From the Publisher

Let's Take Linux Seriously

by Phil Hughes

Back in the early days of inexpensive Unix systems it was a serious uphill battle to sell Unix-based computer systems to small businesses. Why? Because MS-DOS did everything they wanted. By the time these businesspeople realized they had a problem—they needed a computer system that could be accessed by more than one person at a time— Novell was telling them they needed Netware to hook together their MS-DOS systems.

Well, Novell managed to sell these MS-DOS people another operating system. Most probably didn't realize it but they were Netware users with MS-DOS "terminals". It sold because they could keep on thinking they were running MS-DOS and it solved the problem.

I am sure some of you out there want to figure out how to start making money as a software developer or computer consultant. The general answer is "solve a problem". In the rest of this column I am going to show how you can use Linux to solve problems and make money along the way. Get rich quick? No. Guaranteed? No. But, if we all play it right it has a good chance of success.

In the past the hurdles to selling a complete system were:

• Getting the hardware
• Getting the software
• Putting together something that was needed
• Finding customers and selling it
• Supporting what you sell

Has something changed? Not in this list, just in what you need to do to jump these hurdles. Let's look at them one at a time.

In Seattle there are almost as many places that sell computers as sell gasoline. At one time I thought there would be more computer stores than expresso stands but now that dry cleaners, insurance agents and car dealers are selling espresso I would guess not. This means that it is pretty easy to buy a computer—even one that can run Linux. Add to this the fact that computer hardware is just plain inexpensive these days and you have another advantage over the old days when you had to sell your firstborn in order to afford to buy a computer.

If you are not a hardware whiz (or don't want to be) your best bet is to talk to the owner of a computer store. Not a big chain, just a local businessperson who would be interested in your story. Tell him about Linux, about yourself and about why he will benefit. What you want is for him to supply the hardware to your customer and be responsible for it when it breaks. He can make the profit on the hardware and you don't get any of the hassles including financing and inventory.

This deal helps both of you. You send hardware customers to him. He sends people looking for multi-user solutions to you.

The second item on the list is software. This is still a problem with Linux. In fact, Linux is probably not the right answer for a general-purpose business solution yet. But there are lots of things that can be done, and in many cases, Linux is a better answer for the customer.

But first, here is why it is a better answer for you:

• Minimal costs for software—Linux is (almost) free and 100 copies of Linux cost the same as one copy.
• You can get the source code for Linux and other programs such as Ingres or Postgres that you need to work with. You may not want to hack source code but it means it can be done and, as it is available without a $100,000 licensing fee, you can probably find someone to do the hacking for you if it becomes necessary.
• It costs your customer less. If the customer has $20,000 to spend to solve their problem and you can reduce the software cost by a few thousand by using Linux on the systems instead of a commercial OS, you get to put those few thousand in your pocket, or split the difference with the customer.

What should you do?

The best place to start is with something you need or want. If, for example, you currently work for a small company that sees the Internet as a way to expand their business, offer to help them get on the Internet. Get a Linux system, an Internet connection, and get it up and running. Get a news server and an ftp server running for their announcements and bulletins. Develop programs or scripts to translate internal files into announcements suitable for electronic retrieval.

You will have solved a problem, probably got a few gold stars (and a black eye or two) and have learned how to put together something that is needed by lots of businesses. Now you have a product: your skill and expertise in getting a small business electronically connected to the Internet. Start marketing but also start thinking of what else you can do; what would be a logical additon. For me, voice mail and FAX retrieval seems like the logical complement: If the customer doesn't have e-mail they can request the same information via FAX or, if all else fails, voice mail.

Building this system is not for everyone but it is something that is useful and needed. If, instead, you are an accountant with C programming experience, you might want to consider some accounting software. Or start with the Ingres or Postgres database and build an application that solves a problem for your business.

As Linux grows it will start to appeal more to commercial software vendors. This means that you can either get in on the ground floor building applications you need (and making money with them) or wait for whoever the new "Lotus" is to appear as the application provider for the Linux market.

For those of you who just want to use applications, start writing letters to Linux Journal. Tell us what you want and why. There is a lot of talent out there ready to write code. If you can identify what is needed, someone may provide it.