Open Source? Sure, it’ll work…when bees can fly! (Snicker snicker)
Mitch picks up and on Steve Ballmer’s claim that Open Source software is necessarily shoddier than software written in closed shops. Writes Mitch:
What Steve Ballmer doesn’t realize is that his own company takes virtually the same approach, hiring some young coder from Kentucky or Bangalore to write a bit of code that gets appended to Word or Windows, but without the feedback of coders and users that are incorporated into the collective decision-making that produces open source software.
There’s another proof that unmanaged, market-driven projects of some complexity can produce robust, innovative software: the Internet and Web themselves. What the hell, I’m going to out on a limb here and just blurt it out: The Internet is a better piece of software than Windows XP.
Categories: Uncategorized dw
Without wanting to defend either claim that open or closed source projects produce better software, I do think that your statement that ‘The Internet’ equals robust open source, is wrong.
First of all the web works because it is largely user driven, and users are willing to tolarate errors to an extremely high level on the Web. How often have you hit refresh to get the correct response? How often have you checked out a shopping cart only to find that the transaction didn’t go through, because Versign was hosed? These are only a few examples of how we tolerate these errors in the internet. We do not however tolerate these kind of errors from our computer (runnning winxp or anything else); if you open a file and the file system responds with ‘file not found’ your default action is not to hit open 5 more time to eventually make it succeed. We have a very different failure-tolerance model for local computers than for the internet. I have been building large scale distributed system with a focus on fault-tolerant for 15 years now and believe the internet is anything but robust.
Secondly your statement that the Internet is build on open source. This is not true. The original TCP/IP stack was developed for BSD at a university not using an Open Source model. The major web browsers were closed source for a very long time (or still are). Real didn’t put out parts of their code base until it was completely matured. Cisco is not giving you code for their router operating system. The various BGP implemntations are largely closed source. eBay, Amamzon, Google, Swab, Salesforce.com, do not put out their systems for your to inspect. This is the fabric that runs the internet and it is largely closed source.
There are good examples of why in a number of cases Open Source projects produce better software (the 1000 eyes example), but the internet is not one of them.
Open Source software isn’t shoddier, but it is usually a different kind of software aimed at a different audience. Corporate-developed, closed source software succeeds with consumers who favor easy-to-use over flexible. OSS succeeds with sys admins and nerds who favor flexible over easy-to-use. It’s interesting to note how these two groups talk right past each other. The OSS group seems to be lamenting the fact that average Joe’s won’t learn how to compile source at command line, while Microsoft lackeys seem to think that the nerdier set should just let some stuffed-shirted marketing team at Microsoft determine the future of computing. Linux, for example, might be a real contender for the x86 desktop if only some distro would take away some choices and provide a consistent experience for the user. When and if such a distro appears I predict it will be universally hated by the existing Linux community…and an instant hit with consumers.
Setting aside for a moment the question of quality, I want to recognize that the processes of open versus closed source code are very different. That may seem so obvious as to not merit mention. However, as a starting point, let’s recognize the self-organization inherent in open source. This property by itself is sufficient to suggest there is something to be gained by comparing the beginning and the middle and not just the end result.
I think an taking either extreme view, that software should be the product of un-managed or closely managed development processes is a mistake. There are a range of options available at any step along the way and clever people choose well among those options. Shutting out the possibility that someone outside Microsoft could contribute something worthwhile to Windows, as Ballmer does, is just plain balmy. Saying that an enterprise application must be the product of an unmanaged open source development effort is also nutty. The middle way is a wide, wide one and it would be worthwhile to take advantage of it now and again.
Werner, you make many fine points, but notice that I didn’t say that the Internet is open source.
The Internet, the World Wide Web, the “coming soon now?” Semantic Web, highway systems, automobile engines, the telephone system, and many other things, are great for one reason – over time, they have evolved through the efforts of many technologists in UN-related companies. Someinvents something, someone else improves it, and it becomes public knowledge at some point. This is a good argument for both Open Source, and for reverse-engineering – either works, one requires less effort.
One other thought – despite what Ballmer thinks, many many people have added interesting and valuable things to Windows – they are called “applications”. Maybe I’m old-fashioned, but when did we start requiring the operating system of a computer to have so many applications already pre-installed? If Microsoft’s OS folk focused on making the actual O/S great, instead of focusing on featurs, it might actually BE great.
The Net isn’t really “software” in the same sense as an OS, is it ? It’s several key standards at key places and a lot of interrelated infrastructure, isn’t it?
An enabler like software, sure but different. Most software has a designed purpose – it “mimics” or describes some process, and provides means of output defined by constaints, I think. Whereas does the Internet have a purpose? – other than making it possible for a user to say “I can” when wanting to connect to whomever and whatever I can find, and then “do” what softwares will let her or him do.
I like Mitch’s point regarding middle ways – and isn’t it an infrastructure like the Net, supporting opersating systems and software built to collaborate that allows (for example) an approach like Xtreme Programming to take hold and then spread and grow?
The internet is a better piece of software than Windows XP
I bet this will become a classic quote. David Weinberger rants on Steve Ballmer’s remarks that Open Source software is necessarily shoddier than software written in closed shops: “There’s another proof that unmanaged, market-driven projects of some com…
The divergent minds have taken us in all directions. Both Open Source and Closed have merits. Closed will always define “desktop” oriented standards until a group of Open Sourcers can get the standardization and usability issues under control. Unfortunately, they may never without being guided by a much larger force of common goals, management and avoidance of duplicity.
Maybe if Apple stops the proprietary hardware bit, OSX would catch on more. But then, do we want Steve Jobs as our Bill Gates?
You gotta love to hate what’s forced on you.
The divergent minds have taken us in all directions. Both Open Source and Closed have merits. Closed will always define “desktop” oriented standards until a group of Open Sourcers can get the standardization and usability issues under control. Unfortunately, they may never without being guided by a much larger force of common goals, management and avoidance of duplicity.
Maybe if Apple stops the proprietary hardware bit, OSX would catch on more. But then, do we want Steve Jobs as our Bill Gates?
You gotta love to hate what’s forced on you.
Ballmer’s roadmap comment is an interesting one. A roadmap tells you what your choices are, but doesn’t tell you where to go. Longhorn is coming out in 2006.
So our MS option is ‘persuade Ballmer to put it on his roadmap, and wait until 2006 to see it’.
With Open Source I can hire someone to make it now, and submit it to the main trunk for release if others like it.
Ballmer doesn’t provide a roadmap; it is more like a construction schedule for a new railway line.
When a variable is finished with it’s work, it does not go into retirement, and it is never mentioned again. Variables simply cease to exist, and the thirty-two bits of data that they held is released, so that some other variable may later use them.
i am student of 6th sem so plz send the source code for our project that is automatic railway gate control using pic 16f84 plz help me