Even as Linux's popularity and visibility increase, Windows NT is becoming the server of choice for corporate America.
Recently, Linux has been showing up regularly in the mainstream news. For example, there is an article titled “The Greatest OS That (N)ever Was” in the August issue of Wired which is both technically and historically accurate. This article will certainly help increase the visibility of Linux.
Even as Linux's popularity and visibility increase, Windows NT is becoming the server of choice for corporate America. Perhaps choice is not the right word, but NT is fast becoming the automatic answer for companies that don't want to rock the boat.
It wasn't that many years ago that everyone was saying “You can't go wrong buying IBM.” At that time IBM was the 360 mainframe and its offspring. It was hard to imagine that any other company or operating system could take the place of this big iron. Unisys, CDC and others had pieces of the total mainframe market, but IBM seemed to be unstoppable. In fact, people commonly said “IBM machine” when refering to a computer.
There is a lot to learn from this experience. Was it good marketing on the part of IBM that made them the standard? Initially, I expect it was. After all, marketing consists of doing research, finding a need, matching your product or service to that need and advertising the benefits of your solution.
Once other companies started to offer solutions that were as good or better, IBM took advantage of its huge market share to impose itself on the consumer and squelch the competition. For example, even though a standard character set, ASCII, existed, IBM used EBCDIC, their own character set. To enter the market the competition was forced to be compatible with IBM and thus had to use EBCDIC.
As machines became more and more powerful, minis and then micros began to take over the office. At this point the game changed. The new IBM was not a hardware/software giant like IBM but the software-only company, Microsoft. Microsoft offered a better solution which, at the time, meant a more cost-effective solution. As a result, they got the market share and that is where we are today.
Microsoft has clearly overpowered their competition in the IBM tradition and has vendors such as Intel going along for the ride. The standard is to be Microsoft compatible; Plug-n-Play means working with Microsoft operating systems. Virtually all PC hardware vendors bundle a Microsoft operating system with the hardware.
The Linux community is playing catch-up, developing software like SAMBA so that Linux will interact with Microsoft systems, dealing with Plug-n-Play hardware that only works with Microsoft operating systems and so forth. So, we have software that addresses compatability issues. Now, we need the hook—the thing that will make Linux a better solution. That hook isn't going to come from out-spending Microsoft, it has to come from innovation.
I am more of an observer than an expert, but I do have some suggestions. Some of these ideas come from a Seattle-based Linux mailing list, the GLUE list (check out http://www.ssc.com/glue/) and from talking to people about what they want.
We need a certification program—a way to say that someone is a certified Linux technician. Red Hat has started this effort (check out http://www.interweft.com.au/redhat/) with what they call “The Red Hat Commercial Support System”. I talked to Robert Hart about it, and he pointed out that since the effort is needed, Red Hat decided to go ahead on their own to get it moving. Hopefully, other vendors, or possibly Linux International, will turn this idea into an industry-wide effort.
We need a service business package. It is equally applicable to any business where the service provider travels to the homes of his customers—anything from a window washer to a roofing contractor. Software for these types of entrepreneurs needs to run on a laptop, which includes a printer. The issues that must be addressed are customer database, cost estimation, invoicing, inventory and travel route planning. This is not a trivial application, however, once all the pieces are integrated, it offers some great potential markets. While each specific market (Contractors, for example) is not huge, the possibilities are enormous. Also, these various industries have newsletters, discussion groups and, in some cases, magazines. Offer an innovative solution to such an industry and word will spread quickly.
There are packages for FAXing and voice mail from Linux. Putting together a commercial-quality solution for a small business would make a great stand-alone application. As the majority of the interface between users and the system would be via the telephone, there is little concern for users having to learn to use a new computer. The FAX interface could be handled through an office e-mail gateway (which Linux could handle) again isolating the office worker from the actual Linux command line.
Another huge market is government. Governments tend to have limited budgets and want a solution that just addresses their particular problem. Also, governments tend to communicate with each other—particularly at the city and county level. This is advantageous as it means that a good solution would require little marketing beyond a couple of success stories. Government solutions could include meeting-room scheduling, general word processing, accounting, Internet connectivity and web servers. The main difference between this market and the general market is the ease with which you could market a tailor-made solution. It also offers a good chance for local consultants to get involved.
All of these solutions need a truly easy install. By this I mean a bootable CD which installs and asks for the information it needs. While it might be difficult to make a universal Linux system this easy, for a specific market, it should be easy.
Remember, we don't have the money to be bigger than Microsoft. We need to go after markets that better fit our size. If we solve the problems of a few of the smaller markets, people will start coming to us for a solution.