Slashdot Mirror
Lack Of Developers Delays OpenOffice.org
bonch writes "OpenOffice .org contributors spoke this week at a conference in Canberra. Among other things, one of the issues raised was the lack of developer contributions and a source tree that is 'just too big.' Version 2.0 was originally going to be released around this time but will now be delayed until at least June or July."
I can vouch for this... the OO.org 2.0 beta has so far required non-RPM Linux users (or those who want to have a single-user installation of the beta version) to build from source--and I can say that it is a frightening endevour! (This coming from one who has built Gnome from source in the past and who is still daunted by the prospect of building OO.org)[1].
I was recently looking at open source projects that I might contribute to, and-- in my case at least --OO.org was counted out on the basis of build complexity. Cloudscape and other projects are where I've been putting my free time simply because becoming a 'casual contributor' to OO.org seems to be an unduly complicated process.
The solution? Simlplify the build process for the casual coder. This will have the added benefit of helping other Linux distros and UNIX versions more easily support bleeding edge OO.org. And on the development side, potential contributors of the odd functionality (as I would characterise myself) will not be scared off.
As I understand it (probably imperfectly), Linux has gone through the same growing pains in arriving at its current module architecture. I think this is a housekeeping step OpenOffice.org needs to dedicate resources to, and needs to dedicate them NOW, to clean things up to at least the level of the what the 1.x versions had where it was easy to compile[&|]install a single-user version, unlike the 2.Xs where it's a real workout.
----
[1] And this is also from one who also has no problem with contributing Java code despite the recently publicised Java issues in OO.org 2.0.
Yeah, I was just lookin' at their "solver". Pfft. All well and good for developers but the user's options are basically install-OpenOffice-or-not. It's not like GNOME doesn't have the exact same issues. The problem there is solved by the packagers who make apt and rpm dependancy trees. Of course, what's going to happen is that people are going to keep contributing to other projects and they will soon supercede anything OpenOffice has to offer simply because they are easier for people to get.
How we know is more important than what we know.
What, has the world suddnely become a bizzaro parallel universe where it's impossible to componentize frameworks or use shared libraries?
Somebody better make sure that the operating system developers realize that they are all living in violation of the laws of nature, and make them quit before they destroy the fabric of space-time.
I'd like to see the hardest stuff done in C or something else a bit (faster and) more debuggable than C++ and invoked from a Ruby shell.
Wait a minute - you think the solution to a lack of developers is to switch from a mainstream language to what is, if you'll forgive the intrusion of cruel reality, a language known only by a vanishingly small minority of coders, which has a reputation for being very slow (not good for desktop apps) and for not being able to do Unicode (not good for office apps)?
As I suspect you're the only person in the world who thinks rewriting the OOo framework in Ruby would be a good idea... well, it's down to you. You'd better get cracking, it's gonna take a while. I look forward to seeing the results: Ruby needs a big impressive project like this to lift it out of the "toy language" backwater it's stuck in.
It's good to hear X.org is stepping back and refactoring. I think all the major FOSS projects could use a breather, where new development is stalled for a few months while things like memory consumption are addressed, known performance bottlenecks are given a little more attention, etc.
GNOME, Mozilla, OO.org are all useful enough feature-wise, right now, that doing some serious polishing work would take them to the next level against Microsoft. I find GNOME, for example, to be very adequate in most every way except performance.
-- Microsoft is the most expensive commodity operating system and office suite vendor in the marketplace.
They certainly have a worse public image than IBM, and I wonder if people use OO because its free, but don't really feel part of the project because the SUN associtaions.
It seems to me that IBM and others (like Oracle), aren't playing nicely with SUN so much.Which is all a bit of a pity, because OpenOffice is the single main application that advances Linux as a useful OS on the desktop.
Even if all 10^7 lines of code could be thoroughly stabilitized today, it could get completely broken tomorrow because MS would come out with a new version of .doc.
.DOC format
Except that MS cannot do that.
If Word 2005 had a brand-new format, they would lose their biggest selling point to have everyone who uses Word 2000, XP, or 2003--the constant
And MS *did* introduce a new format in 2003. WordML, an ugly XML format, that OOo *already* has an importer for. MS breaking DOC isn't just an urban legend--it's a straw-man argument akin to saying that MS might move to an all-new language for their UI, and everyone who speaks English will be out of luck.
I can't comment on the complexity of OOo, but since there isn't an OSS alternative that even gets the Word-Processing part in the same leauge, I'd say that an argument that "OOo is big because of DOC" is a bit premaure. Maybe they're just big because doing all of the things that folk expect a word processor to do requires a LOT of code.
Also, refactoring isn't an either/or, it's both. Nothing precludes you from refactoring while you add features. In fact, Martin Fowler (who wrote "Refactoring") says:
As for "rewrite vs. refactor", that might be true for relatively small changes, but when you're talking about a large project like OOo, you don't want to throw the baby out with the bath water.
My father is a blogger.
They're using Java, in part, because it makes it easier to write the code.
Well, that's a good reason. But, in other hand, if you look the communities of developers in the FLOSS world, you would see that very few really accept the terms in Java licence. So, by using it, they make easier to write the application, but turn some developers around.
Easier to write or more developers: pick one.
(Note: I'm not saying that FLOSS developers hate easier to write languages, they are more interested in the licence of their works).