Competing as Software Goes to Web

Today’s New York Times Business section has an article by John Markoff Competing as Software Goes to Web that begins as follows:

Can two bitter rivals save the desktop operating system?

In the battle between Apple and Microsoft, Bertrand Serlet and Steven Sinofsky are the field generals in charge of competing efforts to ensure that the PC’s basic software stays relevant in an increasingly Web-centered world.

The article uses a common motif in business articles, framing an issue as a battle between one company and another with representatives of each company featured to make it more personal.

What makes the article particularly interesting — and at times amusing — is to read it with open-source and Linux in mind. The special attraction is what is left unsaid, as neither Linux nor open-source is mentioned in the article. For example, the article continues:

The two men are marshaling their software engineers for the next encounter, sometime in 2009, when a new generation of Macintosh and Windows operating systems is due. Their challenge will be to avoid refighting the last war — and to prevent finding themselves outflanked by new competitors.

Left unsaid is the new competitor on the flank — Linux and open-source.

Continuing:

Many technologists contend that the increasingly ponderous PC-bound operating systems that currently power 750 million computers, products like Microsoft’s Windows Vista and Apple’s soon-to-be-released Mac OS X Leopard, will fade in importance.

Ponderous? Left unsaid is that while Vista and Leopard may be ponderous, Linux is much less so. You can run it on hardware that can’t boot either of the above operating systems.

In this view, software will be a modular collection of Web-based services — accessible by an array of hand-held consumer devices and computers — and will be designed by companies like Google and Yahoo and quick-moving start-ups.

“The center of gravity and the center of innovation has moved to the Web, where it used to be the PC desktop,” said Nova Spivack, chief executive and founder of Radar Networks, which is developing a Web service for storing and organizing information.

Faced with that changing dynamic, Apple and Microsoft are expected to develop operating systems that will increasingly reflect the influence of the Web. And if their valuable turf can be preserved, it will largely reflect the work of Mr. Serlet and Mr. Sinofsky, veteran software engineers with similar challenges but contrasting management styles.

Note the emphasis on the future as they produce systems “that will increasingly reflect the influence of the Web”. Left unsaid is that there already exists such a system, Linux. Linux is the the operating system used to build much of the modern web.

A developer who has worked for both Microsoft and Apple is quoted comparing the different styles:

One software developer who has worked at both companies — and asked not to be identified because he still consults for Microsoft — compared the two men’s approaches to the difference between martial marching band music and jazz.

Mr. Sinofsky’s approach, he said, is meticulously planned out from the beginning, with a tight focus on meeting deadlines — a crucial objective after the delay-plagued Vista project — but with little room for flexibility. In contrast, the atmosphere inside Apple’s software engineering ranks has been much more improvisational.

“Tight focus on meeting deadlines” and “little room for flexibility.” Left unsaid is that while perhaps that approach works for Microsoft it is the exact opposite of the open-source approach.

We learn next these guys are smart:

Both men are the best of a technical elite. “Bertrand is wicked smart,” said Dan’l Lewin, a Microsoft corporate vice president who worked with Mr. Serlet at Next. “He was one of the bright lights.”

Mr. Lewin now works with Mr. Sinofsky, who he said had brought needed discipline to the company’s largest development project. “His ability is in understanding the end-to-end process and architecture and knowing every nook — it’s amazing,” Mr. Lewin said.

“Knowing every nook?” Hello Linus, hello Andrew, hello Ted and all the other core Linux developers. Left unsaid is that they’re amazing too.

There are risks:

The potential risk in the Microsoft approach, he said, is that “they’re like the test pilots who won’t pull up when they see the tarmac.”

Left unsaid is that they also need to watch out for the tar pit that is their development process and the new tar-ball’s produced every day by the open-source community.

They are using a closed process:

Now in charge of the company’s most important development project, Mr. Sinofsky has proved to be far more secretive than his predecessor, Jim Allchin, who retired from Microsoft this year. Shortly after the consumer release of Vista in January, the company took the unusual step of issuing a statement saying that it had nothing to say about its plans for future operating systems.

Left unsaid is that secrecy is anathema to open-source. You can read every day all anyone has to say about Linux by reading all the open mail lists.

There is a plan however:

Microsoft is trying again to reconcile the PC operating system with the Internet. It calls the new strategy Windows Live, an effort to leverage its desktop monopoly onto the Web.

“Reconcile the operating system with the Internet?” Left unsaid is that Linux is already very reconciled. Then again, Linux doesn’t have to worry about leveraging a monopoly.

The company’s chief executive, Steven A. Ballmer, refers to the new approach as “integrated innovation.” But it is less clear yet whether Mr. Sinofsky will have the agility to respond to what is being called an era of “loosely coupled innovation” — an agility that has been the hallmark of nimble Web services developers.

Small groups of programmers have been using the Internet to introduce services far quicker than the slow-moving operating systems projects have been able to respond.

“The challenge for Steve is to get Windows in that mode,” said Michael A. Cusumano, a professor of management at the Sloan School of Management of the Massachusetts Institute of Technology. In the future, he added, Windows will most likely be broken into “smaller pieces released more frequently or put up as a Web service.”

Small groups, agile development process, small pieces? Left unsaid is how familiar these ideas are to open-source folks. This will indeed be a challenge for Microsoft, though they can read all those open mail lists to get some insight into how to do it.

That will mean that the era of big software releases may have come to an end.

“I think that you won’t think about big new releases in the future,” said John Seely Brown, the former director of the Palo Alto Research Center. “You really want to be able to make lots of incremental improvements in ways that things just get better and better.”

Big new releases in the future? Left unsaid is that this is old news for Linux, as for at least a year or so Linux has evlolved so that production-quality releases come out every few weeks, incremental improvement by incremental improvement.

“It’s a very important, longer-term trend of incremental innovation,” said David B. Yoffie, a professor at the Harvard Business School. Software will be “increasingly componentized and offered over the Web.”

Couldn’t agree more.

The article ends with the following:

Still, there are those in the industry who believe that the very nature of software will assure both Mr. Sinofsky and Mr. Serlet comfortable careers for the foreseeable future.

“Software is like the tax code,” said Jean-Louis Gassée, a venture capitalist and a former Apple executive, who in the 1990s developed an operating system called Be. “You add lines, but you never take anything away.”

Left unsaid is that code can be taken away. For example, the UpFRONT column in the February 2007 issue of Linux Journal reports:

The [kernel] systcl call, allowing users to configure kernel parameters at runtime, is likely to go away.

Linus Torvalds and Andrew Morton have both expressed the opinion that taking sysctl out would be the right thing to do — Linus because no one uses it and Andrew because it would be a shame to leave a big wad of such useless code in the kernel permanently, if a viable alternative existed. But, in case it really would break too much stuff, Albert Cahalan has volunteered to be the official sysctl maintainer if one is needed.

Linux will continue to compete as software goes to the web — at flank speed.