On Tue, 2012-02-28 at 01:35 -0600, Devin Walters wrote: > One item that hasn't made the project ideas list that I've seen > numerous threads about is documentation. Does this fall within the > scope of GSoC? > > > It seems like there are a lot of opportunities to either organize, > revise, update, or generate documentation. > > > Some ideas: > - Clojure.org's Libraries section still talks about contrib like it's > first class. > - The Getting Started guide could always use more work. > - StackOverflow contains nuggets of wisdom that aren't anywhere in > official documentation. (It also contains a lot of bad answers, but > still…) > - I've heard it said on more than one occasion that xyz docstring is > out of date. > - This is one of the few communities where you can go back to 2008 and > read a transcript of a conversation between Chouser and Rich about why > map destructuring is the way it is. Some of these conversations hold > some deep wisdom about Why Things Are The Way They Are. > - This list contains truckloads of information that could be organized > for more efficient consumption. > - ClojureScript wouldn't be hurt by more documentation. > - Without making this a laundry list I'd just say: Producing and > organizing good documentation is hard labor, but it is also something > that I think benefits the entire community. Moreover, it might give > someone a chance to learn a ton about Clojure over the course of a > summer, and make it easier on everyone who decides to try out Clojure > in the future as a nice side effect. I'd like to suggest we add an > intentionally vague option to "Make Lots of Things Better" and list > some ideas for how one might go about doing that.
I'd mentor someone willing to do work on the literate programming version. See http://daly.axiom-developer.org/clojure.pdf http://daly.axiom-developer.org/clojure.pamphlet (src) Even more interesting... It appears that the ePub standard allows embedded javascript. So ideally we would like to manipulate a canvas to show the ideas. For instance, I'd like to see a digital book that had a canvas element that showed the red-black-trie evolve, potentially interactively. Even more interesting... Write the generated javascript for the above ePub demonstration using ClojureScript I'd be willing to mentor any of those. > > > More ideas that might bear interesting and desirable fruit: > - Make an album with Overtone. (Kidding (but only a little bit (not > kidding at all, actually (I bet we'd get some passionate proposals > (and maybe even a record deal ;))))) I watched the overtone video shortly after finishing both the stanford machine learning course, the signals course and a genetics course. It would be possible to extract features from your favorite songs e.g ( http://www.ams.org/notices/200903/rtx090300356p.pdf ) using FFT signal processing. (signals) Use the overtone feature set to define the possible features. (overtone). It would be possible to rank the features of your favorite songs by listening to each song and constructing a "like" value for each song (e.g. hit the + key multiple times, or use a number to rate the song from 1 to 11 (ala spinal tap). (machine learning) Having ranked the songs, use the learning algorithm to predict the kinds of songs you like based on features. Use overtone to generate new songs with the most popular features (overtone). Use vector crossovers to generate new songs. (genetic programming). Rinse and repeat. Sort of a "sample" without samples :-) > - The sidebar on the left of the GSoC page lists an opening for a > Community Manager Internship. I think a lot of what I'm suggesting > falls under that umbrella. "creating/editing documentation, helping > migrate projects to newer versions of clojure, developing sample > applications such as solutions for the alioth benchmarks, answering > questions on IRC, administering/maintaing clojure.org, clojure.com, > assemble, confluence, mycroft, etc." > > > I guess what I'm saying is, at the end of the day: Let's add > documentation to the list, but also add some other obviously fun > projects and see what kind of proposals we receive. It doesn't mean we > need to accept them, it just shows (IMO) we're very open minded about > people who are passionate about building what /they/ care about, not > necessarily what we care about. If some musician in grad school > submitted a proposal to make an album exclusively with Overtone and > published the source that would be a boon to the Overtone project IMO. > If a sophomore in college wants to build some crazy parallelized Rube > Goldberg machine with Clojure then I think we should at least > entertain the idea of it. More than anything, I think we need to > present the people who *might* do something like that with the face of > a community that would genuinely appreciate it. I've met many of you > personally, so I hardly think that's a stretch for us. > > > This is getting really long so I apologize, but I'd like to offer up a > bit of personal experience w/r/t GSoC: > I did GSoC years ago for Plan9 (Inferno-OS specifically). I was not > very familiar with their community, and I doubt many people have ever > read a book about programming Limbo. As a result, a lot of the ideas > that were listed were strangely specific from my limited undergrad > perspective. I was interested in learning about Plan9 and > contributing, not necessarily learning Plan9 to make a distributed > authentication system that someone else wanted for reasons that were > unknown to me and/or were not well described in the description. As a > result, keep in mind that we will potentially have people submitting > proposals to write Skynet 1.0 in 3 months who are doing their > undergrad and may have only just had an introduction to lisp or > scheme. Last note (I promise) is: potential mentors, this is not a > small commitment. Trust me on that. It's as much your responsibility > to steer someone toward success as it is theirs. I'm actually working on the ePub idea with a book so the "mentoring" would actually be more of a collaborative effort since it travels into what is, for me, new territory. There are still some things to demonstrate and a lot of reading of the ePub3 standard but it progresses slowly. An ePub book with embedded interactive canvas elements seems to be the best path to inspire people to write literate software. Tim Daly d...@axiom-developer.org -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
