In case you haven’t been keeping up with GNOME—a GPL-licensed desktop environment for Linux and other UNIX-like operating systems—project leader Miguel de Icaza recently quit his day job and co-founded Helix Code, Inc. The company is a Cambridge, MA-based start-up that’s attracted some venture capital, dozens of full-time programmers, and a good dose of media attention (including a recent write-up in the Wall Street Journal). Helix Code’s goal? Without compromising GNOME’s free software licensing, Helix Code plans to realize GNOME’s promise by debugging the existing code and adding new, sorely needed applications. (For example, the GNOME office suite now includes a GNOME version of AbiWord as well as the impressive, Excel-compatible Gnumeric spreadsheet.)

Thanks to an infusion of venture capital, Helix Code has hired a bevy of programmers, and they’ve been hard at work. The result is an impressive new Helix Code version of GNOME, available for download, but you can also order a CD if the prospect of spending eight hours downloading 93 packages doesn’t sound appealing.

GNOME 1.2 is enough to make journalists gush with enthusiasm, and that’s just what they’re doing. The Helix Code version of GNOME downloads and installs smoothly, it’s fast and stable, and it looks great on-screen. But let’s get real. Can GNOME develop into something that could bump Microsoft Windows off corporate desktops?

The answer is no—at least, not right away. As even GNOME’s most ardent supporters will readily concede, GNOME 1.2 isn’t going to lure many corporate Windows and Office users away from the Redmondian fold. The deeper question is whether GNOME’s development trajectory will eventually create a competitive product. As I’ll show in this Daily Drill Down, much depends on whether the GNOME development team continues to focus on what corporate users want and need from their software. And as you’ll see, they’re off to a great start—for the most part. What remains to be seen, however, is how well the GNOME development team will succeed in figuring out what desktop users really need from their software—and how they can create products that will strike corporate users as a genuine and welcome alternative to commercial software.

Easy installation and upgrades (what a concept!)
What do corporate desktop users want, anyway? Clearly, they want a hassle-free installation and upgrade experience. Linux and GNOME aren’t going to provide a credible alternative if users need a high level of technical expertise in order to install and upgrade their applications—period. If you’re used to spending hours or days compiling software, only to wind up with a nonfunctioning system and a raft of obscure error messages, you’ll flip when you try the Helix Code GNOME Installer.

Try it yourself, if you haven’t already. To install GNOME 1.2, visit the Helix Code Web site and follow the links to the download page. The exact procedure you’ll follow depends on which platform you’re running. Packages are available for the following platforms:

  • Caldera OpenLinux eDesktop 2.4
  • Debian GNU/Linux (Woody)
  • Linux Mandrake 6.1 and 7.0
  • LinuxPPC 2000
  • Red Hat Linux 6.0, 6.1, and 6.2
  • Solaris 2.7 on UltraSparc
  • SuSE 6.3
  • TurboLinux 6.0
  • Yellow Dog Linux Champion Server 1.2

Once you’ve selected your platform and clicked Go, you’ll see specific instructions. For Red Hat Linux, you’ll open a terminal window, switch to the root user, and launch Lynx (a text-mode browser that interfaces very nicely with the shell). Lynx access the go-gnome script, which downloads the GUI-based Helix Code installer. From there, the entire process is completely transparent. A wizard enables you to select the packages you want. The downloading begins, and when it’s done, GNOME is installed. It’s as simple as that!

Upgrading your GNOME software is equally easy, thanks to the Helix GNOME Update package. This package contacts the Helix Code distribution site (or one of its mirrors), checks the available packages against the ones installed on your system, and alerts you if there are new versions available. After you’ve chosen the upgrades you want, the whole process is—once again—completely transparent to the user; once the download is finished, your system is upgraded.

The GNOME office: Living with Goliath
As you’ve just seen, Helix Code is well on the way to creating the software installation and upgrading support that corporate desktop users need. But, what about applications? Here, the key word is compatibility. It doesn’t take a genius to figure out that GNOME office software isn’t going anywhere unless it’s compatible with Microsoft Office applications. But compatible is a word with several shades of meaning.

There’s file compatibility, of course, which—in its most simplistic sense—means the ability to read and write documents created by a Microsoft Office application, such as Word or Excel. But there’s another dimension to file compatibility that’s not often given the emphasis it deserves. In today’s corporations, almost all of your work is done collaboratively—and very often, you find yourself opening somebody else’s document, making some changes, and sending it back. Your Word-using collaborators aren’t going to be very happy with you when they discover that AbiWord filtered out all of their styles, footnotes, borders, tables, and index entries.

There’s also feature-set compatibility. To knock Word and Excel off the desktop, presumably, GNOME office software will need to offer users enough of the features of Word and Excel that they won’t miss—or notice—the ones that GNOME doesn’t include. By examining the centerpieces of the GNOME office suite, Gnumeric and AbiWord, you’ll see two contrasting philosophies of what the term “enough” really means. The first approach, reflected in the Gnumeric spreadsheet, concedes that Linux software isn’t going to be seen as a credible contender unless it offers a high percentage of the Microsoft Office features that users have come to expect. The second approach, reflected in AbiWord, denounces the “creeping featurism” that saddles users with an overwhelming number of glitzy features, most of which were inserted for marketing purposes. The AbiWord solution? Develop a minimalist set of features that users really need and don’t bog them down with needless complexity.

Which is the right approach? Let’s take a closer look.

Gnumeric: Can it take a slice out of Excel’s pie?
Miguel de Icaza’s Gnumeric spreadsheet, currently available in beta version 0.56 and distributed with the Helix Code GNOME 1.2 preview, is an ambitious effort: Gnumeric aims to replicate a large number of the feature sets found in the leading spreadsheet applications, such as Microsoft Excel.

In the file-compatibility department, Gnumeric does an impressive job of opening Excel 95, 97, and 2000 spreadsheets, and the program can save spreadsheets in the Excel 95 format. Users who need to exchange spreadsheets with Excel users will quickly discover Gnumeric’s major deficiency, the still-to-be-developed support for charts embedded in worksheets. If you use Gnumeric to open and save an Excel worksheet containing a chart, the chart is deleted—which isn’t going to make your Excel-using collaborator very happy, believe me. (“My chart! I spent hours on that chart! And I don’t have a backup!”)

The good news here is that Gnumeric chart support is coming. The back-end support has already been created, thanks to the Bonobo component architecture and other key technical developments in the GNOME project. When completed, this support will include all the chart types, and most of the chart features, that Excel currently offers.

What about features? How does Gnumeric compare to, say, Microsoft Excel? To provide a rough-and-ready answer to this question, I compared the list of options available on Microsoft Excel 2000’s menus to those currently implemented in the current beta version of Gnumeric (at this writing, 0.56). Table A (shown below) lists the results. This isn’t a perfect comparison method; many important features of both programs are buried in dialog boxes. (I should mention, too, that would-be Gnumeric users are well advised to compare the program’s built-in functions with Excel’s so that they can make sure needed functions are available.) Still, the comparison provides a useful guide to would-be users wondering just what Gnumeric has to offer.

Table A
Menu Missing options
File Save As Web Page
  Save Workspace
  Web Page Preview
  Print Area
  Send To
Edit Fill
  Move Or Copy Sheet
View Normal
  Page Break Preview
  Formula Bar
  Status Bar
  Header And Footer
  Custom Views
  Full Screen
Insert Chart
  Page Break
Format AutoFormat
  Conditional Formatting
Tools Spelling
  Share Workbook
  Track Changes
  Merge Workbooks
  Online Collaboration
Data Form
  Text To Columns
  Group And Outline
  PivotTable And PivotChart Report
  Get External Data
  Refresh Data
Window New Window
  Freeze Panes
Microsoft Excel 2000 Menu options missing in Gnumeric 0.56

So what’s missing from Gnumeric? As Table A shows, Gnumeric doesn’t yet include many of Excel’s menu options, including several that Microsoft is clearly using to try to bind Office-using shops to an all-Windows NT/2000 infrastructure. We can do without those features, thanks. But, there are still some serious omissions in the current version of Gnumeric, including charts, spell checking, a find and replace feature, spreadsheet auditing, and macros. Still, de Icaza claims that his goal, ultimately, is to emulate a large percentage of Excel’s feature set. When Gnumeric begins to take shape, it’s quite likely that a Gnumeric user will be able to open a complex, feature-filled Excel spreadsheet, view and modify it in Gnumeric, save it to Excel 95 format, and return it to an Excel-using collaborator—and without deleting important data or formatting.

What’s more, the completed Gnumeric will free users from Microsoft’s lock-in strategies—and that may turn out to be its greatest appeal. Economists use the term lock-in to describe technical designs that are intended to make users buy more of the same firm’s products. For example, Excel’s macro language locks you in to Microsoft Visual Basic; if you’re serious about developing spreadsheet applications with Excel you develop in Visual Basic. And, if you want to put your applications on the network, you’d better run Windows NT/2000, or else you’ll miss out on all sorts of great features. Get the idea? In contrast, Gnumeric has been designed from the get-go to work with a variety of programming and scripting languages—the choice is yours. If you’re doing a lot of developing but you’re also trying to minimize the amount of money that’s flowing to Redmond, Gnumeric is starting to look like a very appealing and credible alternative.

AbiWord: The minimalist approach
AbiWord, spearheaded by AbiSource, Inc., is a GPL-licensed, cross-platform word processing program. There’s a Helix Code version of AbiWord (version 0.7) included with the current GNOME 1.2 distribution, and it’s being billed as a component of the GNOME office suite—which is somewhat curious, because AbiWord’s development philosophy is in many respects the exact opposite of Gnumeric’s.

Unlike Gnumeric, which seeks to emulate a high percentage of Excel’s feature set, AbiWord isn’t about to knock off Microsoft Word in all its fulsome complexity. On the contrary, the AbiWord ethos castigates Microsoft and the various owners of WordPerfect for engaging in a crazy, pointless “feature war,” in which each attempted to outdo the other by adding complex, meaningless, and ultimately unfathomable features to their products. Users don’t want all those features, the AbiSource people say, and they only serve to confuse. AbiWord takes a minimalist, Zen approach, focusing on providing just those features that writers really need.

But, what about living with Goliath? AbiWord users need to exchange documents with Microsoft Word users. To support file exchange, AbiWord can read Word 6.0/95/97 documents, and the program can also write to the Rich Text Format (RTF), which Word can read. But there’s just one problem: In the same way that Gnumeric lops off unsupported features, such as charts, a two-way trip through AbiWord results in the destruction of anything that AbiWord doesn’t support. The trouble is, AbiWord supports so few Word features (see Table B for a menu-option comparison) that a very great deal of the original Word document will get axed at the hands of AbiWord’s filters.

AbiSource has a solution to this problem, of course, but I’m not sure it’s going to work. Unlike Gnumeric, AbiWord is designed from the get-go as a cross-platform product; code is developing simultaneously for Windows and Linux, and there’s talk of a Mac OS version. In the sweet by-and-by, we can foresee a beautiful world in which AbiWord users will be able to exchange documents across platforms using the native AbiWord file format—and to heck with Word compatibility.

Table B
Menu Missing options
File Save As Web Page
  Web Page Preview
  Page Setup
  Print Preview
  Send To
  Recently Saved Files
Edit Links
View Web Layout
  Print Layout
  Document Map
  Full Screen
Insert Page Numbers
  Index And Tables
  Text Box
Format Bullets And Numbering
  Borders And Shading
  Drop Cap
  Text Direction
  Change Case
Tools Grammar
  Word Count
  Track Changes
  Merge Documents
  Protect Document
  Online Collaboration
  Mail Merge
  Envelopes And Labels
  Letter Wizard
  Templates And Add-Ins
Tables All
Window Arrange All
Microsoft Word 2000 Menu Options Missing in AbiWord 0.7

There’s just one problem with this strategy: I’m not convinced it’s going to work, and it’s not because the AbiWord development effort is being diluted by the need to support more than one platform, as some AbiSource critics have maintained. Once again, the reason lies in recognizing the not-very-pleasant realities of Microsoft’s lock-in strategies—and in particular, Word’s Track Changes feature (also called revision marks), which enables co-authors to make insertions and deletions that can be approved or rejected by the document’s lead author. This feature is a classic example of technological lock-in; once you get people using Track Changes, everybody has to use Word, unless they can find another program that’s fully compatible with Word’s Track Changes feature.

Is Track Changes really such a big deal? Yes. In today’s corporations, almost all writing is collaborative—as many as 80 percent of all the documents currently written in organizational settings involve two or more authors. When it comes to support for collaborative writing, Word is the only game in town. With Office 2000 and Windows 2000, Microsoft is trying to migrate the lock-in to the server side by building proprietary networking protocols into the collaboration picture. Against forces like this, AbiWord won’t stand much of a chance.

Gnumeric points the way
So, which is the best approach, Gnumeric’s or AbiWord’s? I confess that I have a great deal of sympathy with the minimalist approach; I agree with the AbiSource folks that creeping featurism is a bad thing. However, as more people use computers and networks to work collaboratively, mere file compatibility alone isn’t sufficient; Linux users need to be able to exchange complex documents with Office users without having the Office users’ charts, footnotes, tables, and lists lopped off. In this respect, Gnumeric points the way. If de Icaza and the rest of the Gnumeric development team succeed in their ambitious goal of implementing a large proportion of Excel’s feature set, a Gnumeric user will be able to open and save Excel documents without lobotomizing the files in the process. That’s not the case with AbiWord; running a Word document through AbiWord will remove much of the Word formatting and will even trash some of the data, including text placed in Word footnotes. Minimalism sounds nice, but the Goliath-dominated reality requires a rich feature set so that GNOME users will be able to exchange complex documents with Excel and Word users.

Gnumeric points the way in another respect: It offers some freedom from lock-in. Increasingly, managers and users are quite aware of Microsoft’s lock-in strategies—and they don’t like it one bit. Excel presents organizations with an unbelievable thicket of lock-in ploys, which are designed to ever so subtly (and sometimes not so subtly) discourage you from using standards-conformant databases, Web servers, collaboration utilities, and programming languages. By offering users a hassle-free way to select an application development language other than Visual Basic, Gnumeric provides organizations with a way out of the lock-in maze—and as Microsoft’s Windows 2000 strategy unfolds, that’s going to prove increasingly appealing to Microsoft’s customers.

So, where should AbiWord go? To offer a genuinely attractive alternative to Office-dominated desktops, GNOME needs a word processing program with a much richer feature set. At the minimum, AbiWord needs to implement enough of Word’s feature set so that the program doesn’t mangle Word documents. Just as important, AbiWord’s developers need to analyze Word’s lock-in strategies and develop a credible alternative, beginning with a Word-compatible track-changes feature that will free organizations from the need to force all of their users to stick to Word and Word only. What’s more, AbiWord could offer a network-capable, standards-conformant infrastructure for collaborative writing that doesn’t force organizations into an all-NT/2000 mode—and if it does, the GNOME office suite might very well become a credible contender for the corporate desktop.
The authors and editors have taken care in preparation of the content contained herein but make no expressed or implied warranty of any kind and assume no responsibility for errors or omissions. No liability is assumed for any damages. Always have a verified backup before making any changes.