Scor­pio News


July–September 1988, Volume 2, Issue 3

Page 9 of 39

The Use of Overlay Files

By R. Pearce

Undoubtedly one of the biggest advantages of the newer 16-bit micros over our old Z80 systems is the sheer size of the available memory. This extra memory allows more complex programs to carry out more involved operations on larger blocks of data with fewer disk accesses. The shortage of memory space can be a major drawback on 8-bit systems, but there are ways around it.

Perhaps the classic example of memory shortage is the text editor/word processor. Here we have a program with many tasks to perform on a large amount of data. The more options we put in the less text can be held in memory. Some cheap editors such as HiSoft ED80 or GEMPEN simply restrict the facilities and limit the size of text file that can be handled. Research Machines TXED allows long files by editing one section at a time, which works but is very ugly and rather inconvenient. Wordstar, however, manages to do just about everything you could ask of it (except fit on an SDSS disk) and handles 32Mbyte files to boot. It achieves this by continuously juggling both code and text between memory and disk.

In this article I intend to consider only the juggling of code, since data juggling using random access files seems a more popular subject and thus there is much more written about it.

A few months back I wrote a review in this journal of a printed circuit board design package. It so happened that as I was writing the review I was in the process of developing a similar package of my own. Both packages had run into the problem of memory shortage but our approaches differed. I must confess that at least part of the reason for this was that the compiler used for the program I reviewed supported a chain command which my compiler does not have. I would argue, however, that this is fairly irrelevant since my single file solution breaks the file down into only 8 parts compared with some eleven files in the other.

The question to be asked here is not So much how the two programs go about loading the required part but why the divisions occur where they do. Specifically, of course, why doI think my version is better. The logic that led to my approach went something like this:

The program must have facilities for editing, saving, , loading and printing the file. Also useful would be a tidy-up facility for plotter dumps.

This is an OCR’d version of the scanned page and likely contains recognition errors.

Page 9 of 39