Two years ago I was working as VP of R&D for a large telemarketing company. My job involved building and managing very large databases, primarily for the purpose of target marketing. We were spending an arm and a leg on software licenses in the R&D shop and another arm and leg on additional software elsewhere in the company. Although we did some things well, our expensive software was still not providing managers with the information necessary to make sound business decisions. For example, we were too late with information on profitable and nonprofitable campaigns, and we never did have good data on cash flow. I left that company after our R&D office was closed, disappointed that some good work went down the drain. I was also frustrated by the lack of software solutions for the company's basic business problems—and with just a hint of understanding of the board's unhappiness with how much our software licenses were costing.
After a few months of private consulting in database design, I was invited to a job interview by the CEO of a medium-sized manufacturing company named Action Target. (Action Target is in the business of manufacturing and installing target equipment and shooting ranges.) I was shocked at what I discovered. The CEO (an electrical engineer with an entrepreneurial bent) had the whole place running on Linux. Every single user, from sales, shipping and purchasing to production, engineering and accounting, worked on a Linux workstation, most of which were diskless. These workstations were served from a group of servers, which included a web server, a database server and collection of application servers. The application servers hosted all home directories in addition to the applications. The applications were mostly of the open-source or free variety, such as StarOffice, Netscape and TGIF, but also included a few proprietary applications, like Applix. The database server hosted PostgreSQL. Networking was accomplished via NFS between the servers, and the diskless machines were networked via bootp to the application servers.
What I discovered next shocked me even more. Not only was the whole place running Linux, but the CEO had written his own ERP (enterprise resource planning) application. He told me that he had investigated several other solutions, but they were all expensive, ran on expensive and proprietary RDBMSes and, most importantly, did not permit full customization. Therefore, he explained, he decided to write his own. The front-end tool he chose was Wylib, which he had coded. Wylib is a group of libraries written using Tcl/Tk. Tcl/Tk is easy to use, platform-independent, can act as a shell and belongs to and is widely accepted in the Open Source community. For the back-end RDBMS, he selected PostgreSQL.
This customized ERP system runs the company. Employees use the system because they can't do their jobs without it. The ERP is based on open-source software and is fully customizable. In order words, if we're not happy with it, we change it. Management estimates that this system has saved the company over $1 million (US) in operation expenses, not to mention the huge savings in software licensing.
After working at this company for a year, recoding the ERP using Wylib and restructuring the database, I've been spun off into my own company, WyattERP. WyattERP's mission is twofold: 1) to show companies how to save money by using open-source software and 2) to show companies how to create an ERP solution using Wylib that is inexpensive and completely suited to the needs of the business.
Wylib and the sample code can be downloaded from the web site, www.wyatt-erp.com. Wylib is open source and distributed under the OPL (Open Public License) agreement. I can be contacted via e-mail through the web site.
It's no news that Linux is finding a home in everything “from household appliances and consumer electronics gadgets to controllers used in aviation, factories and automobiles—application-specific devices”, as Evans Data Corporation (www.evansdata.com) reports. But it is news to discover how high the Linux embedding rate appears to be among those many devices.
In its “Embedded Systems Developer Survey”, Evans Data reports what it calls “a dramatic change in the software content of the microprocessor-controlled devices in the workplace and in our homes”. As a result, the firm predicts a triple increase in Linux-based projects over the next year, which it says is “part of a shift toward commercially available operating systems rather than those developed in-house.”
Tools are an issue. Without giving numbers, the study showed “keen interest in tools and products that can help developers get their projects completed correctly and on time”. Two-thirds of the surveyed developers said many embedded projects are started and never completed, and many more are delivered late.
What is it that developers like about Linux? Just what you'd expect: “Open-source code, royalty-free licensing and a large community of knowledgeable developers were cited as Linux's key benefits.” On the down side, Evans Data reports, “Factors slowing Linux adoption include lack of specific board support packages, availability of device drivers and fragmentation of the code base.”
Percentage of early-adopting PVR (personal video recorder, like the Linux-platformed TiVo) owners who are watching more TV than before they owned the device: 63
Percentage of current PVR owners who have “no idea” which channels the shows being watched were originally on: 12
Percentage of current PVR and VCR owners who like to scan past ads: 25
Number of channels on Dish Network's $31.95-per-month Dish 150 plan: 215
Number of open-source licenses on the Open Source Initiative's approved list: 22
Percentage of 74 leading brands that have lost brand value over the last ten years: 55
Percentage drop in value for all 74 leading brands over the same period: 5
Number of new products introduced last year: 31,432
Royalty charge paid by recording equipment manufacturers, according to the American Home Recording Act (AHRA), as a percentage of the manufacturer's revenue: 2
The Sound Recordings Fund's share of AHRA royalty payments: 2/3
Percentage of that fund allocated for “non-featured musicians and vocalists”: 4
Percentage of remaining royalties allocated for “featured recording artists”: 40
Percentage of remaining royalties allocated to “owners of the exclusive right to reproduce sound recordings distributed during the relevant year”: 60
Red Hat percentage on machines surveyed by Linux Counter: 28.39
Slackware percentage on machines surveyed by Linux Counter: 22.40
Debian percentage on machines surveyed by Linux Counter: 19.30
1-3: Electronic Media, from NextResearch
3: The Wall Street Journal
4: Dish Network
5: Open Source Initiative (www.opensource.org)
6-8: Financial Times
Jan Schaumann has begun a new software project that will serve a dual purpose: 1) to produce MakeMan, a project that will provide several GUI and non-GUI front ends to an XML interface for writing man pages and 2) to document the steps of an open-source project, from registering at SourceForge to releasing packages in great detail. See mama.sourceforge.net for details.
Show me a web app that can't be served from a Pentium 100 and I'll show you a dead dot-com.
If you want to be a platform vendor, and you want developers to invest alongside of you, your platform needs to be open source. This in fact has been true for the last few years.
All power is derived from the barrel of a gnu.
—Mao Tse Stallman (from the .sig of Alan Olson)
Technological progress is like an axe in the hands of a pathological criminal.
They're the next dead thing.
—Grady Hannah of Linuxcare on a company other than his own that he'd rather leave nameless.
The operating system wars are over, and operating systems lost. What we are now witnessing is the growth of environments where applications span multiple operating systems, rather than vice versa. Individual computers are the new object, with a handful of exposed interfaces and complicated internals hidden from the caller by standard protocols.
The real problem we face with the Web is not understanding the anomalies, it's facing how deeply weird the ordinary is.
This is an era when nonsense has become acceptable and sanity is controversial. Too many people fail to see a distinction between “the rule of law” and the edicts of judges. Unfortunately, these people include many judges.
—Thomas Sowell, in 1996
One can never consent to creep when one feels an impulse to soar.
Today we turn our dagger around and point it at ourselves. As goes the Internet Economy, so goes the magazine founded to cover it.
—Jimmy Guterman, uttering the parting words of Media Grok, the e-mail newsletter of The Industry Standard on the day it went out of business.
What lever did the invisible hand have to throw to start up the ghost in the machine?
The news caught everyone by surprise. Hewlett-Packard was buying Compaq for $25 billion (US) in stock. After both stocks were hit the next day, the deal was worth about $20 billion—even though the announcement boasted that the combined company “Will Have Number One Worldwide Positions In Servers, PCs and Hand-helds, and Imaging and Printing; Leading Positions In IT Services, Storage, Management Software.”
The numbers involved were staggering. Combined revenues exceeded $87 billion over the past four quarters, which would place the new HP second only to IBM ($90 billion) in size. The new behemoth would also have 145,000 employees in 160 countries—prior to staff cuts, which are sure to come mostly at Compaq's expense. If the deal goes through, Compaq as a company will only persist, in the manner of Netscape after AOL, as a “brand”. Together HP and Compaq brands account for about 70% of PC sales through retail outlets; but Dell now leads in sales overall, having moved ahead of both HP and Compaq. This, pundits mostly agreed, was clearly one strong motive for the deal.
Reviews in the mainstream media were mixed at best. Dan Gillmor of the San Jose Mercury News wrote, “It's hardly thrilling to be the leader in a market that is dull, nearly devoid of innovation and barely profitable.” In the UK, The Register wrote:
A braver choice for HP would have been to prise its way into new infrastructure partnerships, swallowing a Nokia or a Nortel. Or even an ARM. But instead of looking forward, HP has looked back and fallen on an acquisition target that looks agreeably like itself only financially weaker.
But in the Linux community, the mood was more upbeat. Linuxcare cofounder Dave Sifry said, “Gee, everyone said that there'd be consolidation in the Linux space, but this is a bit bigger than I expected!”
Dan Kusnetsky, VP System Software for International Data Corp. and one of the leading Linux analysts, says:
In the recent past both companies have been looking for ways to lower their overall software development costs by moving to high-volume, third-party software. This is quite reasonable. It is quite expensive to be a world-class supplier of operating systems.
He adds that the combined company would be supporting as many as eight different OSes, and offers these predictions: 1) since Tru64 UNIX is the lower-volume product and is tied to Alpha, the best features of Tru64 UNIX will be merged into HP-UX and Linux, and then Tru64 UNIX will be retired; 2) the best features of OpenVMS and MPE/ix will be merged into Windows and then one or both of those operating systems will be retired; and 3) the combined company will focus on HP-UX for high-end enterprise tasks, Windows for workgroup and desktop tasks, and Linux for Web-centric computing tasks.
Which isn't exactly bad news for Linux.