BBC Falls for Sun's Open Source Software Spin 9 comments
an article to
-
Font Size:
-
Print
- TweetThis
Some technical editor at the BBC has been snookered three ways by Scott McNealy of Sun (JAVA) in this January 21 “news” story. The open source blogoblatherers in the European Union are multiplying the effect.
First, there’s the sentence in the article that says McNealy has been asked to “prepare a paper” by the Obama administration. Assuming the U.S. government works like the rest of the world (not necessarily true as U.S. taxpayers are well aware), being asked to “prepare a paper” is probably like saying “Don’t call us; we’ll call you.”
Second, the gist of the BBC article is that, according to McNealy as quoted by the BBC:
“It's intuitively obvious open source is more cost effective and productive than proprietary software. Open source does not require you to pay a penny to Microsoft (MSFT) or IBM or Oracle (ORCL) or any proprietary vendor any money."
In fact that statement’s not at all obvious, nor is it even true (in multiple dimensions). There is no reason you should expect open source to be more cost effective since the open source concept relates to “free as in air, not free as in beer.” And since software is software, how can the way it is licensed make it “more productive”?
Third, it’s Scott McNealy promoting open source!!! He’s arguably the most proprietary guy in IT industry history since…. I dunno, maybe old Tom Watson. He's glommed onto open source to try to save Sun and its billions of lines of "proprietary software" from going down the tubes.
He’s backed in the BBC story by quotations from a Red Hat (RHT) vice president. See a trend here: Red Hat and Sun want to sell hardware, storage, software and services to the U.S. Federal government.
I hope they are successful. And I don’t have any concerns that IBM, Microsoft, Oracle, SAP and other open source software suppliers larger and smaller than Sun and Red Hat don’t know how to play the game with the U.S. government too. (Actually some of the smaller suppliers of open source software might think things happen “before someone sells something.” But not the smart ones.)
But this article makes the BBC staff look like pretty naive.
Related Articles
|






















You say, "He’s arguably the most proprietary guy in IT industry history."
Let's look at this. Sun's first systems were based on the open source BSD distribution of Unix using the commodity Mot. 68020 processor. Each system we shipped from day 1 had an industry standard ethernet port and TCP/IP stack when our competitors were shipping DecNet, SNA, Vines, and Netware. We introduced the Sparc processor and freely licensed it to anyone with $100 (sparc.org) as well as made the Sparc architecture an IEEE standard so that companies like Fujitsu could compete and partner with us. Last year we open source the CPU firmware for our T series of processors. Can Intel or IBM say the same about their X86, Itanium or Power CPUs?
We invented NFS as a network file sharing system and licensed it liberally. When we switched to System V Unix, it was considered an industry standard (opengroup.org). When we created Java, we also created a community to build it for every device in existence and let IBM, Microsoft and other contribute.
Now, what part of this strategy do you find proprietary? Compare that to Apple who has a propriety photo DB software (iPhoto), User interface (Aqua), System API (Cocoa) and is the most secretive of development organizations. Compare us with Microsoft, Oracle, SAP, CA and other large software organizations and I think you will find that we are quite open about our code, intentions and strategies.
It is true. One ought, if one is running some kind of enterprise with any software, have access to support for that software. One can get that for free, if one knows how. Or one can purchase same, from multiple sources. "With enough eyes, all bugs are trivial". If M$ understood that, its software would be better than Vista.
>> There is no reason you should expect open source to be more cost effective since the open source concept relates to “free as in air, not free as in beer.” And since software is software, how can the way it is licensed make it “more productive”?
For those who are willing to "use" but not "support" O/S software, the $$ cost is very low, if not quite zero. And, if you've been paying attention, Linux machines are far more stable than M$ machines. Whether industrial strength relational databases from O/S will ever match DB2 or Oracle is a different question. For operating systems, and basic infrastructure (and desktop applications), O/S is far more productive on a $$/user basis. How much of the funky Word functions (and are not in OO) do you use? If the answer is few or none then linux. How many viruses to you worry about? If the answer is Lots, then linux. You see where this is going. You either can't read, or lie a lot.
IBM - I don't think so.
Microsoft - Not so much
Red Hat - Nope
Oracle - doubtful
Cisco - nothing
NetApp - nada
EMC - not a chance
HP - empty
Then consider the directions of open storage solutions and the games are about to change again as they did with the advent of NFS. Consider your facts before trying to make a point about the person in the middle of the story.
Back in the day, Sun was selling UNIX/ Solaris machines, and the operating system complied with open standards, such as POSIX, X11, and TCPIP There were even free GNU compilers, so you didn't have to buy a Sun C compiler. At the same time, Novell and Microsoft were selling proprietary platforms with proprietary interfaces, proprietary networking stacks, proprietary programming APIs, all based on MSDOS.
Proprietary is fine sometimes, and sometimes not. In certain setups, proprietary is a non-starter. Imagine if Google was trying to scale out its search engine infrastructure having to negotiate licenses from Microsoft one-by-one. Or if someone was building a 500-node supercomputer for academic research purposes. Each of the 500 nodes would need a license for the O/S, and whatever applications it uses. In many cases, the licensing costs would be back breaking, and sink the project. In a university setting, or even overseas, labor is nearly free, so that leaves the cost of just hardware and software licensing.