gnotebook: The Desktop War: A Separate Peace
By: Michael Hall
Friday, June 1, 2001 08:21:59 AM EST
URL: http://www.linuxplanet.com/linuxplanet/opinions/3428/1/
Killing Emacs, Progeny Gives the Gift of GNOME 1.4
By way of a quick bit of GNOME scoop before launching into the column today, Progeny Debian GNU/Linux users will be happy to know that apt-get update && apt-get upgrade will net them GNOME 1.4 today. The distribution's made an update available that provides a current GNOME desktop and Nautilus 1.03 (with the very nifty newsfeed sidebar). It also means Progeny users can enjoy a current GNOME without going through the contortions they might have had to involving Ximian's release, the notable missing elements being the GNOME doorkeeper, Monkeytalk, and Red Carpet, all of which are specific to Ximian GNOME. That noted, onward.
Killing Emacs
I'd like a quick and easy way to record Emacs' "uptime." I managed to crash it for the first time in a decade a few weeks ago and it was, frankly, a stunning experience vaguely akin to being told Bozo the clown secretly hated children or learning Bill Clinton was actually a Republican sleeper agent. That left me also wondering how many hours of uninterrupted productivity came before this particular aberration.
A little background helps:
Since last December, I've been running GNU/Emacs 21. Most distributions are shipping with version 20.7, and the jump forward represented by the newest (and still unreleased) version is good enough that I've become hooked and I've taken to following along with Takuo Kitame's experimental packages, which are built around Debian's unstable/Sid branch. Thanks to the magic of Debian and its children, and the availability of a source archive from Kitame, it's pretty easy to to do this, too. In fact, here's a mini-tutorial, the steps of which I've used on Debian 2.2 (Potato) and Progeny 1.0:
- log in as root
- add this line to /etc/apt/sources.list:
deb-src http://people.debian.org/~kitame/emacs21 ./ - run apt-get update
- make sure you're in a directory on a partition with plenty of space (ie, not your root directory if / happens to be smallish
- run apt-get source --build emacs21
- install the resulting .deb's once the build is through
For the most part, the resulting packages are very solid. I didn't have any trouble until I decided to track down how to get Emacs' browsing package, w3, working under Emacs 21. A little googling later, I located a patch to apply to the w3 source that took care of that as well. As much as it was nice to once again flirt with the sort of absolutism Emacs encourages, it wasn't so nice when a pre-release running on top of a pre-release eventually ran across a bit of agressively nasty web design and locked Emacs hard.
I can't blame Emacs, really. In fact, in a world of uncertainty, Emacs is one of the few absolutes I'm willing to point to in terms of reliability. My recent crash represented the first time in over ten years of use that I managed to figure out a way to take it down, and I think it's safe to blame my own insistence on riding the cutting edge with it.
Regardless of where blame is to be assigned, it instilled in me a sudden and healthy loathing for the urge to live on the bleeding edge where my daily work tools were involved.
A Penguin's Progress
In a lot of ways, Emacs (and my own ignorance at the time) was the killer app when it came time to shuck Windows. In 1995 I bought a clone, the space issues involved with keeping it in a crowded Fort Bragg barracks be damned, and it represented the first chance I'd had to run Linux at all. Remembering pleasant days on an Ultrix machine years earlier and dismayed at how hard Windows 95 could be on an AMD 5x86 with 16MB of RAM, I decided to forego the pleasures of shrinkwrap and out-of-the-box hardware support and throw my lot in with Linux: it was a way to get a Unixy workstation and return to Emacs. Had I known about the Win32 Emacs ports, I might not have been so eager. As much as I loved the tools Unix (in the generic) provided, the one thing I remembered with genuine adoration and longing was Emacs, which I'd experienced over an ADM3a+ and VT100's years before. I knew little enough about Linux, in fact, that when I discovered during the initial install that X Window came with it I was pleased with the "extra feature." The early versions of the GIMP were absolute nirvana considering how much fun I'd had with Paintshop Pro under Windows.
I learned my way around fvwm, struggled through the hassles of mastering setserial and ppp in making a recalcitrant modem work, got sendmail configured to not send all my system logs to root at my ISP, figured out plug-and-play support enough to get my SB16 PNP working, and generally built up the sort of sweat equity you build. Life was pretty good, and I was well entrenched by the time the real explosion of interest in Free/open source software happened. Looking back on my days as an Ultrix user at a university years earlier, I decided I was about as well off as I had been then.
That mini-history wasn't meant to serve as some sort of breast-beating. By the time I got around to Linux, the bar to entry was low enough for the moderately motivated and curious. The local Books-a-Million had a few books on installation and configuration (I work for the author of one of them, actually) and while Linux wasn't as "nice" as it is now, it was manageable. Rather, it's meant to set the stage for where my own expectations are anchored when it comes to my Linux desktop. In many ways, the minimal configuration I had to deal with in 1995 remains "good enough" for what I do.
In fact, a sure recipe for frustration of late has been spending too much time trying to cram the desktop paradigm around tools I've been using for a while, or attempting to "remain current" on the latest applications for its own sake. It makes more sense to use established procmail recipes (and the mindblowingly effective spambouncer) with a variety of context-appropriate clients than it is to introduce IMAP as a way to let Evolution benefit from the sweat equity I built up learning procmail when there wasn't a mail client under Linux that had decent filtering built in. It makes more sense to remain productive, frankly, than it does to remain "current."
Accompanying the launch of the two major desktop projects several years ago was the notion that "the desktop" was the next logical target. To that end, we saw installation and configuration become magnitudes easier on the premise that a hurdle Linux needed to clear was how effectively available it was. The desktops were to extend that effective availability further.
It's clear that they've already achieved a lot of success. Linux installation and daily use is now, for the most part, truly simple. It gets easier on a daily basis, too. The landscape has changed to the point that where Red Hat was derided not three or four years ago as a "newbie" distro, you can now scan talkbacks and see readers identifying it as a "serious server distribution," while Mandrake typically receives the nod for a good beginner distro. In fact, I'd go so far as to argue that everybody who wants to try Linux probably has. They certainly have the means, the out-of-the-box tools made available courtesy of the usability consciousness raising the desktop campaign engaged in, and a Linux book market so saturated with assorted "dummy," "idiot," and "fast-n-easy" titles you're reduced to just making sure someone isn't trying to push a two-year-old release of the included distro before collapsing in the face of the incredible selection.
What we're left with is, in fact, the problem with applications availability and some attendant contributing elements: the widespread perception that the Linux desktop community doesn't want to pay, remains hostile to close-source software, and is tardy to embrace the sort of standards mainstream commercial software producers want before they'll venture into the Linux market. I know as well as the average Linux advocate that some of these points are being addressed, and I understand the imperfection of some of the arguments against Linux that point to these concerns. On the other hand, I've decided I'm overwhelmingly apathetic when it comes to whether my neighbor embraces Linux for his day-to-day work. In fact, in terms of Linux's advancement, my hopes largely center on the issues of open standards and a platform-agnostic Internet.
Closing the gnotebook
I've read a lot of assertions that without the desktop market, those hopes are in trouble, and that given a generally homogenized client space there's little chance Microsoft will remain in check when it comes to the slow crush of "embrace and extend." This is a forward-thinking argument that's a reasonable one to make. I once bought it, in fact, to the extent that I spent a little time helping out as best a non-coder can. Over the past few years, as I said in my last column, things have gotten better and better on this front, enough so that arguments regarding Linux's genetic inability to ever serve Joe Enduser well can be written off: it's obvious that the building blocks are present, and the potential to take things home is there. I like to believe, though, that even without a major share of the desktop market, the Free Software community possesses the will and talent to keep computing open and free. Apache certainly provides an example of this strength, as do several other well-established and vibrant projects. Given the ability to participate in the 'net community with any open-standards-compliant software, people can have as much choice as they want, even if that means making tradeoffs that come at the expense of the Linux desktop's market share. If I have faith in anything where Linux and Free Software in general are concerned, its in their ability to assure this future. Thinking so, the war for the desktop becomes more of a skirmish for a hill in a broader battle that's being fought in the valleys, out of sight of end users but every bit as critical to the campaign.
So, limiting writing about the Linux world to what's happening in GNOME is ignoring my own happiness with a rich variety of tools that aren't tied to the rubric of a unified desktop and my own curiosity with the wider world of computing. Consequently, it's time to close down gnotebook. Though GNOME will be a project I'll continue to follow with interest, and report on when appropriate, I hope to diversify a little over the coming months, out from under the limitations reporting solely on a single project can impose.
Beyond that, I hope to focus on the issues surrounding open and free standards as they play out in broader contexts. These things have a direct impact on the Linux desktop, even if they aren't as easy to grasp or as apparently glamorous as the rapid improvements in user interface and ease of use we continue to see coming out of GNOME and KDE. It's all still about a future where computing remains free and open.