From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: To: corey@bitworthy.net, 9fans@9fans.net Date: Sun, 25 Apr 2010 13:55:54 -0500 From: blstuart@bellsouth.net In-Reply-To: <201004242314.24052.corey@bitworthy.net> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Subject: Re: [9fans] Mars Needs Women Topicbox-Message-UUID: 0d5d87a2-ead6-11e9-9d60-3106f5b1d025 > It infers that "what Corey wants" is to bring GNU and Linux into > Plan 9. > > Which isn't true. I must admit to jumping to that conclusion too easily. I guess it's because the most common discussions here start either with "I've been beating my head against a wall; has anyone seen this or can tell me what I'm doing wrong?" or "Here's a little something I whipped up, so if you're interested grab it and play with it." or "You guys need to add features X, Y, and Z to Plan 9 because it's losing the competition (presumably for the largest number of users)." And anything that doesn't start out like the first two does seem to sound like it's a discussion of the third type. > I think there's lot's of potential out there for Plan 9 in "Consumer Space"... > I honestly thought that I might see more folks interested in imagining > what glenda looks like outside of the research and data center[1]. Well, I can't speak for everyone, but I personally find end-user applications and experiences to be the least interesting part of computing. I'm only half joking when in talks or in class, I say that computing would be a lot easier and more fun if we didn't have to deal with users. Also, I don't really like the things that seem to be the most widly accepted in the consumer space. So for me it's a virtue that the community doesn't find popular acceptance to be an important factor in design decisions. > Personally, I think carefully identifying, then porting/forking just a few, > _select_ pieces of software from the *nix space, then maintained natively > in an alternative Plan 9 based distribution, is a more interesting and direct > route. Well, a few select ones have been ported over the years, ghostscript and TeX coming to mind first. There was even a late unlamented X11 port. > Unfortunately - that means noisy discussion and collaboration amongst > people from a variety perspectives and skillsets/experience. Which is > anathema to 9fans lone-ranger aesthetics. I don't really understand this part. Why does porting something or even building an alternative distribution require discussion and collaboration? I can certainly understand how discussion could deveop as a result of such a thing being done, but I don't see how discussion is a prerequisite for it. When discussion of a more philosophical nature does happen here, it tends to be a) the evolution of some other conversation, and b) focused on a fairly specific question, such as "Would there be any value in replacing the traditional system call mechanism with a 9P system call service." Also, it's not that collaboration is unheard of in the Plan 9 and Inferno cultures. But part of the shared aesthetic is an appreciation for simplicty and elegance. One result of this is that few particular "things" (driver, application, library, etc) are so complex as to benefit from collaboration as it's typically viewed. One last observation regarding the idea of discussing before doing: A lot of us are old enough to remember the difference between the way the ISO approached networking and the way the IETF approached it. At least from the outside of the process, the ISO approach certainly seemed to be one of discuss and nail down lots of details before ever coding where the IETF had a basic policy that without code there wasn't a lot to discuss. Although we can all pick at certain aspects of TCP/IP, I think most here would say it's better than whatever part of OSI ever became reality. That experience tends to feed the "show us the code" type of response that pops up frequently. BLS