cs.SE in nondescript beta style
Since my post in December, Computer Science Stack Exchange has gone through private beta and is now accessible for everybody. Head over to cs.stackexchange.com and take a look! We are waiting for your questions so do not hesitate to post anything you have always wanted to know in computer science, be it theory, applied or even practical!
We also need more expertise: out off almost 190 questions sixteen have yet to receive a good answer. among those are one about equivalence of Büchi automata and linear μ-calculus and another about efficiently learning regular languages.
You might also want to check out our hottest questions, including why Quicksort is often considered the best sorting algorithm, encryption using NP-hard problems and connections between Gödel’s incompleteness theorem and the halting problem.
I am very excited about this site. One one hand, it gives me the opportunity to share my knowledge in computer science with a wide variety of people, and I like teaching. On the other hand, I get to learn a lot. Not only do people ask about things I know next to nothing about so I can learn from the answers, but there are also original concepts. My favorite is Patrick’s proposal of heap automata; I have spent several hours contemplating his question about their power and follow-up questions.
I hope to see you on cs.SE soon!
I have posted about Stackexchange before. Helpful folks on various sites have solved many a problem I have had since then. I have posted a bunch of answers which—hopefully—helped out others in return. To say the least, I am convinced of the StackExchange model and am continually amazed at its effectiveness.
The wealth of knowledge saved on the StackExchange network has become so extensive that if you google for programming or LaTeX related problem almost certainly a question on Stackoverflow or tex.SE comes up. If not, asking there is often faster than searching the webs and/or trying around. Especially on tex.SE you can expect great answers in a matter of hours.
Now an exciting thing has happened: A proposal for a computer science site has reached commitment phase! That means that a number of people have to vote for the site in order to move it to a beta phase after which it will be a full-fledged member of the network. The new site is to complement Stackoverflow and its derivates on one and cstheory.SE on the other side, filling the massive gap in between. I am very excited about this; if the community on cs.SE can be only half as good as on some other sites we are about to create the best resource for computer science students, researchers and users the web has seen so far.
But we are not there yet! First we have to have enough people commit to using the new site, then we need a successful beta. If you want this to happen, head over to area51, commit and be part of the community from the start!
When I started roaming the internet back in 2003 most communities I visited were using discussion boards. For discussions they were perfectly fine and are, in my opinion, still unrivaled. But they were of course also used for sharing trivialities and problem solving. While the former works well enough, the latter does not work well on boards at all. Assume you post a question and receive a number of disagreeing answers from experts that then argue in a language that you can not understand at all. How do you find out which answer is the best? How do you keep track of the pages-long discussion?
StackExchange provides a very good alternative for this. Here, both questions and answers can be down- and upvoted, thus creating a measure of confidence the community has in a particular statement. You can change your vote later if a discussion in the comments changes your mind or answers are edited. Futhermore, answers are sorted by score, not date. Therefore, you can visually observe convergence towards a meaningful answer assuming that at least some people around know what they are talking about.
A reputation based reward system keeps people behaving; you earn reputation by getting others to upvote what you write and lose it by getting downvoted or downvoting yourself. You can also set a bounty on questions whose answers you are interested in and hope that somebody is motivated enough by some reputation points to take the time to answer. The influence of trolls and spammers is kept minimal since many features (e.g. closing, retagging) are only available if you have certain amount of reputation; the more impact the feature has the more reputation you need. This is necessary because there are no moderators or the like; each community manages its site on its own.
The most notable subsite is Stack Overflow which has been around quite a while, but a number of other sections has come up lately. They cover a broad range of topics from Cooking and Bicycles over TeX – LaTeX and Ubuntu to Mathematics and Theoretical Computer Science. Check it out; even if you do not ask questions yourself, reading and answering those that match you interests can be both interesting and entertaining.
Due to a weird idea for a project that would include uploading many videos I checked out how subtitles are supported by different video hosters. The results were rather disappointing; if support is implemented only the original uploader can provide the source files. This causes a lot of work even if the crowd provides subtitle files. Also, there is apparently not a superior solution for creating subtitle files, let alone with videos on the web.
Then, I stumbled on Universal Subtitles. It is a Mozilla powered project still in its infancy that aims at providing a subtitle service that is independent from video hosters but supports them all. Everyone who watches a video can add or correct subtitles in his native language(s) directly in the browser. I love this concept and had to try it. I subtitled this video in German:
Try for yourselves! It is not entirely bugfree and definitely resource-hungry but for a beta, I am impressed!
Neulich bin ich auf Ubuntu Brainstorm gestoßen. Die Plattform ermöglicht es jedem Nutzer, seine Wünsche und Ideen für sowohl die Distribution als Ganzes als auch einzelne Programme schnell und einfach gesammelt zu publizieren. Es wird immer zunächst ein Problem definiert, um dann Lösungen dazu anzubieten. Andere Nutzer können diese Lösungen dann bewerten und auch eigene Vorschläge einbringen. Mit knappen 15000 Ideen und fast 100000 Kommentaren ist doch schon ein reichhaltiger Stoß an Material zusammengekommen.
Ich halte dieses System als Ergänzung zu Bugtrackern für eine sehr gute Möglichkeit, sich auch als Nichtentwickler in den Entwicklungsprozess von Ubuntu einzubringen. Außerdem dürften viele Ideen, die man hier in einer klar definierten Form statt, wie üblich, in losen Foren- oder Blogeinträgen vorfindet, wertvolles Feedback, aber auch kreativen Input für die tatsächlichen Entwickler darstellen – gerade weil sich auch (und vor allem?) solche Leute einbringen, die eben nicht coden.