I am a daily user of OpenOffice, and I mean it literally. There is hardly a day when I don't use it for something and usually I have it opened at all times, working on one document or the other. I have been a daily user of OOo since the late 1.x versions so I've seen versions come and go with key features remaining unimplemented and ignored.
Here is, for example, a list of most wanted OOo features, maintained by the project itself. Note how many of the highly desired and literally easy to implement features are dated years ago, some even before OpenOffice has been called that! All these years users have gone unheard.
Here's a list of features I consider imporant based on my experiences with OOo and the above list (in order of importance):
- Support for MS Office XML file formats. While ODF has been a great idea and I wish it happened earlier and ended up on top, I think it is time to admit that it has failed in practice and the dominant Office formats are .docx, .xlsx and other OOXML formats. Please bite the sour apple and implement full 1:1 support for OOXML formats, allowing document interchange between OpenOffice and the rest of the world. Please show that all that talk about open formats was not just empty rhetoric and that you value users above politics. The complexity of OOXML and the fact that it's in some points closely bound to MS Word, Excel, etc. behaviour cannot be an excuse since the same relation exists between ODF and OpenOffice (as is clearly demonstrated by there not being a single compliant implementation of ODF outside OOo, and even OOo uses plenty of extensions and "unspecified" behaviour").
- Support for SVG import and export. It is the year 2011, and by its importance SVG is a format right up there with HTML (for which btw also isn't a decent full-format export filter!). SVG images are becoming not only standard but also popular. Having a decent support for SVG would allow, among other things, to use Inkscape (which is a very nice tool) for drawing illustrations for documents!
- SQLIte for the main database engine. The only reason a non-ACID, memory-only database engine written in Java (HSQLDB) was imported in OOo was that Sun wanted to justify the requirement for Java in OOo which would, presumably, make it more enterprisey. See how that worked: OOo Base is one of the most hated parts of OOo. SQLite is all HSQLDB isn't: ACID, small and portable to everything from Android to mainframes. As such, it is ubiqutous and has became a defacto standard for small databases everywhere. The SQLite engine around 500 KiB in size *alltogether*!
- Java. Just throw it out. It was brought in for political reasons and has brought OOo only pain. Java as a scripting engine? Really, when Lua's interpreter is around 100 KiB in size. As a *requirement* for a *database engine*??? Really! That's ridiculous. Even nowadays I hear potential users (i.e. those who might use OOo but don't because of this and other reasons in this list) how they're avoiding OOo because "it's written in Java and it's slow", "it uses Java", etc. Seriously, just throw Java out and don't look back. It's a dark page in OOo's history and should be forgotten.
- Multi-monitor support. A non-brainer: just do what MS Office does (especially in PowerPoint) and be done with it.
- Finish the Bibliography tool. As a writer of scientific and technical documents, the Bibliography tool is simply one of the tools of the trade but I keep constantly running into its limitations: no import / export support (e.g. for BibTex), a funny table schema, hard or nonexistent ways to change / switch databases and absolutely *no* way to create new ones. And of course: no way to link in-document references to in-database references so that the database overrides document data (really, really needed as that's the point of having a database). The tool is not atrociously bad but is coming close to being so - it simply needs finishing.
I'm sure there are other bugs and wishes (I come across new ones almost every day,) but these I think are the key for making LibreOffice have any chance of being accepted by normal users.