Measuring Measures

Any chance of getting back contents of Bradford's old blog http://measuringmeasures.com/ ? There were a few brilliant articles on becoming data scientist, learning statistics etc. It would be much appreciated.

You're speaking to where we want to see all of this go. There's no easy answers, but you might want to check out: http://techcrunch.com/2009/11/23/flightcaster-flight-delay/ http://eaves.ca/2011/03/07/launching-an-open-data-business-recollect-n… http://measuringmeasures.com/blog/2010/7/2/research-driven-startups.ht… I have also ordered Sarah Lacy's new book, about info entrepreneurs in the developing world: http://www.amazon.ca/Brilliant-Crazy-Cocky-Entrepreneurs-Profit/dp/047…

I came to this thread to check if Bradford replied. Not yet. Hopefully he will pitch in. The uninitiated will benefit from digging into his blog: http://measuringmeasures.com/ Also, the new ML Q&A site: http://metaoptimize.com/

Bradford cross who does machine learning work at flightcaster addresses this question in the following post: http://measuringmeasures.com/blog/2010/3/12/learning-about-machine-lea…

Linear Algebra by Axler Matrix Computations by Golub Collaborative filtering papers by Koren: http://research.yahoo.com/Yehuda_Koren Stanford paper which is the source of Mahout project: http://www.cs.stanford.edu/people/ang//papers/nips06-mapreducemulticor… Map Reduce algos book by Jimmy Lin: http://www.umiacs.umd.edu/~jimmylin/book.html Pattern Recognition by Bishop (or Theodoridis) Linear algebra by G.Strang Graph Theory by Gross Hadoop the definitive guide by Tom White Natural Language Processing by Manning Information Retrieval by Manning for additional references here is a great compilation: http://measuringmeasures.com/blog/2010/3/12/learning-about-machine-lea…

It's kinda like the firehose drink, the language and compiler are changing rapidly, but I think what's described in Stuart Halloway's (excellent) book is still applicable, except for some very small number of bugfixes and breaking changes in 1.1 and 1.2. The next release of Fogus/Houser's draft "Joy of Clojure" book should be butt-kicking, don't know when that's scheduled tho. Also, installing swank, slime and paredit, and getting one/more of leiningen/ant/maven to work seem to consume a lot of bandwidth (this might be one of the times that the MS windows install is more straightforward). just saw this for Debian: http://measuringmeasures.blogspot.com/2010/01/my-new-clojure-emacs-set… the IRC and google group are very friendly, ask questions!

Bradford sorry for dismissing your post so fast but I don't agree with the way you try to conciliate the two approaches. Those are two different problems at different levels. Minimum Viable Product is a principle to follow at a tactical level to drive the development. The strong vision you refer is required at a strategic level. Depending on your role you should rely more on one or the other but trying to conciliate the two may just create more ambiguity because they guide decisions at different time-frames. A product manager without an holistic vision will not understand the drawbacks in adding yet another customer requested feature. A team lead focused on the strong vision will try to integrate the new feature with the whole system, slowing development cycles. The two principles are fundamental for developing a great product but they define the mission of different parts of the system. By the way, I really liked your blog. I don't know if it's new here on HN but it's new for me. http://measuringmeasures.com/blog/2010/6/3/software-product-granularit… Wow, love to see how you explained, it really ressonated with me. Yesterday in the end of the scrum sprint retrospective a big discussion arose where this would have been helpfull. Liked this as well: http://measuringmeasures.com/blog/2009/12/2/the-decline-of-the-agile-c…

"Is there a job site targeted at skilled international workers like you?" As far as I know there isn't. Most people are (rightly) trying to outsource just the grungy bits for which they (rightly) seek low cost folks through elance etc. "How do you look for work, and how would people find you?" I can't answer the general question but as for me specifically, I have a bit of a unique skillset (AI/ML/compilers, reasonably strong math skills), which, while not in wide demand in Bangalore/India is sufficiently compelling for the few people who do want it (e.g some folks in the local Defence Research labs, telecom companies, financial trading folks and so on). More importantly, over the years I've made a bit of a reputation for straight dealing and delivery on time/below budget. I have never actively looked for work since I left ThoughtWorks (6 years ago now , yikes! where did the time go?). And I left TW because I was tired of working on enterprise sw. The company and my co workers were great - the work was terrible. I have enough work for the next five years. Research + coding projects tend to be like that. I constantly turn down projects. So far so good. The problem is that the "rare skillset" bit cuts both ways - I've been working alone for the last 6 years and would love to work with great people (ideally much smarter than I am) again. That part doesn't seem to work out so well hereabouts. There are some interesting startups coming up but still very early days…

reminds me of Research-Driven Startups by Brad Cross http://measuringmeasures.com/blog/2010/7/2/research-driven-startups.ht…

I have this: http://www.amazon.com/Introductory-Graph-Theory-Gary-Chartrand/dp/0486… and quite enjoy it, although it's fairly beginner if you're a math person. There's also a more thorough list by a real pro under the 'network theory' section here: http://measuringmeasures.com/blog/2010/3/12/learning-about-machine-lea…

You are right about the main point. When I first started looking into this topic, I thought it was that simple. "You just have to tell the world about your work," I told myself. Then my friend Ross laid the smack down on me: http://measuringmeasures.blogspot.com/2009/12/build-brands-with-luck-a… If you read through Hamming's statements, and my reflections on them, you can see that it is not as simple as just promoting - it is specific recipes about how to do so effectively, and how to avoid certain mistakes.

Correct link: http://measuringmeasures.blogspot.com/2010/01/you-should-waste-50-of-y…