally is a path: the (shortest) one that we would follow on the directory tree to go from one directory to another. For example, consider the files illustrated in the first diagram. Suppose we ask what the "path" is from the file index.html to the file file3.ext. The file index.html is in the directory Dir2 and file3.ext is in the directory Dir3. From the tree diagram, to get from index.html to file3.ext we must first go to the directory Dir3, which is contained in Dir2. We say that (relative to the file index.html) the file file3.ext is at the location Dir3\file3.ext. Likewise, relative to the directory Dir1, the file file3.ext is at Dir2\Dir3\file3.ext, and so on. Examples such as these are termed "specifying the relative path to a file" from some starting point.

Specification of Paths

Thus, the complete path to a file that lies in a subdirectory of your present directory is specified by listing, in order, the directories (omit the one that you are in) that you must pass through to get to the file, with backslashes after each directory name, and followed finally by the file name. Pretty simple, once you get the hang of it!

But what if the file is not in a subdirectory of your present directory? For example, where is the file file6.ext (which is in Dir6) relative to the file index.html? The same rule applies as before, but the adjacent figure illustrates that to get to the directory Dir6 from the directory Dir2 we must first go BACK in the directory hierarchy until we reach a directory that contains Dir6 (and, of course, Dir2) in its subdirectory structure.

Thus, to reach file6.ext from index.html we go back to Dir1, then to its subdirectory Dir5, and then to its subdirectory Dir6. The notation that is used to go back one level in a directory hierarchy is "..\" (two dots, followed by a backslash), and we may say that the location of file6.ext relative to index.html is ..\Dir5\Dir6\file6.html, as illustrated in the preceding figure.

Absolute and Relative Addressing

The examples of addresses we have just been considering are termed "Relative Addressing", because the location is specified relative to our present location. However, notice that we could also specify an "Absolute Address" for files by going back to the root directory and writing the complete path to the file from there. In the previous example, the absolute address of file6.ext is \Dir1\Dir5\Dir6\file6.ext, where the initial "\" stands for the root directory.

Notice the difference between absolute and relative addresses: if I move the file index.html to a new directory, the address of file6.ext relative to index.html changes, but its absolute address remains the same.

On the other hand, suppose I move the entire directory structure to the right of "Root" in the directory tree intact to a new location on the same computer or a different computer, as illustrated in the adjacent diagram. Then generally the absolute address of file6.ext would change, but its address relative to index.html (indeed, relative to any file in the directories that I moved) would stay the same.

Thus, in the preceding diagram the absolute address of file6.ext is \Dir1\Dir5\Dir6\file6.html in the upper example and \new\www\Dir1\Dir5\Dir6\file6.html in the lower example, but since the part inside the gray boxes is the same in the two cases, the relative address as referenced from index.html is unchanged.

The Connection Between Paths and URLs

So why should you worry about paths, whether absolute or relative, if you are creating web documents? Because the path is the part of a web document's URL that gives the document's location on the server on which it resides. Just as the location of a file can be given relative to the location of another file, or can be given as the absolute path to the file (which is really the file's location relative to the computer's root directory), so the path portion of a document's URL can also be specified relative to the location of another document or in terms of an absolute location. In the latter case, one refers to an absolute URL, and it is given, not in terms of the location relative to the computer's root, but in terms of the base directory for all the html documents. This directory is often called the html document root.

Thus to determine the appropriate path portion of a document's absolute URL, one needs to know its directory path on the server, relative to the html document root. If the server is a Windows machine, it has a path of the form that we discussed above. To get the path portion of the document's absolute URL, simply:

1. write the directory path of the file relative to the html document root directory

2. replace every backslash in the path with a (forward) slash

To obtain the complete absolute URL, combine this path part with the host name and the string (http://, ftp://, etc.) that gives the method of file access.

To determine the URL of document A relative to document B, just write the directory path of A relative to B and replace each backslash with a (forward) slash.

So why, you may well ask, do I have to worry about both slashes and backslashes? Why couldn't we just use one or the other for both directory paths and URLs? The answer lies in the fact that URLs and web browsers were first developed on Unix machines. There the forward slash is used in directory paths and they were naturally carried over into URLs. Unfortunately, however, the backslash already had a long history of serving as the delimiter in directory paths in the PC world, so it was retained in this role even after Windows machines began to be used as Web servers. Consequently, in the Windows environment one has to use two different symbols, while you can get away with only one in the Unix world.

Use Relative Addressing Whenever Possible

Absolute and relative addressing clearly each have advantages and disadvantages, but in producing Web pages it is advisable to make addresses to files on the same machine relative addresses unless there are circumstances dictating otherwise. Then, if the directories containing your Web pages are moved intact to another machine the links among these files should still work properly. If instead the files are specified by their absolute addresses, the links to local files in your directories will often "break" when you move the files, because the absolute addresses will generally change. Repairing these broken links can be a time-consuming task if your site involves many files.

Homepage Exercise: Relative and Absolute Addresses

If you wish, you may now use your homepage folder and files to illustrate some of these ideas with this exercise

The elementary examples in this exercise illustrate the difference between absolute and relative addressing, and how to implement relative addressing for the two simple (but common) cases where the file being addressed is either in the same directory as the HTML file (the next.gif example), or in an immediate subdirectory of the HTML file (the gifs\previous.gif example). As we have noted above, it is generally advisable to use relative addressing in your Web pages to increase portability.