All three word processing programs handled normal-size documents quite well on my laptop, a reasonably configured system with a 598MHz Pentium M processor, 512MB of RAM, and a 5400rpm hard drive. But sometimes you have to work with very large files, so I tested the performance of the three word processing and spreadsheet programs when processing large files.

I created a 1-million-word text file, copied the text into each office suite's word processor, added 20 embedded graphics, and saved the file in the program's native format. The plaintext file was 9.5MB. The graphics ranged in size from 4.8MB to 33KB, averaging 740KB and totaling 14.5 MB. Although this example is probably larger than files you typically work with, it demonstrates how each word processor behaves in the most extreme case.

I tested how long each word processor took to open the file, navigate to the end of the file by using Ctrl+End, and generate statistics for the file. Statistics in all programs included counting the number of characters, words, lines, and pages. Web Table A shows the average of three trials for each test. I closed and reopened the program and file between each trial. Microsoft Office and OpenOffice.org both have Quick Start assistants, which I didn't shut down.

I also tested Microsoft Excel, Corel Quattro Pro, and OpenOffice.org Calc with large files. I created a tab-delimited text file consisting of 250 columns by 20,000 rows. Each cell contained a random number between 0 and 1 million with 10 decimal places. The text file was 82.5MB. I imported the file into each spreadsheet program and created a formula to calculate the sum of each row and average of each column. I timed how long it took to paste the formula across the bottom row and rightmost column, forcing the calculation to be performed for each row and column. Finally, I saved the file in the program's native file format. Web Table B shows a comparison of the three spreadsheets.

Excel did well with this large amount data. Calc was a bit slower than Excel at importing data and was very slow at both saving files to its native format and opening them again. Quattro Pro was very fast when working with its native file format, both in performing calculations and in opening and saving files. However, Quattro Pro was very slow in importing the text file.

Your performance will, of course, vary depending on the activity you're performing. You can find some other performance-test results on the Web. See, for example, the OpenOffice.org 1.1.4 and Office 2003 Standard comparison at http://www.matt13.com/computer/open_office_or_ms_office/index.html.