OSNews

Ok, I think I finally understand what happened, thanks to the article.

1) Oracle Buys Sun.
2) Oracle kills Open Solaris.
3) Non Oracle Open Office devs get nervous, try to take over by creating an umbrella organization that would oversee all Open office development, but not have Oracle in charge.
4) Oracle just sees it as a fork, kicks everyone related to it out.

I'm not really sure about the last one though, either Oracle saw it as a fork, or they're just using that as a excuse to cover for the fact that they really don't want to willingly give up control over the project.

The graph on this page paints a picture: http://cedric.bosdonnat.free.fr/wordpress/?p=734 Guess what happened on Week 38.

What, a for-profit corporation interested in money? Holy shit, say it ain't so!?!?

Say it better: "What, the corporation lead by the 6th richest person in the world is interested in money?"

... And yet, RedHat is also -very- interested in making money, which doesn't prevent it from working (almost) exclusively on OSS projects.

You could claim that there's more money to be made out of making OO proprietary (I doubt it), but this doesn't negate the fact that external OO contributors had every right not to trust Oracle.

- Gilboa

Edited 2010-11-02 13:41 UTC

Out of curiosity, from where you got this number?

He's probably referring to the commit logs.
http://people.gnome.org/~michael/blog/ooo-commit-stats-2008.html

Of the 33 only a few actually contribute code. Digtizor reported this story incorrectly.

Only programmers would place more importance on code than anything else. 33 people is still 33 people.

More people doesn't mean more productivity, ask Greeks for that or anyother European country.

So are these people idiots to you or something? They all had their part to play in the success of OpenOffice.org. It’s highly disrespectful to just brush aside their hard work because they don’t all write lines of code. You would make an awful product manager with an attitude like that.

Greeks? What do the Greeks have to do with whatever point you are trying to make? I asked a Greek and he does not know what you are talking about.

When it comes to this type of open source project yes I would place more importance on the code.

OO has a massive, unwieldy C++ codebase that few developers would want to work on in their spare time. A project like FreeBSD offers more independence and satisfaction.

But I wasn't trying to dismiss the significance of the loss. From what I have read they were responsible for the German version. Given the size of OO that is like losing a floor of people. For any commercial ISV the German port is the most important.

Damn, Oracle has thousands of OOo developers?

Don't look now, but KOffice is about to split as well.

That is sad, but as there are groups with different priorities competing for different targets, it looks like it will be unavoidable. As long as the oars pick almost the same direction the boat will advance. It was not happening on KOffice lately, I guess.

More info on that: "" rel="nofollow">http://lists.kde.org/?l=koffice-devel&m=128812911619277&w=2"

The thread is long and looks like they do not agree on why they are splitting. Perhaps, there is still a change of reconciliation.

No, no chance at all. After five or six years of interaction problems arising from the personality problems of the author of that mail quoted above, a number of people in the KOffice community decided that enough is enough. He didn't want to leave, so a separation between him and the rest of the community was the only way out.

That's the only real issue, all the rest is spin.

Sad, but even I could see there was something wrong. KWord has seriously stagnated over the years and I always wondered what that libflake crap was about.

Well if LibreOffice picks up steam and becomes a better alternative than OpenOffice then it was a good decision to fork, but in order to do that they need to attract developers and this is the achilles heel as I see it. I just can't imagine that there are many developers out there who wants to program an office suite 'for fun' but I may very well be wrong. Also chances are that some companies will provide LibreOffice (doesn't exactly roll off the tongue, does it ) with backing (code, monetary).

I think you're not being fare. Comparing OpenOffice to GNumeric is not valid as GNumeric "is less than" OpenOffice (a spreadsheed application vs. a full-featured office suite). GNumeric, just like Abiword, is a "more stand-alone" program, allthough to be seen as part of Gnome's office suite (but quite independent of Gnome).

Still, I can partly agree with your statement. Modern software tends to get slower and slower, that's true. If you would compare today's OpenOffice (version 3) to an older version (version 1), you would see that the older versions often load faster even on older systems. If you then maybe compare more backwards to StarOffice... well, seems to be a lot faster.

The reason? Because there's much more functionality inside, like script processing engines, extensive input and output filters, and all the other litlle things - MANY little things - users are interested in. With the implementation based upon newer sets of libraries that need certain dependencies that need certain dependencies that rely on other dependencies... cont. ad. n... it's obvious that such a program CAN'T be fast unless you upgrade your hardware and software along with that program.

Another point I'd like to mention in this regards is the trend of modularization. For example, OpenOffice version 1 and 2 contained a german dictionary in the default localized install. Today, those doesn't seem to be part of the localized builds anymore. I think that's sad - spending hours on compliling and then still missing essential parts that have been present in the past.

I'm not sure Oracle is to be blamed here, as the trend I pointed out is visible across many (most?) software products. With today's more-than-enough resources on hardware, there seems to be no need for efficient programming. "Go buy a new PC!" seems to solve everything.

For example, if I compare the startup times of OpenOffice 2 on FreeBSD 5 on a 300 MHz P2 vs. OpenOffice 3 on FreeBSD 7 on a 2 GHz P4, I can see that the older system loads faster. On the other hand, it's not so important how fast a program loads - this is O(1) - but how well it performs when in use - that would be some kind of O(n), "scientifically" spoken. :-)

Coming back to your initial statement, I found myself more and more often using Abiword and GNumeric instead of OpenOffice because of speed issues. The only times I fire up OpenOffice if I reach the limits of the other programs (mostly related to importing nonstandard file formats).

OOo took the English dictionary out of the base install too, if I remember correctly. They reasoned the dictionaries should become a plugin since they are different between localizations, which would be fine if the automatic plugin installation process actually worked.

Modularization is a good thing; it makes it easy to create system which only contain necessary items. With that being said, the system also has to have an easy way to add those items back in when they are needed.

Removing the dictionaries from OOo wouldn't be so painful if the Plugin finder actually worked. It should let you find the plugin and install it without any more intervention then a click, Firefox does this well, but that's not the way it works.

I think you are missing the key factor. If there’s no commit access anymore, then nobody can replace those who have left. Oracle are locking down OOo so they can better profit from it.