Computer file management systems can be classified by the type of file organization they provide (Flat File or Hierarchical File Systems); and the type of user interface they present (Character/Command or Graphical/Event Driven). These file organizations and interfaces greatly affect how users manipulate and interact with their information.
Early computer file management systems (MSDOS 1.0
, IBM CMS, etc.) only
allow a linear list of the files. In order to group related files, users were forced to
name the related files very similar, usually varying only the last 1 or 2 characters of
the name or the type (extension) of the file. To be fair CMS also allows a user to
maintain several 'virtual' disks which do provide a minimal level of hierarchical
organization. A contributing limiting factor in these file systems was the restriction on
the name of the file length (usually 8 characters -- still present in DOS and Windows 3.1
today). This serious drawback often requires the use of cryptic file names, where in the
worst case each character is a code with a specific meaning. Fortunately these systems
have declined in use due to the popularity of hierarchical file systems.
Hierarchical file systems (MSDOS
2.0 or greater, UNIX,
etc. ), allow the creation of files which can
contain other files, (directories). This allows users to group related files
together in directories, often nested several levels deep.
Physically files are stored on a disk which is ordered linearly.
It is only through the interface that the user perceives the hard disk as being a hierarchically ordered device. The advantage of transforming information is clear in this example. Users are given a powerful organizational tool through a virtual transformation of the file layout data.
One side effect of this organization that has arose is that some users tend to duplicate files so that they appear in several directories. This leads to problems of data redundancy and file synchronization (aka data integrity). These problems can be avoided through the use of an alias (link), feature which creates a dummy file in one directory to refer to the actual file in another directory. Changing the dummy file in reality changes the actual file because multiple copies of the file do not exist.
If a picture is worth a thousand words how much is an icon worth? The following sections will contrast some of the effects upon information organization that character and graphical interfaces exhibit.
Character/command driven interfaces arose out of necessity. The early processors were not powerful enough to handle the large amounts of data required for image processing. Some computer users argue that an experienced person can accomplish similar tasks faster with a character interface than a graphical interface. While this claim may be true, it requires a high learning curve in order for someone to become experienced. The high learning curve is due to the heavy cognitive memory load required of users. There is little doubt that inexperienced users can accomplish tasks faster with a graphical interface than a character interface.
Many of the complaints against character based interfaces are due to the lack of standards. This requires users to learn different commands for accomplishing the same task in different programs. More standardization has occurred for graphical interfaces due to the lessons learned previously; and the fact that graphical programming, being more difficult, allows standards to be more easily enforced.
Command driven interfaces only allow a user to execute a particular set of operations at any given time. This restricts the types of information sharing between tools that users can easily perform. It forces a linear task ordering pattern upon people who naturally tend to switch from task to task as they freely associate information with different tasks.
Because the human mind is much more adept at processing image information quickly than textual information, graphical user interfaces (GUI) such as Macintosh, MS Windows, NeXT) are rapidly becomming defacto standards. A GUI is sometimes referred to as a WIMP (Windows Icons Mouse Pull-down menus) interface. Although somewhat derogatory, this acronym accuractely gives the elements that comprise a GUI.
While many of the advantages of graphical interfaces are apparent there are
disadvantages. Iconic representation of information is not the most efficient in terms of
compactness. Iconic languages such as Chinese and Japanese (Kanji) , have extremely large icon, i.e. character, sets. GUIs are
event-driven interfaces. This forces programs to be able to react to multiple events at
any given time. Events such as menus being accessed; text being typed; icons from palettes
being chosen; switching to other programs, etc. Allowing users to quickly shift between
applications provides them with the ability to interleave tasks and work concurrently.
A person may or may or may not be able to actually perform work in parallel depending u
pon whether the underlying operating system is multi-tasking (UNIX, OS/2, Windows 4.0),
suspended tasking (Macintosh, Windows 3.1), or single task (DOS).
