7 Mar 2004
(updated 7 Mar 2004 at 01:01 UTC) »
Following hp's blog about GNOME Office I think that I should have a few comments. Basically he is right. One of the reason GNOME Office is not that successful is that OpenOffice arrived with a decent feature set and integration beetween each module, which led the major vendors like Ximian giving their support to it instead of investing a lot more on GNOME Office. It is not really a missed opportunity as he state, because the GO (Abiword, Gnumeric, etc.) developers do all they can do: their best. The lack of "integration", which would come by the lack of embedding each other thru something like OLE/Bonobo is something we know is missing. One of the biggest problem comes from the cross-platform approach. I need to recall Abiword history to explain this:
Abiword was started in 1998 by Sourcegear. The goal was to write an cross-platform and free (libre) office suite whose first brick was a word processor, Abiword. But more modules where planned: spreadsheet, presentation program, etc. The cross-platform approach that was chosen was to not write a custom widget set like it is done in OOo, but rather provide a more general framework and let the GUI be implemented for each platform, to some extent. It currently supports Gtk2/GNOME, QNX, Windows, BeOS and MacOS X (not using X11), looking like a real app developped for each platform.
The task appeared to be harder, and the business plan of Sourcegear related to AbiSuite changed: they threw the towel for Abiword in 2000. But development continued, lead by the people that are still working on it: cinamod, uwog, msevior, etc. And the good platform integration of Abiword, as well as the GNOME variant that cinamod brought us made Abiword the definitive pickup for a GNOME word processor.
cross-platform approach on the other side is to port it using GTK+
ports on Win32. All in all this complexify the embedding approach as this has to be done a in cross-platform way in Abiword.
Concerning the presentation program, hp you are a little bit too enthusiastic. Criawips, the project shortly mentionned in the Abiword 2.2 roadmap is no longer under this form. The original plan as we discussed about when at GUADEC4, was to use Abiword framework and make it the second brick of AbiSuite, but finally herzi's approach moved and he did everything by himself, outside of our code base. GNOME Office is in desperated need of a good presentation program, and that is probably the bigges point. GNOME presentation program have a real dark history.
(update) after posting this, I found uwog blog (he is one of the active Abiword developers) and must say that I solely agree about what he said. The meta project nature indeed did not help. That is why we released 1.0, one year after we talked about that at GUADEC3.