ntoshev

77 ms·

691.
by ntoshev 18y ago
Can you give an example of something you can implement with a Lisp macro, but you can't implement with Haskell?
692.
by ntoshev 18y ago
You can keep all your data in RAM, replicate it for fault tolerance (still in memory, but on another machine, another rack, another datacenter) and just dump the entire thing on disk once in a while which would be fast.
693.
by ntoshev 18y ago
I don't think Google is that good with integrating all the data they have. For example they still show me Google Chrome ads on pages I open with the Chrome browser. I also doubt Google AdSense would show me ads based on the contents of the
694.
by ntoshev 18y ago
I don't think a formal system with symbolic inference is useful for describing the knowledge of any reasonably complex domain that doesn't have a mathematical model. And most of the human knowledge tends to be like this. I'd love to be prov
695.
by ntoshev 18y ago
Java & Javascript http://en.wikipedia.org/wiki/Gmail
696.
by ntoshev 18y ago
TV/Movies talent is a winner-take-all business right now. If people start making cheaper productions the talent payment tail would get heavier. And/or talent would be payed with equity.
697.
by ntoshev 18y ago
I would approach this problem differently: tokenize the text and diff the stream of tokens (as opposed to a stream of characters).
698.
by ntoshev 18y ago
This kind of storage should work very well with SSDs instead of disks. Is this the case?
699.
by ntoshev 18y ago
Nobody knows a good way to measure this and you know it. Just because we can't measure it doesn't mean it is not real (I'm not claiming either way).
700.
by ntoshev 18y ago
I guess no one has come up with a better way. If you introduce a new system, it has to be at least as good as what we have now: no one would want to sacrifice relevant results. It is wrong to think about pagerank as a single algorithm: it i
701.
by ntoshev 18y ago
Google's ranking blends together hundreds of variables, just one of which is pagerank. If it is so prominent, there is a reason for this: it is effective in delivering relevant results.
702.
by ntoshev 18y ago
I don't think "make a product for yourself" captures the entire philosophy behind 37signals' products. A better discussion of how to copy can be found here (it is about Picasso's famous phrase "good artist copy, great artist steal"): http:
703.
by ntoshev 18y ago
97 conventional wisdom recipes in an awful, screaming font, linking to explanations. I didn't read them all so maybe there are some gems, but I highly doubt it.
704.
by ntoshev 18y ago
It's interesting to compare this to the "google is reading my email!" reaction they got for showing ads in gmail. People get easily upset on stupid stuff like this, but it doesn't mean the company should ignore it.
705.
by ntoshev 18y ago
I submitted a similar poll a while ago. Results here: http://news.ycombinator.com/item?id=198272
706.
by ntoshev 18y ago
You are probably joking, but I will repeat anyway: I asked about usability benefits.
707.
by ntoshev 18y ago
What's the usability benefit over simply having an iPhone friendly web UI?
708.
by ntoshev 18y ago
If you treat the number of ratings as an indicator of quality, you need to account for the time the item has been available for rating as well. Otherwise you underestimate newer items.
709.
by ntoshev 18y ago
A term like "lookup engine" is certainly not commonly used in the field of information retrieval: http://www.google.bg/search?q="information+retrieval"+"looku...
710.
by ntoshev 18y ago
Why get the lower bound: this means you systematically underestimate the items with fewer ratings. Also this formula assumes normal distribution. There is another solution called 'True Bayesian Average' that is used on IMDB.com, for example
711.
by ntoshev 18y ago
I'm looking at the source code right now, and it seems Whoosh does use Porter stemmer to include word forms for English. Not sure yet if it weighs exact matches heavier. EDIT: actually it uses query expansion for word forms for English. Too
712.
by ntoshev 18y ago
This story has been on HN before ~3 months (of course it is still news that the patient is well and without any trace of HIV): http://news.ycombinator.com/item?id=357144 Also see an earlier speculation by a HN reader that such a cure for
713.
Picasso: Good Artists Copy; Great Artists Steal(businessofdesignonline.com)
2 points by ntoshev 18y ago|0 comments
714.
by ntoshev 18y ago
You are correct, as long as you don't get too many arguments.
715.
by ntoshev 18y ago
There are many ways to write a small piece of software so that it conforms to certain specs. Rules, even if arbitrary, tell a developer what is a "better way" and reduces the doubts he has about how to write that code. So developers, especi
716.
by ntoshev 18y ago
I disagree. Obsessiveness with small methods means less opportunity to keep your transient state in local variables, meaning you are going to make it object state (use object fields), which would actually increase your side effects. Edit:
717.
by ntoshev 18y ago
Thanks a lot, this is pretty much what I wanted to know. Looks like there is not too much overhead and the results are amazing.
718.
by ntoshev 18y ago
These are automated functional tests. It would be interesting to know if the author uses unit test too and how he writes tests in general. Also some code:test ratio metrics would be indicative of how much weight does all this testing add.
719.
by ntoshev 18y ago
In my experience most often the public interface is an instance of the Facade pattern and its state is spilled in an object graph behind it.
720.
by ntoshev 18y ago
Most people treat packages just like bigger-than-classes chunks of related functionality and state. Typically there is a 'common' package for supporting stuff that is used all over the place, but these are not the data types most important

More ›