Databases
The word "datum" is considered, according to dictionary definition, to be "an item of factual information." Multiple instances of this are denoted by the plural, "data." However, although the word should strictly be used only for plural forms, it is not uncommon in contemporary usage to apply it in the singular sense, as in one piece of data or such. A "base" or storage in which data are kept, would be a data base. As with many compound phrases and terms, it is now common practice to use "database" as one word, to denote any collection of meaningful information stored in some organized fashion.
A database may consist of a single file, or of many of them. In the most general sense a database could consist of physical artifacts such as a card catalog in a library or a collection of reference texts, but it is far more common to find the word used almost exclusively to refer to a computer database of files that are referenced by specialized software. The data stored in a database must be persistent, i.e., must not merely consist of temporary values generated during computation and then destroyed; there must likewise be efficient and well-defined means by which data may be queried and updated as required.
A database file consists of records, each record being a unit of information about something. A physical analogue is a library's card index, which consists of cards, each one providing information about a single publication. Within a record, the information is organized into fields, each being a separate entity devoted to one kind of information. To follow our example, a card in a library's card catalog has fields for author, title, call number, location, et cetera. Each field has to have only one kind of information, called the data type for the field. Some fields may take numeric values in a certain format, such as a field that denotes the year a certain title was published. Some may take a text entry, such as a name or title. Yet others may take a Boolean entry, such as a field denoting whether a certain title is or is not checked out at the present time.
A database would ideally be expected to provide fast and convenient access to its data. A database is also expected to preserve data integrity, data independence, and data security, while avoiding data redundancy and providing means for data maintenance.
Data integrity is the quality of validity as pertaining to the data found in a database. This may be affected adversely in a number of ways, such as by human error in keying in the data, by hardware errors, erroneous algorithms or faulty software programs used to update, or errors in data transmission. To avoid compromising data integrity, it is important for a database to use validation procedures that ensure maximum possible integrity. For instance, database entries should be allowed only be certain trusted personnel, not by everyone; likewise, if a user erroneously keys in an input different from the kind expected (such as a name in a date field), the software should flash a warning.
Data independence is the quality of data to be available to various applications rather than to just one. For this reason, it is desirable to avoid proprietary database formats where the data stored can only be read by a certain application. The risk in that case is that only software provided by the vendor who owns the proprietary format can access it, thus rendering the user a slave to the vendor. For maximum ease and freedom, it is desirable that data should be stored in formats that various applications can read, and which allows the user to change the software used at a later date.
Data redundancy is the quality of data to avoid needless repetition of certain parts or the whole because of poor design. For example, if customer addresses and customer orders appear on different, heterogenous databases used by a supplier, then the supplier is likely to have trouble collecting all the billing information cleanly and sending invoices for payment. Such a problem also increases the chances of there being an error.
Data security is the quality of data to be secure from unauthorized access or modification. For example, employee health records stored in a company's database should not be accessible to anyone, including their superiors, and it should not be possible for an employee to access any part of a database other than that which she is actually required to access in the course of her duties. Likewise, the database should be stored (backed up) on a second, distinct medium such as a tape archive, so that in the event of a catastrophic loss due to power failure, system malfunction, or the like, the data is still recoverable.
Instead of a plain database stored on single files, it is now common to use a database management system (DBMS) which can link data from several distinct files. DBMS applications are more specialized than common file management programs, but have several advantages. The most common type of DBMS currently in use is called a relational database management system, or RDBMS. An RDBMS can work with many separate files and can relate data (such as, for example, finding an employee's health records as well as her payroll information, given only the name), thus helping solve the problem of data redundancy.
Another recent kind of database technology is object-oriented databases, which are likely to find much application in future, especially with multimedia data. (Current applications focus for the most part on text data.) This type of database allows the retrieval of data in the form of an "object," within which are multiple aspects such as audio, video, text, and graphics. For instance, a query for a particular film title might bring up an object that included a brief description and review of the film, a short clip with audio content, and some still pictures.
This is the complete article, containing 984 words
(approx. 3 pages at 300 words per page).
