[plug] OpenOffice.org to the rescue

[Leon Brooks]

On Mon, 21 Jul 2003 21:29, J Michael Gilks wrote:
> I did have the thought that perhaps Microsoft programmers know the
> format and therefore don't check as rigorously for what really is
> there. ie OpenOffice.org will check that the bytes it is reading are
> really what it wants to see, whereas Microsoft just assumes that the
> bytes are correct, gets out of step and cannot recover.

Surprisingly close. An MS-Office file is basically an OLE2 memory dump. 
When an MSO product reads the file in it basically undumps it and fixes 
up some address chains. If the file's slightly frotzed the normal 
reaction of the MSO component is to die, occasionally taking the entire 
OS to hell with it.

OOo actually thinks about what it's reading, and still manages to read 
it quickly.

Cheers; Leon