Slashdot Mirror
NYT On Open Source
KOHb writes: "The NY Times has a front-Web page article on using Open Source to bridge the "Software Gap." Mentions Mozilla, Apache, and other friends." Well-researched piece, talks of potential changes in software development -- both positive and negative -- as the Open Source model spreads. (Free, reg. req. to read.)
While Slashdot discussed the government's reluctance to accept Open Source in Linux -- Government Acceptance vs. Actual Use, apparently our Nation's proud warriors have no problem putting our nation at risk at the hands of a closed-source operating system as evidenced in USS Yorktown dead in water after divide by zero. The mighty Yorktown had to be towed back into port after its NT network crashed when a sailor entered a "0" into a data entry field.
-- Could you use my software consulting serv
"This issue of open source cuts to the core of the software business," said Jim Gray, a Microsoft researcher and a member of the presidential advisory group. "It is a real challenge, masked by a great deal of hype."
For those that don't know, Jim Gray is an ex-IBM, ex-DEC database guru who won a Turing Award for his book "Transaction Processing: Concepts and Techniques" (among other things)
It's an amazing book (on my desk as we speak) and in it Jim bemoans the fact that DBMS technology has been hampered by the fact that most of the innovation was being done behind closed doors, thereby forcing coders to reinvent the wheel rather than advancing the state of the art.
So then DEC implodes and Jim Gray goes to work for MS. I seem to recall that he got a 7 digit signing bonus. And now he's describing open source as a "challenge"
I'm not suprised, but I'm more than a little saddened.
--Shoeboy
"The truth is, great software comes from great programmers, not from a large number of people slaving away," Mr. Joy said. "Open source can be useful. It can speed things up. But it's not new, and it's not holy water."
I've seen this come up a couple of times before; there seems to be some resentment on the part of the old guard towards a younger, somewhat arrogant crowd who think just because they can slam out an ls frontend with the gtk toolkit that they're wizards(no, of course this doesn't apply to everyone, or even most people) Not meant as a flame, I'm not part of either crowd, just an interested observer.
Eric S. Raymond, an open-source evangelist, observed that Mr. Torvalds was "the first person who learned how to play by the new rules that pervasive Internet access made possible."
This is probably one of those statements that irritates the older crowd as well. The internet was founded as a scientific and technological collaboration tool. Linux is the new kid on the block.
--
I thought the most interesting part of the story was the idea that the project methodology of Free Software (or open source if you like) is becoming recognized as not only a "challenge," but a better way of doing things.
(The licensing issues of Free Software are not explained well in the article, but it has other virtues.)
IBM discovered long ago that programming teams don't get the job done faster if you throw more bodies at it. Yet Microsoft and other proprietary software makers really stick to the same old project methodology. In one way or other it is quite similar to the old "waterfall" methodology, with a schedule driven by market needs, and features and bug fixes dropped at the last moment so the buggy release goes out the door.
Consequently, Microsoft attracts highly-skilled, highly-paid engineers to enlist in these "death march" projects and rewards them with stock options based on performance. No doubt few at MS feel able to work 40-hour weeks.
But the result is more and more massive software, more and more legacy code to maintain, more and more bugs to fix, more and more releases to fix the bugs, and software never seems to get radically better.
This is the old "software crisis" of the 1970s, back again with a vengance. Now that Microsoft has gained a monopoly in certain areas, they have little incentive to innovate or find better ways to do things.
Many of you will argue about this point. But again I say the projects are market-driven by the bosses, not customer-driven. The aim never seems to be delivering code that is elegant or really functional, but only shipping it out the door and charging money later for a release that promises to fix those bugs (and introduces more). And the software gets bigger and bigger, with an idea to include as many functions as possible. The customer more and more relies on the behemoth manufacturer to fix things and customize them and support them, and those jobs aren't being done better now than ever.
On the other hand, a team of variable size that is distributed over the Internet does seem to work in certain situations. The code it produces is closer to what customers want and need. Even though in many cases it is not finished, it provides enough for a custom programmer to fill in the missing parts. The C and Unix environments have become a lingua franca for professional programmers.
The eXtreme Programming model does not use teams distributed over the Internet. Instead, it puts the much smaller teams near the customer. Like the open source model, it releases in stages, and it reuses components and other programs well. And it leverages software as a profession. Programmers should not work more than 40 hours a week. Instead, they need to get out of the office, become human beings, relate to their families and societies, and become inspired once again why programming is important, and why it is so important for all of us to do our best job.
When Microsoft programmers start working 40 hours a week then I know they will have learned something from these new methodologies. But they won't be able to do that unless their bosses learn it first. It might indeed taking a lot of shrinking of the big software factories before this happens.