.comment: A Dead End and a Milestone, or "What's Up, .doc?" - page 2
Documents Check In, But They Don't Check Out
But just as we cannot in good manners complain over the lack of a particular feature when we are using the fine programs of developers who work in large part for the sheer love of the art, we also cannot be certain that our heart's desire won't appear in the CVS tree tomorrow. The aspect of Linux that is at once both maddening and delightful is that it is constantly being developed, with the results instantly available. I build the KDE CVS tree every week or two, and as a result do not know if I ever actually had KDE-2.2. I know I had versions very close to the release version, and may have had the actual release for a week or so, but I'm not sure. I'm currently running a version that identifies itself as KDE-2.2.1pre.
And if the schedule holds, on Friday the KDE CVS tree's "head" branch will take a step unlike any it has taken in more than two years. On that day, the main branch of KDE development will officially switch to QT-3.0 (well, the current pre-release version of QT-3.0).
This is a Big Deal. It is the first externally obvious step toward KDE3.
I remember when, a couple of years ago, the switch from QT-1.44 to a beta of QT-2.0 came. There was no shortage of confusion as people discovered that they couldn't build the daily snapshots (CVS wasn't as widely used back then) using their existing QT, and they couldn't use their existing KDE-1.x if they upgraded to the QT-2.0 beta. Recipes for getting the two to run on the same machine were published.
Fortunately, I was in the middle of another book then, about KDE-1.x, so I didn't have the opportunity to play with KDE2, as it was then called, in the really early days. (For what it's worth, my superb editors, Gretchen Ganser and Hugh Vandivier, were extremely helpful in porting their template to StarOffice, so I didn't have to use the dreaded Windows for that book.) But once the author review circus had left town and my part of the book was completed, I instantly downloaded and built both the new QT and the beginnings of KDE2. A lot of stuff didn't work yet, that second week of November 1999, and a lot of stuff was broken. Sometimes a build seemed more like a step backwards, so keeping the most recent working version was a very good idea. But from the first day I built it, I never had to return to KDE-1.x. For a few months there it seemed almost as if using a system where everything worked would be boring and unexciting.
Then, gradually, almost imperceptibly, things began to work, and work well. It just got better and better. People produced amazing new features (some, like themes, seemed pretty pointless at the time, but today I would not like to be without Mosfet's "Liquid" theme which, though sadly no longer part of the KDE distribution, is still available from him). Sound support, pretty rough at first, got reliable. It's too bad that there's no software equivalent of time-lapse photography, for if there were, watching KDE2 grow from its beginning to where it is today, as we can watch a tomato grow from hole in the ground to ripe fruit, would be a delight.
And now it's about to begin again. There will still be versional confusion, and for awhile stuff will be unreliable, but it will probably be a lot smoother than it was last time. Many of these guys have been through this before. I just now dropped by the developers gallery on the KDE page and looked at the pictures, most taken years ago, of the audacious young developers who got together -- was it four years ago already? -- and decided to build the first fully integrated desktop for Linux and any other operating systems that cared to tag along. Looking at their mailing list posts today, they seem every bit as full of piss and vinegar as they were in the days before the release of KDE-1.0 in 1998.
What innovations will KDE3 bring to the desktop? It's impossible to know. QT-3.0 will have enhanced support for threading, and there is much room for exploration in that direction. People are already posting their wishlists for KDE3 on the KDE developers list. Certainly, many of KDE's legacy applications will be ported over, and it appears that the path to porting is fairly straightforward. But we return to that maddening, delightful Linux phenomenon, where marketing studies and careful plans for achieving market share take a back seat to those things that people, limited only by their imaginations and skills, want to write. KOffice will become more sophisticated. And, as with that first, pea-sized green June tomato, it will all always seem closer than it will turn out to be.
My work on this book ought to have gotten through the hateful author review by early November. I'll celebrate by downloading and building the head branch. A lot will be broken, and some stuff will be absent entirely. But somehow I doubt that I'll even look back.
And oh, yeah, I have a wishlist item: a WinWord export filter.