|
Editorials and Opinions
EDITORIAL: SOAP Suds
Microsoft Corp. is the Rodney Dangerfield of corporate behemoths. Not only is the company reeling from the antitrust trial and smarting from slower-than-expected sales of Windows 2000, but the company’s initiatives in the XML arena are widely rebuffed because they’re seen as being tied too closely to its proprietary Win32 APIs and DCOM architecture. No matter what it does, the giant can’t get any respect.
Case in point: the dismissal by the ebXML working group of the SOAP standard as an acceptable transport mechanism for XML messages.
Microsoft is the prime and most visible advocate of the Simple Object Access Protocol, and Microsoft intends to tightly integrate its BizTalk XML server platform and the SOAP protocol.
Does the fact that Microsoft is enthusiastic about SOAP mean that the protocol should be suspect, or that it should be viewed as a Windows-only protocol or as a scheme by Gates & Co. to make XML proprietary? No.
For once, Microsoft has done things the right way. The company submitted SOAP 1.0 to the IETF back in December 1999 as an Internet Draft. SOAP is based purely on the XML and HTML specifications and doesn’t require any specific features of Win32, Windows DNA, .NET or DCOM.
Yet SOAP remains widely derided, and ultimately the ebXML working group has shunned it as a potential message-passing protocol, choosing MIME-XML instead.
Were there technical reasons for avoiding SOAP? Yes, though some of SOAP 1.0’s weaknesses are addressed in SOAP 1.1. Was bias against Microsoft another reason for avoiding SOAP? In our opinion, that was a major factor.
There’s no doubt that Microsoft is committed to genuine standards only when it’s convenient, and has mastered the art of seducing developers with proprietary APIs and closed extensions to open standards. Look at what the company did with Kerberos key exchanges in Windows 2000 and its attempts to “embrace and extend” Java.
But does that make Microsoft substantively different from any other large software company? Not really. Its competitors use the same business plan, but they just haven’t been as successful in executing it.
Yes, Windows is closed, but so are NetWare and MacOS X. Yes, Windows CE is proprietary, but so are VxWorks and OS/9. Yes, DCOM and .NET are designed by both Microsoft’s technologists and marketers, but Java 2 Enterprise Edition is tightly controlled by the executive committees of the Java Community Process, dominated by Sun and a few large vendors with strategic relationships with Sun.
In fact, take a look at what Sun and its Java Community partners are doing with J2EE. Each new release, such as J2EE 1.3 and EJB 2.0, not only introduces new features, but also adds new requirements to its partners, mandating specific features, technologies and APIs. The rest of the Java industry, who need Sun’s blessing to sell their products, have no choice but to keep pace.
It’s time that industry leaders mature beyond their instinctive anti-Microsoft, pro-Sun posture and begin to evaluate the company’s initiatives on their technical merit.
Guest View: Taking a Byte Out of Crime
Theft. That’s what happens when someone takes your software without paying you for the privilege. You as a developer are as one with the movie maker, the music maker and the book writer when someone rips off the fruits of your efforts and makes them available to others at his profit, without paying for the right to use the intellectual property he sells.
You know this phenomenon as “software piracy,” and controlling this theft is a problem. From the early days of user groups to today’s auction sites, piracy has been around for a long time.
It’s a huge problem around the world. Some countries don’t recognize the concept of intellectual property as other countries do. Some people don’t bother to read the shrink-wrapped licenses, and so don’t realize that in some cases, they may be committing theft when they move software from their old system to their new system or remove the software and sell the whole package to someone else without getting the developer’s permission. (The other side of that argument is that the U.S. Copyright Act doesn’t restrict subsequent sales.)
The Software and Information Industry Association (SIIA) and the Business Software Alliance (BSA) commissioned the firm of International Planning & Research (IPR) several years ago to model the cost of piracy of business software. The result: a claim of $12 billion in lost revenue for 1999 worldwide. North America’s share of that loss was $3.6 billion. In spite of a continuing decline in the rate of piracy, according to IPR’s model, the dollar amount keeps rising because computer usage worldwide continues to increase. (Interestingly, there is no mention of the effect GNU/FSF and shareware is having on the apparent piracy decline.)
Remember, though, it’s a model. It makes assumptions, such as everyone who steals software would still want it if they had to pay full price for it. Who pays the full retail price for software today, anyway, when the mail-order houses discount software highly?
While different companies have different definitions for types of piracy, Microsoft’s four classes of piracy appear to be the most accepted (www.microsoft.com/piracy/basics/default.asp):End-user piracy is the casual copying of software, including disk swapping. As long as this was limited to small numbers of friends, the owners of intellectual property fumed but could do little to stop it. But with services like Napster and the Internet in general allowing widespread end-user piracy, publishers believe their losses due to end-user piracy are greater and are trying to find ways to combat it. OEM piracy is the sale of commercial software bundled with a new or used computer, but included without a license to do so. Counterfeiting is the duplication of software, at varying levels of quality, for sale. Mischanneling is the inappropriate sale of software destined for a specific market, such as selling to anyone an academic version of a product.
Not included in this list, but more and more frowned upon by software vendors, is rental or loan of software.
LONGTIME PROBLEM
Software piracy is nothing new. Software piracy was a problem when the first mainframe computers were installed, as programmers would move programs wholesale from computer to computer without the author’s permission. The reason Bill Gates negotiated the distribution of Microsoft MS-BASIC in the ROM of the original IBM PCs in 1981 was to stop cold any piracy—Bill and Co. wanted to be sure they were paid for each and every copy of any software sold, and they became militant about it.
User-group meetings were the main source of pirated copies, with people bringing in disks of new stuff to give to friends. This illegal copying was lost in the maze of people bringing in their own contributions and making legal copies of them, and also copying programs that were distributed exclusively at user-group meetings for no charge. Today, you see little if any software piracy activity at computer user-group meetings.
You didn’t have to leave home to steal. In the late ’70s, some RCPM (“remote CPM”) systems operated as pirated software repositories, providing a limited number of commercial titles to those savvy enough to find them and download them. The growth of the Bulletin Board System craze in the ’80s led some people to run pirated software BBSes, openly and with little attempt to cover up the activity. The Internet has also been an interesting conduit for pirated software: So-called “warez” newsgroups, exchange via Internet Relay Chat (IRC) file transfers, and archives accessible via FTP and Gopher were popular before the explosion of the World Wide Web.
Finding sources for pirated software meant getting into the inner circle. The reason so many BBS systems, warez newsgroups and FTP/Gopher archives could operate so openly was that there was no way for people not plugged in to know of their existence, let alone how to get to them. This inability to readily find piracy pools worked in favor of the panderers by making them smaller and harder-to-find targets for law enforcement.
The search engines make open piracy dangerous to the pirate operator, but there are pirate Web sites—and law enforcement has learned how to find them and shut them down. For that reason, pirate Web sites are falling out of favor, with the Jolly Rogers moving to that new darling of the Web, auction sites.
It was eBay where the author kept tripping over pirates selling their wares, and was sucked into the purchase of pirated software. (I did get my money back and bought legitimate copies from reputable sources—on eBay, I might add.) There are two major types of piracy on the auction sites: mischanneled products and counterfeits, particularly products or collections on CD-R disks.
LINES OF COMBAT
Combating piracy has been a long-standing game as well. One early way of keeping your software from being stolen was to write it for hardware esoteric enough that the user had to buy the hardware from you. Some software today is indeed sold in that way, requiring an expansion card to operate properly.
For expensive or special-purpose programs, a hardware key provided with each copy of the software limits running the program to the computer on which the key was installed. One type of key plugs between the keyboard and the system unit; the key’s appearance gave forth the term “dongle.” Quite a few packages still use the key-protection method. One drawback of the key protection method is the quality of the dongles and the ability to get a replacement dongle when the one you bought stops working. Another method is to use some unique physical attribute of the computer to come up with a hardware-based key, which can be used to generate an unlock code that allows the software to work on one, and only one, PC. Intel’s Pentium III serial number was an attempt to create such a unique physical attribute. Another one is the one-of-a-kind MAC address burned into each network interface card.
Many software developers have moved to different business models to minimize the effect of piracy. Borland International—now Inprise Corp.—championed keeping the traditional first-sale model but at a price that discourages piracy; indeed, it was considered crass to steal software priced lower than dinner at a good restaurant. Some businesses that used software to support a service or other product would “give away” the software, making it up elsewhere; Netscape is a good example of this model. Shareware was invented by the late Andrew Flugelman, then editor of PC World magazine. His PC-TALK program was widely distributed as freeware and proved to be a popular and well-received way to purchase software.
The open-source model is the most radical departure from traditional software sales, engaging a large volunteer development horde to produce quality software. The Linux operating system and the Apache Web host system are the two shining examples of community efforts that have gone well.
Finally, there is a new model being born on the Internet: the networked application and application service provider (ASP). This is to programs what thin clients are to networks: the program lives on your server and runs on the customer’s machine. In many cases, the business model for ASP programs is pay-per-use or pay-per-subscription-period, reducing the capital cost to the customer and loading it into an operational budget instead. One advantage of networked applications is that updating is easy and almost automatic, because the customer never has to load a local copy of the application. To software developers, additional advantages of this concept are that updates don’t require shipment, there is a steady revenue stream, and the piracy model changes from theft of code to theft of authorization tokens.
As a developer, you need to take piracy seriously, but don’t be stampeded into an unworkable control measure because of the hype. Look at your application, your applications base and your customers. Remember that technical support is a lever for keeping pirates at bay.
Stephen Satchell is a technical writer based in Incline Village, Nev. He can be reached at satch@concentric.net.
And Another Thing...: Death of the PC Programmer
Corporate microcomputer programmers used to have it easy. Organized loosely into little teams, many of them worked on small, easily definable projects. Their culture and style was very different from that of mainframe programmers, with their highly structured applications requirements, huge teams and “bet-the-business” projects that were measured in man-decades, not man-years.
The PC programmers were independent, and they were held minimally accountable to timetables and requirements because no one really knew what they did and everyone accepted that applications were most likely going to be late anyway. The greatest area of independence—and the area under great attack today—was in the selection of tools. The individual programmers used to pretty much hold the power over the tools they used. Like an independent plumber or auto mechanic, they chose their own code editor, C compiler, debugging tools and utilities. Because the tools were relatively inexpensive, usually $150 to $800, and because each project was distinct, the programmers had the authority to buy whichever tools they wanted.
Much has changed.
In the microcomputer world, application development projects have been growing in size and complexity. That is evolutionary. Companies involved in e-business (that would be about all of them) need Web sites that tie into old databases and legacy applications plus integrate new applications written specifically to operate on the Web. This is revolutionary. Programmers who were coding away in the basement are now up near the top floor as they work feverishly to get the Web site working and then add functionality. And they have bosses now. Lots of them.
PROGRAMMING IN TEAMS
The corporate programmer of today is likely to be part of a much larger team, organized and managed by some version of the title Software or Application Development Manager. There are frequently many teams, and managers representing different priorities loom like shadows over the formerly independent and free-spirited developers. For one thing, there is a huge tools standardization process going on in companies. Why have programmers use six different editors and a dozen debuggers? The corporate benefits of standardization are manifold, and they are burying the independent’s cry of “one programmer, one tool”: minimized incompatibilities on team projects, reduced training and support costs, and the ability to swap out programmers from team to team as emergencies warrant.
Today’s programmer is more likely than ever before to be told, “Son, this
is a Codewright and Visual SourceSafe shop,” and be asked to conform to a set of programming tools and practices that have been selected and approved by the programmer’s bosses. After all, that’s the manager’s job. The best a senior programmer can hope for is to influence the choice of tools. Much of the programmer’s penchant for finding and using new tools that are only 10 percent better (“Hey, that’s still 10 percent!”) has migrated to a late-night activity called downloading—but we’ll come back to that. If the loss of power to choose one’s own tools is not the death of the PC programmer, it’s certainly the end of the Old Ways.
The death of the hot-dog coder—or the demise of his or her influence—is also killing the market for low-priced tools in the $200 to $800 range. Symantec hurriedly sold VisualCafé to WebGain; Microsoft wants to sell you tools by subscription, much preferring that you buy the whole Visual Studio suite; and Inprise is just confused. It’s not just tool vendors scrambling to reinvent themselves. Visual Basic component vendors are scrambling by putting their parts together and selling suites and enterprise editions for $1,500 to $2,500. Why is the market for sub-$1,000 tools drying up? Partially because the programmer can’t choose and buy his or her own tools anymore. If a manager is defining a standard and authorizing a volume buy, they want superior functionality, support and upgrades thrown in.
EVOLVING MODELS
In parallel with the changing dynamics of the programming team, some interesting business models are appearing to serve the rearranged market. One of the more common models for software tool vendors today works like this:
1. Give away a working version of the product for free, via download off the Web site.
2. Then sell high-end versions of the product or support packages for the product at high prices.
The thinking is that the free downloads will seed the market with tens of thousands of programmers who will discover how the product is changing their lives and begin proselytizing the later adoption of the commercial version or support package to their bosses.
This is not a bad idea. But just who are the downloaders? Who actually has the time and inclination to download free development tools and tinker with them?
Managers generally don’t download and play with development tools. While they might ask one of their programmers to do this, the managers don’t have time to be experimenting like this, and frankly, neither do real working programmers. Only the most enthusiastic among commercial programmers is going to want to top off a 14-hour shift with some recreational bit-twiddling at home.
At work, these people are cranking out code like there’s no tomorrow—there’s scarcely time for them to be experimenting with new tools at work. Since I doubt we have become “a nation of downloaders,” I would suggest that most tool downloaders are garden-variety hackers. Yes, hackers. They still exist (see: open-source movement). If you don’t like the word hacker, you’ll also find ready acceptance of these free downloadable development tools among students, self-employed consultants and in undercapitalized start-ups.
While there is something noble about the attempt to seed a great sea of programmer enthusiasm for a tool, transferring that to the managers may be more of a challenge than simply selling the full-featured, fully supported version to the management decision-makers directly in the first place.
The bottom line is that there is now a bottom line. Many programmers inside companies used to roam freely with power over their own domains and their own tools. The e-business revolution has thrust them into the spotlight and into the clutches of managers who rightly insist on standardization and coordination of development efforts and tool selection.
Software tool vendors have watched the sub-$1,000 tool market dwindle away along with the death of programmer power. The smart ones will put less emphasis on reaching the hacker downloaders in search of elegance, and increase efforts to sell to you, the development managers, directly.
Ted Bahr is the Publisher of SD Times, and can be reached at ted@sdtimes.com.
What Do You Think?
SD Times welcomes feedback. Letters must include the writer's name, company affiliation and contact information. Letters may be edited for space and style and become the property of BZ Media.
Send your thoughts to letters@bzmedia.com, or fax to 516-922-1822. Please mark all correspondence as Letters to the Editor.
|
|
BACK TO TOP
Back to Top
Back to Top
|