GUI Bytes

It's All In the Name

a logical filename structure can make it easy to find files

One of the real improvements in the DOS/Windows world was the introduction of "long filenames". For the benefit of those who are relative newcomers to computers, we used to have to deal with filenames in the "8.3" format - eight characters, a decimal point (or period), and three more characters. So we had files with names like SWJK0601.DOC and MTGSCHED.TXT.

Under this system, only the first eight characters were really useful, as software programs quickly appropriated the three characters after the decimal for their own purposes. Standardization of these characters was also beneficial for the user, as it helped identify the type of each file. For example, dbf indicates that a file is a database, txt that it is plain text, wpd that it is a WordPerfect file, and so on.

In theory, even the remaining eight characters would allow a very large number of unique filenames. Given the use of thirty-six alphanumeric characters (twenty-six letters plus the numerals 0 through 9), we could have thirty-six to the eighth - a few trillion - filenames. Because some other characters, such as the dash or the underscore, could be used, the number would actually be higher.

In practice, though, users generally chose filenames that meant something to them - or at least as much as eight characters would allow. Examples would be ltrtojohn.doc (letter to john), or ltr2mom1.wpd (letter to mom no. 1). Obviously, such names would use up all of the meaningful "words" quickly; what would the name of the eleventh letter to mom be?

Some companies instituted naming conventions so any user might be able to determine something about the file, even if a certain amount of decoding were necessary. Thus, bd00001.doc might be the first document from Bob Dole. Directories could also be used to allow more filenames; if all of Bob's letters were in one directory, then all eight characters could be used to identify the contents. Still, filenames were cryptic at best.

raising the limit

Most users now regularly use long filenames, and it is now much easier to determine what is in a file. Even with the new limit of 255 characters, it is no longer necessary to use a code table to name files. But - is this really a good thing?

This paragraph has less than 255 characters, so it could be used as a filename. But how useful would it be? You can't read more than a few words in any dialog box. From the many computers I have seen, it appears that file naming has gone from cryptography to chaos.

A standard naming convention is still useful, and you should be using one. It will let all users find things more quickly, it can make files automatically sort in a preferred order, and searches will be more effective.

Consider the following examples. The first part of the filename indicates which committee is involved, the second part indicates the document type (agenda, minutes, etc.), and the third part is the document date.

TC agd 010605.doc

This would be a Technical Committee agenda dated 05 June 2001.

BOD min 010729.doc 

This would be Board of Directors minutes dated 29 July  2001.

If several files of this format were in the same directory, they would automatically sort by committee, then by document type, and then by date. A similar system can be used to advantage in any office, with project number first, followed by type of document, then date, and project manager.

Use mixed case to make the names easier to read. Avoid the underscore as it gets lost when filenames are used in text or in hyperlinks. In order for dates to sort correctly, the date must be arranged first by year, then by month, then by day, using leading zeros.

If you have a filename convention that works particularly well, send it to me. I'll include interesting examples in a future column.

2001 Sheldon Wolfe, RA, CSI, CCS, CCCA, 
on the web at 
July 2001

home page

Web site design and content Copyright  1995-2004 Sheldon Wolfe

Material from CSI Chapter newsletters used with permission.