Off the Beat: Bruce Byfield's Blog

Kodak's recent announcement about the release of ScanTWAIN, a desktop application for scanners, was quickly picked up by the free and open source software (FOSS) media. After all, here was another large company now supporting the cause. Moreover, Kodak was doing so with TWAIN, the standard protocol on most operating systems, rather than just SANE, the FOSS protocol developed for scanners when TWAIN support for free operating systems was unavailable. The announcement seemed too good to be true -- and, unfortunately it was. Instead of being a minor cause for celebration, it raises the perennial question: When will traditional companies get serious about porting their products to FOSS?

Hints of a problem
The first hint that something was wrong is the note in the announcement that ScanTWAIN requires the installation of Mono, the GNU/Linux version of .NET -- not just because Mono is a technology that many people worry could be the cause of patent infringement cases, but because it suggests that the application might be a quick port from Windows.

The second hint arrives when you try to install the program. The .RPM package for Fedora requires an unavailable library. If you uncompress the source code instead, you find it in a directory labeled "trunk" with not even standard install instructions, as if the code was hastily copied from a repository without any effort being made to conform to standard source code conventions. This impression becomes even stronger when you compile and find that the code installs to such non-standard places as /usr/local/lib, and /opt, directories traditionally reserved for applications for the local administrator.

But the largest hint -- contained in the announcement if you read carefully, and obvious as soon as you install -- is the fact that, while ScanTWAIN might be released under the second version of the GNU General Public License, its backend isn't really a backend, and you still need Kodak's proprietary drivers. The supposed SANE driver is little more than a pointer to the proprietary TWAIN drivers. In other words, the FOSS involved is not an independent application at all.

You are left to wonder: Are Kodak's programmers inept, or is ScanTWAIN intended more as publicity than as a serious piece of free software? When you consider that the ScanTWAIN project was only added to SourceForge a few weeks ago, the idea that it is only publicity seems especially plausible. ScanTWAIN was probably developed in-house at Kodak, and is only a FOSS project by virtue of its license -- and in no other way.

At best, ScanTWAIN is comparable to projects like Adobe AIR, which flirt with the idea of FOSS, but contribute little to it directly. But, at worst, you might be forgiven for wondering whether, like so many business schemes in the Dot-Com era, it is based on the idea that the FOSS community are naive amateurs who are easily deceived. Whether this suspicion is actually true is irrelevant -- just the fact that it arises at all means that, by creating false expectations, Kodak may have done itself more harm than good with ScanTWAIN.

The same old story
You can understand why companies want to appear FOSS-friendly these days. Unlike most technology sectors, FOSS has momentum. It has been gaining in popularity steadily for the last decade, and, with the current recession, it appears more popular than ever.

Yet for all the desire to cuddle up, traditional companies are often reluctant to make any commitment to FOSS. Companies that sell software, of course, worry that they won't have a market if they make proprietary versions of their software for operating systems like GNU/Linux. Yet even companies like Kodak, whose business is hardware and that post drivers free for the download, seem reluctant to take the last step and use free licenses. They act as though they're afraid of that last step, as though still worried that genuine participation in FOSS might contaminate everything else that they do.

What makes the hesitation of traditional companies so ironic is that many FOSS users have an imperfect idea of software freedom and are begging them to port their products. For instance, three years ago, when Novell surveyed users to find what Windows applications they wanted to see on GNU/Linux, the most in-demand application was Adobe PhotoShop. Never mind that the GIMP is a solid alternative for working professionals (no matter what detractors say) -- the lure of name-brand software was apparently stronger than any ideals, or even the wish to receive software at no cost.
Yet even such invitations aren't enough to encourage traditional companies to get involved. For the last decade or more, they've been hesitating -- and, while they do, their opportunities are lost.

Bit by bit, the FOSS community has been building its own alternatives. Microsoft won't port MS Office, and Adobe abandoned a GNU/Linux version of FrameMaker after a half-hearted beta nine years ago, so FOSS has OpenOffice.org to replace both. Xara announced FOSS ambitions, but held back key libraries, leaving the field to Inkscape and KOffice's Karbon 14. TWAIN support for free operating systems was lacking, so SANE came along, doing TWAIN one better by separating the user interface from the device driver and making efforts like ScanTWAIN functionally irrelevant. In case after case, by refusing to enter the market, many traditional companies have given FOSS alternatives the space to develop their own alternatives.

Today, these FOSS alternatives are equal or superior to their proprietary counterparts. Where they are lacking, frequently they are still adequate for most users and evolving quickly. By the time the majority of traditional companies finally muster the courage to enter the market (if they ever do), they are going to find some serious competition. Basically, the FOSS market remains a lost opportunity for them.

Of course, as a free software advocate, I am content to have things this way. Had traditional software companies showed any courage in opening new markets, they might have delayed or blocked the maturity of FOSS for years. All the same, when I see traditional companies making the same mistakes about FOSS that they did ten years ago, I still have to shake my head in amazement. What, I wonder, will it take for traditional companies to understand FOSS? Or are they simply incapable of getting it?

Comments