As far as I can tell, author is talking about FM-Index. It compresses the search data into a much smaller index memory footprint. I tried using it few times, but never figured out how to use it as a key-value data store.
If anybody is interested, here is the code: http://pizzachili.di.unipi.it/indexes/FM-indexV2/fmindexV2.t...
I guess FM index is just not the right thing to use when you need a key-value data store. It's a full text index -- a data structure, which allows fast substring queries over a fixed text corpus.
I just want to clarify something here, I am not an idea guy or a business guy. I am a hacker but my specialties are systems programming in C/C++ and I am learning web dev on the side when I get time, but haven't gone that much far. The reason for this post was partially to motivate me as well, since its great motivator once you get something going.
Didn't understand how the sentiment analysis works, since results look quite not what I expected. I did a test search on 2012 presidential candidates:
Obama: 44% positive ("barack obama" is 62% though)
Mitt Romney: 67% positive
Michele Bachmann: 50% positive
Herman Cain: 71% positive (?)
Ron Paul: 71% positive
Rick Perry: 81% positive (??)
I am not sure the sentiment calculation is just based on Positive/Negative dictionary comparison or something else. Is there a logical explanation of above percentages? Do results vary for other people?
So currently this is based on taking a sizable rolling window of textual posts for each topic then running it through our synt library (http://github.com/tawlk/synt) which does sentiment classification with a NaiveBayes classifier trained against iterations of a couple million samples. This setup averages about 80% accuracy against new labeled sample sets.
We however only have the server resources to have a large rolling window for a limited number of topics (currently the ones in the side bar).
When you searched for "Barack Obama", it based the sentiment score on just what your browser collects live on the fly, whereas when you searched "obama" you got sentiment calculations based on our server collected rolling data set which is far more stable.
Both are fairly accurate, however they differ in how large of a window they are averaging against.
As topics become more popular and as server limitations allow we automatically migrate them server side for more aggressive collection to provide more reliable sentiment.
We also only provide scores and reach assessments for topics we posses enough server-side data for to justify it.
One particular company I was interested in had few puzzles on their website. I once worked the whole weekend to solve them as good as I can. Spent lot of time writing a custom cover letter, resume and attached the C++ solutions to the puzzles.
Its been several months and I am still waiting for the damn reply!
The site looks great and very user friendly, but as others have pointed out, the paid-tier-only seems too restrictive. Specially given the motto of treehouse is "Millions of people can't afford a quality technical education, or if they can, it's out of date immediately when they graduate. We aim to change that."
How can this compare to MIT (and other) Open Course Work and hundreds if not thousands of good free tutorials on the internet Vs $25/$49 a month?
FWIW, I learned basics of web development/design totally free, just took a bit of searching to find good resources, and once I was confortable with basics, the further education came through just googling-on-demand, videos only helped in the beginning.
I too learned Dev and Design through free resources and googling (It's still how I learn), but a lot of people find value in having a path laid out to them, and having consistency in the resources they learn from.
Of course a lot of people on HN are going to be people who taught/teach themselves everything, but a lot of people want some hand holding, and that's what we aim to provide with Treehouse.
Interesting that you say a lot of people want hand-holding - I agree.
But don't you think in an industry like software development, you need to be able to self-educate. Software just moves too fast. So a person that needs hand-holding is not going to "get" what it takes to be a good programmer.
The article says that over the next 5 years, about 7 Million laptops will be purchased. The price of windows mentioned in the article is about $110 at current rates, which means $770 million order for MS. That definitely is a big order for MS, considering the full laptop is only expected to cost ~$300.
I really can't see any good reason to spend that money to put windows in laptops for poor children. It would be much wiser to use linux and buy more laptops.
Even now, RedHat Linux is used in local government offices. Wondering what happened now, for switching to Windows 7 for freebie laptops. Huge drain on the exchequer.
No way this is right. $100 for Windows 7 Starter Edition? Not a chance. W7 Starter probably wouldn't cost that much retail, and there's no way OEMs are paying that much. I've read that HP/Dell/etc pay more like $25-$30 for Home
Premium, which is several tiers above Starter.
Just so you know, I work for Microsoft, but not on Windows. I have no inside knowledge of Windows OEM pricing, but I'd be very surprised if Starter cost an OEM more than $5-$10.
the last place i worked paid $35 per desktop for office, so if they're actually paying $110 for windows, i'd suspect there's a kickback happening at the OEM level, not unheard of in many countries.
Not sure if it'll add more credibility to the claim, but interestingly the wikipedia page for Anakam doesn't talk about Founders. And the company's official leadership page is empty!
On a slightly different note, I don't remember top 5 posts on HN all about deaths! I wonder why so many people upvote so many stories related to same subject (i.e. Steve Jobs' death)?
You've done some nice work on the project. Since I follow most showHN posts, I think I am seeing this submission 4-5th time in last 1 and a half month with different titles. I tried to search for previous submissions but seems like they were taken down later.
Please don't get me wrong, I wish best of luck to this startup, but would you really still call it a 'weekend project'? I think not.
Yes, this has been submitted a few times. And I'm finally getting the useful feedback that I respect HN readers for. Maybe what you see as spam is what I see as persistence. But I think a lot of getting good feedback from HN is submitting at the right time with the right title so your submission doesn't die on the vine so to speak.
When I learned to use vim years ago, one interesting observation was that the biggest thing to learn in vim is efficient navigation.
i.e. if you know how to get to a particular logically defined place in the file, you will always be able to carry out copy-paste-delete to and from that place to where to cursor is.
One would be surprised how easy thing become once you know the basic navigation rules due to marvelous mix-n-match nature of vim operations.
