Computers have revolutionised our lives in so many ways, bringing exciting new digital technologies into our homes and transforming our productivity at work. Just occasionally, however, their limitations become all too apparent. Occasions when old text documents need to be digitised are an example in point.

Converting large, sprawling reports into digital format is a nightmare, but help is at hand in the form of optical character recognition software, or OCR as it is more commonly known. The fact that OCR software can take scanned pages of text and graphics and convert them into editable documents should be manna from heaven – after all, it has the potential to save users hours of tedious typing. The problem is that OCR software has always been a bit hit and miss. One program that promises a breakthrough on the accuracy front, however, is Readiris Pro.

Now in its ninth incarnation, Readiris Pro works by intelligently analysing a scanned document or a PDF file and turning text areas, graphics and information contained in tables into zones. It then converts user-defined selections into editable documents. With 104 languages supported, Readiris can recognise characters from the Cyrillic alphabets and even Chinese character sets. According to the program’s developer, the OCR engine works on a linguistic basis, resolving unclear characters through recognition of the letters surrounding it. The program can also be taught to recognise characters as you use it which improves its accuracy.

So what are the results like? Well, a good way to get started with the program is to use one of the sample documents provided. Load one into the interface, click on the Page Analysis button and select the areas of the document that you wish to convert. This takes a good bit of fiddling because you’ll want to make sure that text documents laid out in several different columns are converted in a logical fashion. When you’re done you select a language, click on the ‘Recognise’ button and let Readiris get on with the character recognition process. Once this has finished you can then open the document in a predetermined word processing package – both Microsoft Word and TextEdit are supported. We were surprised by how good the results were when using the provided sample files; not a single character was out of place and all the text was correctly formatted and presented in the Times typeface.