Friday, October 07, 2005

OpenOffice

I’d also like to say that in general OpenOffice.org is a pretty lame piece of software. is rather good but has quite a number of areas for improvement (sorry about the previous negative attitude ;-)

First off, does it even use a standard GUI toolkit? No. Why not? Because it’s legacy software: OO.o has a long history, dating back to its StarOffice days before coming under the auspices of Sun Microsystems. First order of business: move over to Qt or Gtk, both of which are ported onto every significant platform OO.o would want to be involved in.

Second: it’s huge and slow. The “Options” window is cluttered worse the the KDE Control Center (and that says quite a bit). Some of the components depend on Java. The “Draw” program can’t even import SVG files and never prints correctly. Because it doesn’t use a standard toolkit, it has issues with some Unicode characters, even when using a font that supports those characters. Usability and UI design also seem to be a low priority for the project (I know it’s not, but it appears so).

So what do I recommend? Since the likelihood of OpenOffice ever actually being ported to a standard GUI toolkit is essentially zero, I recommend that development effort be focused on bringing apps such as Gnumeric and Abiword up to speed with the featurefulness of OO.o.

My dream recommendation, given any amount of resources being made available, is that OO.o be ported to Gtk. Then, it would be restructured to be the Firefox of office software - a lean core functionality and a flexible “Extensions” mechanism. This would hopefully lower the barrier-to-entry for programmers sufficiently as to encourage wide participation by people interested in improving the office suite. The lean core could then be focused on and maintained more effectively than the massive bulk of the do-it-all office program.