Several characteristics of most commercial database packages make them unsuitable for many applications. First, prospective database developers frequently want to distribute the data they have collected and inputted, because others would find it useful. But while the raw data files can be distributed legally, the commercial package required to search and access the data is copyrighted.
The second problem is that most of the popular database packages are relational databases, and are therefore appropriate only for specific types of information. A relational database requires that every field length be defined before inputting data. For inventory or address lists, where fields tend to be close to the same length, the loss in disk space is tolerable when shorter fields are padded out with spaces to equal the defined length. But for other types of databases, such as those with annotation fields, what happens when the typical entry is 50 characters, and you still have to define for the worst case, which might be 250 characters or more?
Data files quickly become too big to manage efficiently, which leads to another problem. Very large files require special programming to break up the file into chunks that are small enough to fit on a backup floppy disk. Moreover, large relational files are delicate -- one missing or extra character can throw off the entire database.
These days, ever since Microsoft's "Visual" packages, database software has become so bloated, and so locked into Windows, that one's distribution options, as well as cross-platform conversion options, can be limited. And if you standardize on SQL, you've already exceeded the capacity of most microcomputers. A better approach might be a flat-file, delimited ASCII database that is so simple, that development can be done on a single-floppy laptop. Years later, when data entry is finished, it can be scaled up to a different search engine with a little custom programming in BASIC, or searched directly with a C-language program on a Linux server. ASCII won't be obsolete ten years from now, but (hopefully) Windows will.
MiniBase solves all these problems. It runs under DOS, and requires only 130K of memory and a single floppy or hard disk. It is not copyrighted or copy-protected; once you register it with Public Information Research, it becomes yours to redistribute as you wish, with the assumption that any fees you charge for distribution reflect the value added by your data.
It is not a relational database, so only the characters used in the actual inputting take up room in the data files. Because it uses 52 alphabetized data files, the DOS copy command is all that is needed to back up a large database onto floppies from a hard disk. The program itself requires less than 100K of disk space, leaving room for data files even on a single floppy system. Each record allows a heading, date, and up to a full screen of annotation, all of which can be searched effectively with powerful Boolean AND/OR combinations.
The major limitation in MiniBase is that the maximum record length should not exceed one full screen of text (additional records may be added under the same heading if one screen is not enough). The editing commands for modifying a previously-entered record work on a single screen, on one record at a time. This editor can handle cursor movement, backspace, deletions, overtyping, and insertions, but cannot do cut and paste.
MiniBase searching is powerful enough to zero in on any record in several seconds. After years of inputting, it might take 15 seconds or so on today's microcomputers, if you don't specify a heading and force an A-Z search of the entire database.
A heading is typically a proper name (if an individual, last name first), a subject heading, or a mix of both of these in a single database. By specifying either an exact heading (by adding a period) or the leading letters of a heading (default mode) in a search, you have already excluded 98% of the database, so searches will be almost instantaneous. You may also require that an additional "string" (a keyword or phrase, or portion thereof, as long as the characters, including spaces, are contiguous) be found in the annotation field before the record is produced. Beyond this, you may also specify a second string in the annotation field, in addition to a date, or a date alone. In other words, after a heading any combination of one or two strings or date may be specified.
If you like, the heading may be skipped, and the same combinations of strings and date specified. If the heading is skipped, the entire database will be searched. Searches with or without headings can all be stacked together up to 50 deep at a time to allow for various possibilities and combinations that may be of interest. An "Esc" entered for a heading will back up one line so that a mistake can be corrected, or a solitary period will close the stack and begin the search. The results of a search always display on the screen, but may also be printed or sent to a new user-named file. A search for a single line of search terms may also be done by passing them as a command-line parameter when MiniBase is first loaded.
MiniBase is freeware. Download the complete package as a 74,123-byte zipped file.