On Tue, Jan 3, 2017 at 9:29 AM, Juliusz Chroboczek <j...@irif.fr> wrote: >> Well, the rabel repo exists for the crazy ideas... and unfortunately >> testing them at scale tends towards being a PITA. > > Exactly my point. Ideas are cheap. > > -- Juliusz
Ideas are not cheap. Most people have few of them. I average a few good ideas a year, and one great one a decade. If I didn't write the ones down that pass my internal tests for "don't know", I wouldn't know which ones were which, to even start towards the next phase. It helps a lot to find one person that thinks your ideas are not crazy, but finding that one person is hard. Bouncing them off of people is one way to winnow down what is terrible, decent, good, or great. Often the best ideas are met with intense rejection initially. If one has enough self confidence to take feedback and rework the idea from decent to good, it helps. Far too many people are stopped cold by that initial rejection or indifference. ... Implementations are hard. It helps to have established that something is a good idea before trying... and sometimes it's ok to just bull ahead blindly in the hope that by merely trying you'll end up somewhere unexpected. Prototype stuff, even a little, and get stuck, but make sure you can go back to the work later, leads to incremental progress over years, sometimes... And: Why climb a mountain? "Because it's there." Sometimes an idea requires a confluence of other ideas to become feasible, or requires an accident to happen... Here's the story of one: http://the-edge.blogspot.de/2003/06/wireless-connection.html I (theoretically) live next door to Edison's estate in florida - where he is famous for trying 3000 different kinds of filaments in a light bulb - but what is less well known is the decades he and Ford spent looking for an alternative to South American rubber, and failing. (he nearly went broke, trying). I wrestle with big problems[1], batting at the walls of the cage until something gives, somewhere, that makes for a better world. (and going broke, trying) ... There's a concept called a "brainstorming session". This is what I tried to kick off at the start of this thread was that, a few folk rose to the challenge... and my follow-on message recently was asking if anyone had ideas for a good experiment on the two features recently added that you would like data on. I have an issue in that *everything* has to work in order to get anywhere in networking, and a zillion different features all have to work together in order for the whole to work *at all*. It doesn't matter how perfect theoretically one component is when something else is incorrect or malfunctioning... This week I'm - after exposing and helping fix a dozen+ bugs in the stack over the last 9 months, while adding one new major feature) - trying to get to doing that all-up test - restructuring tests from what I learned from the last set of experiments with what is hopefully a near final version of the kernel and lede.... Did we get them all? Is it going to work? What else broke? Did we win? I will end up getting stuck on fixing other bugs than the ones I set out to fix in order to have something that barely works. I don't mind that you sometimes react badly to my chaos, but I outlined some of the problems I was having in my deployment at the start, wrote down some ideas for fix-ing 'em, shared em, and sat back for feedback, and ideas on other fronts, and got some. With that in hand I can go build up the testbed again, and see what blows up this time around, all over the stack, not just in babel. there are always too many unknown unknowns, my goals are generally to find those and reduce them to knowns. If you'd like me to explicitly tag a message "[CRAZY]" or "[BRAINSTORM]" so you needn't be subjected to it, on this list, I'll gladly do so. [1] http://the-edge.taht.net/post/gilmores_list/ -- Dave Täht Let's go make home routers and wifi faster! With better software! http://blog.cerowrt.org _______________________________________________ Babel-users mailing list Babel-users@lists.alioth.debian.org http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/babel-users