Affirmative action for open source

Should a public agency use privately developed software when equally suitable open source alternatives exist? Maybe a better question is, should politicians set software evaluation guidelines for IT managers, or should software compete on its own merits? That’s the question at issue in two US state legislatures this year, as well as in various nations around the world.

Last month Oregon, with House Bill 2892, and Texas with Senate Bill 1579, proposed requiring state agencies to consider open source products when purchasing computer software and to “provide justification whenever a proprietary software product is acquired rather than open source software.”

Neither of these proposals attempts to regulate private companies. They apply only to government agencies, which of course are funded by citizens’ tax dollars.

There’s nothing wrong with asking states to consider open source software as well as proprietary. Every organization ought to consider all the products that might meet their needs. I’m a little wary of provisions that force written justification when purchasing one particular category of software over another. That seems like unnecessary bureaucratic red tape if we presume that our public employees are trying to do the best possible job.

In California, advocates are trying to go a step further. Attorney Walt Pennington last summer proposed a Digital Software Security Act that would force state agencies to buy software only from companies that do not place restrictions on use or access to source code. To me, that’s a step too far. Forcing any organization to buy only open source products seems overly restrictive at best and foolish at worst; what if there is no comparable open source program?

The Texas bill offers some good guidelines for any organization trying to evaluate which software provides the best value. Its criteria include purchase price, compatibility with existing data and systems, scalability, reliability, training costs, support costs, and compliance with organizational standards and policies. The product that does best in most of these categories is the one that deserves to be implemented.

Consider this quote:

We don’t believe that governments should pick winners and losers. Technology should compete on its merits in a free market. Let the government look at all the options and then make a decision.

That sounds right to me. Do you agree? But when I tell you the statement is from Microsoft’s Ricardo Adame, quoted in News.com, it sounds more than a little self-serving.

Microsoft, as one of the largest makers of proprietary software, has a vested interest in countering open source software. Developing nations, where money can be tight, naturally find open source software attractive. In countries such as Peru, India, and Mexico, Microsoft has made substantial donations of cash and software to counter the natural inclination toward free and open source software. It’s hard for these governments to turn down hundreds of thousands of dollars in free money, but sometimes you have to make sacrifices in the short term to gain in the long term. Specifically, any package that IT managers can prove a product does a better job at a lower total cost over its lifetime should get purchasing approval.

Microsoft is attempting to ensure its future by donating software to schools. It knows that when students become comfortable with software, they’ll tend to prefer it when they enter the workplace. The company tried taking the same tack in the US. In 2001 it offered to provide a billion dollars’ worth of software, hardware, training, and support to 16,000 poor US schools as part of a proposed settlement of the government’s antitrust case. The government rejected the proposal, realizing that the company’s largesse would only help increase its market dominance.

Government mandates on social issues are always tricky. Take affirmative action, for instance. Do you approve of hiring policies to remedy historical racial inequalities, or should each potential employee be evaluated on his or her own merits? Consider your feelings about open source legislation in light of your feelings about affirmative action. You don’t have to be consistent, but if you’re not, you should understand why you feel as you do.

Do legislative attempts to level the playing field for open source really serve the open source community? Maybe, in that they may lead to more governmental adoption of open source software. But it’s a little inconsistent to find open source proponents trying to pursue their ends by advocating governmental coercion.

Linux not ready for the desktop? Give me a break!

I’m tired of reading on an almost monthly basis articles asserting that Linux isn’t yet ready for the desktop. Nonsense! Linux is about as ready for the desktop as Windows is. It’s simply a matter of corporate and user inertia that’s keeping Linux marginalized.

There are really two markets for desktop Linux. One is individual users. I’m ignoring them here to address IT departments. If you’re a Linux user outside of an organization, you probably face many of the same issues as corporate users, but you can (and must) deal with them yourself, and therefore decide for yourself whether Linux or Windows or MacOS suits your needs best.

The other market for desktop Linux is business organizations. Their OS needs are well-known: reliability, security, compatibility with existing data and devices, and the availability of applications they need to accomplish their business objectives.

To compete on the desktop, Linux must score at least as well on most of those criteria as the other choices, and it does. It’s at least as reliable and secure as Windows XP. Microsoft might have the edge in compatibility with existing data and peripherals, but Linux has the edge in compatibility with existing clients — you may have to buy a whole new PC to upgrade from Windows 98 or 2000 Professional to XP.

Application availability can be a killer. While there’s now a greater range of open source applications for users, the number still doesn’t stack up to what’s available for Windows. But more important than what’s available in shrinkwrap is what’s been written for special purposes. While simple Visual Basic for Applications macros may run in OpenOffice, the highly customized ones that many businesses rely on probably won’t. The same is true of applications programmed in a high-level languages. Virtually any such program will require more than a simple recompile of the source code to work under a new operating system.

There’s one other key component that I thought was hindering general deployment of Linux in large organizations — lack of a good desktop management client for server-based network management frameworks. Applications such as LANDesk Software’s LANDesk Management Suite, Veritas Desktop Management Suite, Novell ZENworks for Desktops, Computer Associates Unicenter, Tivoli, and Novadigm Radia Management Suite are vital to keeping costs under control when managing large numbers of clients. These programs must be able to perform hardware and software inventory, software distribution, remote control, and desktop policy management.

As soon as I did a little research, however, I learned I was wrong. According to a recent buyer’s guide in Network World, leading products like Unicenter, LANDesk, and Radia already support Linux clients. It seems reasonable to assume others will follow suit.

Yes, there are other issues for organizations that switch operating systems, including training for users and system administrators. You certainly have to figure those costs into the decision whether to change desktop operating systems. But that doesn’t negate the assertion that Linux is ready for desktop deployment now.

Of course just because Linux is ready for the desktop doesn’t mean users are ready for Linux. Users get very comfortable with their computers, and quite naturally fear change. But new systems come into production all the time as the needs of a business change, and when new client operating system better fit business needs, employees have to change with the times. The best way to assuage users’ fears is by a phased-in approach, migrating one department at a time (starting with the most technically savvy) and learning from each move. Also, don’t take away every Windows PC at once; leave one or two available in each department, if no longer officially supported, for those that might have legitimate need for them.

Any network manager facing a desktop upgrade decision ought to factor Linux into his calculations. I haven’t seen any hard figures (can anyone point me to any?) but it wouldn’t surprise me if it took just as much effort to upgrade Windows 98 clients to Windows XP as it would to migrate them to Linux. The initial expense is less, and the total cost of ownership on an ongoing basis is lower. What’s not to like?

Ellison an oracle for open source?

Larry Ellison is putting his money on open source development. Over the last couple of weeks headlines quoting Oracle Corp.’s chief
executive officer have included “Oracle
ups Linux efforts with ISV program” and “Linux
will wipe Microsoft out of data center.” What’s his rationale?

This is a shrewd move on the part of a man not known for making too many major mistakes in business. (Though I’m still scratching my head over his New Internet Computer Company, established to sell Linux-based “network computer” hardware. Maybe the fact that it’s still in business means he was right after all.)

Ostensibly Oracle has little to gain by backing one platform over another. Its eponymous database runs not only under Windows and Linux, but also a host (pun intended) of other operating systems, including the three major commercial variants of Unix. Since Oracle is operating-system-agnostic, Ellison would seem to have nothing to gain by backing Linux.

But if you view the market from the bottom up, things look different. The Linux database market is far less mature than the Windows server database market. Today there are several good choices, both open source and proprietary, for organizations that need a Linux-based DBMS. Ellison wants to grab mind and market share in an expanding market. That way, if his predictions pan out, the dollars he invests today will accrete to the bottom line for years to come.

Oracle isn’t trying to put MySQL and PostgreSQL out of business. It’s aiming its budget at the likes of IBM’s DB2 for Linux and Sybase‘s suite of Linux products — the commercial vendors Oracle has been going head-to-head with for years.

One of the strongest of Oracle’s competitors is SQL Server, marketed by Microsoft, a company that also sells a server and a desktop operating system. One way Oracle may be able to hurt SQL Server’s sales is to lower the sales and market penetration of the operating system SQL Server runs on. That’s another good reason for Oracle to provide incentives to Linux software developers.

What a great time to be a Linux software entrepreneur. According to IDG, “Oracle will provide two dollars for every dollar an ISV invests of its own money. It will also share its customer list with them to help generate sales leads and will allow them to market their Linux offerings through its Web site” and “provide access to migration tools for free.” It may not be raining money, but I’d have a bucket handy just in case.