The Most Influential People In Open Source 189
mmaney writes "As part of its 2009 open source best practices research, MindTouch asked C and VP level open source executives who they thought are the most influential people in the industry today. The list is ranked by the effect these individuals have had on the open source industry. Over 50 votes from executives in Europe and North America were cast. There were a few surprises from outside of the open source industry. Steve Ballmer got a mention because of his negative remarks on the open source industry and its subsequent positive impact. Vivek Kundra was mentioned because of his contributions to the industry inside the US Federal Government. Notably absent, however, are any influential women."
Relatedly, Matt Asay (who is also on the list) writes about the decreased need for open-source evangelism, noting that several people on the list are there "not because they're open-source cheerleaders, but because they have helped vendors and customers alike understand how to get the most from open-source investments."
Execs, etc (Score:4, Informative)
Re:WTF? (Score:2, Informative)
What's so hard to understand? When Ballmer started mouthing off about open source it was probably the first time lots of people heard of it. Just because he wasn't influential in the way he would have liked doesn't mean he didn't have an influence. They aren't pretending that he's deliberately helping.
Re:WTF? (Score:5, Informative)
Comment removed (Score:3, Informative)
Re:Fifty votes from "executives"? (Score:4, Informative)
Speaking as a former Apple employee: so _that's_ why a bunch of senior engineers in the hardware devision were let go or put in such a horrible situation that they left. Apple isn't being that great to their engineers and is focusing more on how hard they can drive them to produce new products every 6 months.
Re:Hear Hear! (Score:4, Informative)
Why even make files a UI metaphor at all? I did a little experiment a few years ago. I got around twenty people, some computer scientists, some completely nontechnical, some from a scientific background but not directly related to computers, to define a file for me. Only two of them gave me the same definition, and they were from a UNIX background so defined a file as an untyped stream of bytes with a name associated with them. Almost half answered with something along the lines of 'I don't really know'. Then I asked people what a document was. There answers weren't all the same, but they were close and people were a lot more certain that they could define a document than a file.
A UNIX file is a nice abstraction for the OS to present to programmers, because it's simple to build complex things on top of it. It is a terrible abstraction to present to users. Try explaining to a user why a Word document can contain images in the file but an HTML document refers to images in an external file, so dragging one to a disk works fine and dragging the other to the disk loses all of the inline images some time and you'll see quite how bad an abstraction files are. NeXT-style bundles go a little way toward improving the situation, but not far enough.
I totally agree on the pipes concept. You should take a look at System Services on NeXT / OS X, which are a good step in the right direction. While streams of untyped bytes are fine for persistence, they are horrible for communication. Something trivial, like sorting the output of ls -l by file size (displayed in human-readable form) is insanely complicated on a UNIX shell relative to the complexity of what you are actually trying to achieve. If, rather than a set of lines of text, ls emitted an array of objects, then you would just sort them by the size attribute and pretty-print them. Depressingly, this was actually solved nicely in Smalltalk-76, where the Transcript window gave you exactly this kind of interaction.
Re:Not influential to me (Score:3, Informative)
I find it very hard to disagree with your list. I am not completely sure that I'd put Ian Murdock on the second list - most of the things you dislike about Debian seem to have collected later and he's done some good work on OpenSolaris that makes up for Debian. Marshall Kirk McKusick and Bill Joy both deserve to be near the top of the list for their respective achievements.
One person I'd add is Keith Packard. He doesn't get much press coverage, but he is largely responsible for the fact that X.org is now feature-competitive with Apple's Quartz. He's been involved with the project for a long time and has evolved the X system from a remote display system for 1-bit displays running multiple terminals to something with support for compositing, accelerated rendering, and so on. Most importantly, he's done this without sacrificing backwards compatibility and without hard-coding policy decisions into the display server. You can still fire up a xterm on a VAX running 4BSD (or whatever) today and display it on a handheld machine running Linux with X.org and have it render correctly. Only now the drawing commands will be sent to an off-screen buffer and then composited in the display server according to a policy decided by something complex like Compiz or something simple like xcompmgr. I can't think of any other project that has achieved that long a period of continual evolution without breaking (binary) compatibility or throwing away the entire codebase and starting again.
Re:Hear Hear! (Score:3, Informative)
Something trivial, like sorting the output of ls -l by file size (displayed in human-readable form) is insanely complicated on a UNIX shell relative to the complexity of what you are actually trying to achieve.
ls -l | sort -n -k 2
Oh, the complexity!
influential Dutch (Score:3, Informative)
Bram Moolenaar (vim)
Wietse Venema (postfix)
Guido van Rossum (python)
Stephen R. van den Berg (procmail)
Not the smallest programs either. Yeah i'm proud of my country :)
Re:Hear Hear! (Score:2, Informative)
ls -lSh
Even shorter :)