Michael Meeks: "OO.o is a profoundly sick project"

[RGB]

With the lowest number of active developers on OOo since records began (24) it seems that the project have severe problems. Read the whole Michael Meeks analisys here:
http://www.gnome.org/~michael/blog/ooo- ... -2008.html
What do you think?

[Villeroy]

interestingly, a good third of contribution to Linux comes from external (or un-affiliated) developers, but the rest comes from corporates. What is stopping corporations investing similarly in OO.o ?

The Linux kernel turns a device from an amorph mass of silicone, copper and plastic into something useful. It can turn on the geeks so they work hard to get everything out of the device and learn more about microelectronics.
OOo can help to get rid of the tight bondage between certain file formats and a particular software vendor. But the whole paper-analogy, feeding "documents" into "desktop" applications with menues and buttons is not sexy. It's soo 90ies, still assuming an untrained, or even computer illiterate user. I think there is a broad gap between the "nature" of the software itself on one side and the personalities with the required skills on the other side. Money can fill this kind of gap, which may be the reason why most of the code comes from a small group of paid developers. The "true geek" (TM) is not engaged with another version of a virtual type writer.

[bobban]

Very interesting article RGB. I had no idea things were this bad in the OOo dev community. I wonder if this project would grind to a halt without Sun's heavy involvement, or would the release of that control open the gates for more independent developers to enter?

I am fairly newbie programmer myself and am just trying recently to make a few hacks. I found that just getting the code to build was a challenging task, mainly because the wiki has too much information, a small part of which is relevant and a lot of which is outdated. I found one good guide and a heaps of others that had snippets of useful stuff. I have found less information on how to work most productively with the source, and right now I am trying to figure my procedures out. I would think that having good manuals on this stuff would be a good start to making it an attractive project for new independent developers.

At least I have had some good responses from some Sun developers on the mailing lists to assist me.

[TheGurkha]

The "true geek" (TM) is not engaged with another version of a virtual type writer.

Based on nothing but my gut feel, I would say I think this is true. The true geek isn't busy working on some other word-processor or office suite, they're probably not atracted to such a project in the first place. But that might be because other things appear sexier to the true geek, as Villeroy says, or it could be that they mistakenly think OOo is already well catered for with many developers and they don't realise that they could make a big difference to the project.

[acknak]

That there are few programmers outside of Sun participating, or that OOo is a drain on Sun's already-strained finances, is neither new nor disputed. Whether that makes OOo a "profoundly sick" project is just one person's interpretation of those facts.

I suspect that the project fails to attract coders not because of the nature of the software, but for two other reasons:

One, because of the details of it's implementation. For a potential volunteer coder, OOo is difficult to build--as bobban points out, difficult to code for, poorly documented, and in many cases, ancient and arcane. In other words, it's not fun.

And two, because Sun maintains a level of control over the project that means they make all the interesting decisions, and those decisions are made for the benefit of the overall OOo product, not "to make the best software we know how to make". That situation is a severe turn-off for would-be volunteers to contribute something important to OOo, but it is also necessary for maintaining the current quality and effectiveness of OOo. Would it make sense to stagnate OOo's development for who knows how long, while volunteers re-write it using better and more "interesting" technologies?

Sun is slowly replacing the worst parts of the OOo infrastructure (the replacement of the graphics layer should be coming this year) and I think that's the right approach even though the pace is maddening.

I would love to see an "OOo-TNG" project that would re-design the suite from the ground up, but I can't see that being very practical due to the large investment and lead time required to produce anything useful.

[bobharvey]

What do you think?

It's not the first time he has said something like this, and those of us who have been submitting bug reports for most of the decade know that the whole project is sclerotic, disorganised, and slow to innovate.

A couple of years ago the decision was taken to 'chase' M$ compatability over innovation. I argued strongly at the time that this amounted to letting your competitor design your product, and that we would only ever be in catch-up.

The biggest impact of OOo has been when they did innovate - I am thinking in partucular of ODF file formats.

I think the analysis we have seen is fair. The code, the management structure, and the decision making are all megalithic.

None of the 'integrated suites' on the market, including OOo, are truly integrated, in that the same code handles text layout in all applications. Or maths, or tabular layout, or printing.

There is a case for starting again, but who would want to put in that much effort? People have such huge expectations of what a wp or a spreadsheet should do now. Back in the 80s WordPerfect had vertical tabs and wordstar did not. Now we have some sort of consensus that page layout should be a sort of flowing stream based on font metrics. There is no concept of vertical measurement at all. Frame-based layout tools are called DTP and become a minority thing.

And the fearsome inaccessability of the macro languages - they are all a disgrace in this regard.

No OOo is sick, but so are its competitors, struggling under a weight of compatability issues and attempting to feature match each other instead of asking what the user wants. Does anyone remember Locoscript? The application that allowed Sinclair to sell CPM computers for 10 years after other people were going gui? There were plenty of parish magazines and club newsletters banged out on them with no-one from the IT industry noticing at all. (http://en.wikipedia.org/wiki/LocoScript).

Abiword is gaining ground with users, I hear.

[RGB]

I'm a sort of "compulsive open-source tester". Abiword is fine only if you need to do simple works, but very complex layouts (the kind of documents in which Writer shines) are simply out of the reach of that software.
But there is right now a project that is "reinventing everything": koffice.
Version 2 will have support for all mayor platforms (Windows, Linux, *BSD, Solaris, mac... and palmtops: some time ago I read in the dot a comment about koffice beta running on Maemo).
I looked at the last 2.0 beta 4. The whole package is light years behind OOo in key features (kword 2.0 will be released without table support...), but they built a very interesting base for the future. The "flake" concept is neat, and the UI design is fresh.
In fact, right now koffice beta have out-of-the-box support for otf fonts, antialiased vector graphs, multimedia for Linux (through kde's Phonon) and twin monitor presentations.
koffice 2.0 will not be ready for "general use" (nor 2.1, I think). At least, it will not be ready for me. But in a couple of years there will be an interesting "competition" in the open-source office software, I think.

[Aubéron]

About the number of developers of OOo :
http://eric.bachard.free.fr/news/2008/1 ... opers.html