LibreOffice is a wonderful, welcome, huge step forward for what was People are complaining about the name and stuff, but really, that stuff doesn’t matter: what does matter is that this now unfetters developers to do anything from the firing of drive-by patches to more fundamental work, and get the project going at the speed it deserves.

Lots of people have talked about the direction the project ought to be going in; I fear to some extent many of those people think that is a better app than it actually is. I’m a heavy user of it, and there’s a lot of it which is really hard to recommend to people. As an example, I’ve been working on a document that’s some sixty pages long over the past week: every time it saves, it takes an absolute age, and the entire suite – including spreadsheets or any other part of it you have open – just locks up entirely while that happens. And because it auto-saves every now and then (a necessity, sadly, because it’s not rock-solid) it means every ten minutes there’s virtually a minute of down-time while it saves the various files. And yes, this was an OpenDocument file – which should be the best performing backend (though I haven’t checked if this is actually true!).

However, with all the faults the suite has, it is still extremely valuable, and root-and-branch fundamental change to the codebase is exactly what it doesn’t need right now. I can’t see a Firefox-alike style project doing much other than throwing away what makes it so worthwhile: of course, everyone could name features that could be junked (“Who uses StarBasic?!”), but for every person who thinks a given feature is basically worthless there are likely to be at least two people for whom it’s crucial.

For what it’s worth, I think these points are the most important, though:

  1. OOXML support. Yes, OpenDocument is the native format, and should get the most love. But the output from Microsoft Office should come a close second. Let’s not kid ourselves: interop with Office is the key reason people can use LibreOffice at all. This feature is absolutely, 100% crucial for large numbers of users.

    Support is already in there, of course, but it needs to be better. There’s no technical reason LibreOffice can’t support these files as well as Office.

  2. User interface. It’s not as bad as people make out, for the most part, but some bits of it are clearly stinky sewers. Setting custom colours being the obvious pain-point which hits me every time. There are key bits which need significant work.

Other stuff, like resource usage, speed, etc. will come naturally with time one hopes, and in most ways the suite is basically usable now. New features are also definitely required – online collaboration being something which comes immediately to mind – but with an easier route into the codebase which doesn’t require libraries of specifications to be filed before work begins, it’s much more likely that people will commit to starting such features. 3.3 is a pretty lame release feature-wise – for most users, they won’t identify anything new beyond the “fast find” feature (which on its own is hardly rocket science) – and in a sense, it really doesn’t matter what features LibreOffice starts to deliver. It just needs to deliver new features consistently.

This has, of course, highlighted Oracle’s up-to-now lack of participation in the community. Presumably the web-based Oracle Office will be a tremendously exciting product, but it’s a totally different track. This also puts IBM at something of a cross-roads: itself appears to be on life-support (and potentially dead, depending on what Oracle decide to do). It’s conceivable that they could ship a version of Symphony on top of LibreOffice, although one would suspect it would be harder than what they do now. I hope they take the plunge and donate large parts of Symphony to LibreOffice – though I find it difficult to believe that they might.

In the same way that big business like Oracle and IBM ought to step up to the place, I think it’s also incumbent on pretty much everyone using LibreOffice to contribute. Maybe that contribution could be in the way of templates, graphics, documentation improvements and other non-code stuff. Hopefully, more excitingly, it will be a lot more straightforward to download LibreOffice, compile the bits you’re interested in and make code changes that can be sent upstream. I suspect for most people this will, initially, be little more than correcting obvious faults and one-liners, but it’s certainly something I’m going to have a bash at again and hopefully the success of the project will encourage many other people to do that too.