HNHacker News
TopNewBestAskShowJobs

tmarthal

204 karma · joined March 9, 2009

web analytics / math / data / java / python / awesome

los angeles, ca

submissionscomments
tmarthal··on Alaskans vote to legalize marijuana
Actually, I think when Tobacco and Pharmacy companies start to get a piece of the financial incentives in the states where it is legal, you will see their lobbying arm start to influence federal elections on the national level. It is assumed that Marlboro/Camel or whichever tobacco company, could expand their market tremendously with legalized marijuana federal legislation.

I am not an expert (anyone want to chime in?), however it seems right now, the industry does not seem to have any real monopolies and that individual dispensaries (which also have farms?) are making most of the money, rather than any corporations.

tmarthal··on With $30M More in Hand, IFTTT Looks to the Internet of Things
They added a Surfline channel last October, so a whole slew of non-technical surfers have started using it to get surf alerts. Granted, that is just one audience, but it is something that people are using it for.

[Why Surfline doesn't provide alerts directly is another conversation]

tmarthal··on With $30M More in Hand, IFTTT Looks to the Internet of Things
What is interesting to me is that IFTTT is running as a cloud platform, yet some of these devices won't be accessible outside the wireless LAN. I wonder if they are going to have to create a hardware piece, software service or something that communicates to their servers and translates the scripted commands and relevant statuses.

But also, this is a huge market and someone needs to do this.

tmarthal··on Scalable A/B Experiments at Pinterest
Did you guys look into Bayesian Bandits? They seem much more tradeoff-able.
tmarthal··on Why Genius may struggle to annotate the world
Annotations work particularly well in two contexts; first, when explaining the meaning of a single word

It's almost like people do not remember the mid-2000's where the New York times had the dictionary popups on their website. The worst user-interface that ever existed - double clicking (or highlighting? I don't remember) would display a modal popup with the word definition. This happened on every word, so oftentimes they would popup definitions for 'for', 'and', etc. It was highly intrusive (the NY Times removed the behavior after widespread consternation, turned it into a small '?' behavior).

Maybe if everyone expects that behavior all of the time, it would be worthwhile, and people could get used to it. As it stood on a single website, it was not very valuable.

If you are curious, here is an article from 2011 about the types of words that were looked up using the service: http://www.niemanlab.org/2011/07/nytimes-coms-most-looked-up...

tmarthal··on Show HN: Sex, politics, and pay gaps in the U.S
If anyone is interested in a similar basic crossfilters written in d3, you can check out dc.js - https://github.com/dc-js/dc.js/wiki (edit: i have no relation to the project, but it's pretty neat)

Here is a list of community contributed examples: https://github.com/dc-js/dc.js/wiki#examples-contributed-by-... including some geographic visualizations.

tmarthal··on Sixth grader's internet-famous science project may have been plagiarized
For the last eight or so years, I have been a judge at the California State Science fair. At this level, the projects must have won the county or city level (not quite the same for all of them, but it is usually their second or third fair presenting their projects).

There is a a Senior division (for anyone in high school) and a Junior division for everyone younger.

Let me tell you, even though it is usually the kids second or third career fairs, all of the Junior level kids that I have judge have only a very basic idea of the science that they are doing. They are usually very unprepared to talk about their methods or accurately understand the scientific method and what they are measuring. Granted, it is usually the first time talking to a "scientist" with industry or academic experience (rather than elementary school teachers), but usually the conclusions that they draw are rather unlikely. The experiments are never setup properly, they do not measure anything really, they get their ideas off of the internet or out of a "science experiments for science fairs" book (and usually tweak it to win their way to the state level).

This may be different with a professor for a father, but I am skeptical that any child ,even at the high school level, can come up with independent ideas for applicable science experiments.

tmarthal··on One second per second is harder than it sounds
I know that offhand gaming and financial services rely on very accurate clocks at the application level. And you can't necessarily rely on the clients to tell you what their time is, as it can be spoofed.
tmarthal··on An Opinionated Guide to Modern Java, Part 3: Web Development
Actually, it is basically Pivotal versus TypeSafe, not just Spring versus Akka or anything like that.

I put my money on TypeSafe.

Pivotal still (05/2014) doesn't have Java8 support for some of it's frameworks (grails https://jira.grails.org/browse/GRAILS-11063 ). TypeSafe pre-released versions of Play that take full advantage of Java8 features.

Given that if you need to run a Java 1.4.2 app on top of JBoss (or Glassfish!) then the types of technologies that these two companies are pushing are solving problems in a different domain.

And arguably the PermGen change in Java8 would benefit Grails/Groovy the most.

tmarthal··on Moves joins Facebook
Moves.app was free up until November of last year when they released version 2.0. Here are the release notes http://www.appannie.com/apps/ios/app/moves/
tmarthal··on How much would unlimited access to movies, books, magazines, and music cost?
Honestly, the only thing that a local well-funded library does not give you is instant gratification like Netflix streaming. If you are willing to setup a queue and allow most content 2-4 weeks for delivery, you should not have to pay for any music, [hard|soft|comic] books, magazines, or movies on DVD.

Granted, this is also for non-current releases.

Also, you may be underestimating the legal protections that libraries have, and also the voting power (and uncanny monetary resources) of the common library patron.

I am not an expert in the subject, but it would be interesting to see recent legislation for/against library systems (budget cuts notwithstanding).

tmarthal··on Getting started with Django, Pipeline, and S3
Note: there is a great Django template at https://github.com/tommikaikkonen/django-twoscoops-heroku-s3... that will automate the creation and uploading of the s3 assets. It uses `collectstatic` to upload the static resources to be hosted from s3 (which seems to be the more typical way to do this). This template also uses heroku to host the dynamic portion of the app.

You can start your new project with the command (the hyphens break the line, but it should be executed with the settings in place): django-admin.py startproject --template=https://github.com/tommikaikkonen/django-twoscoops-heroku-s3... --extension=py,rst,dotfile,rb --name=Procfile,Gemfile,base.html,404.html PROJECT_NAME

Follow the rest of the README and/or buy the rest of the Two Scoops book (http://twoscoopspress.org/) if you're using django for any sort of large project (and want a standard/sane project layout).

tmarthal··on Multiple Simultaneous Ajax Requests (with one callback) in jQuery
A 'count-down latch' is commonly (at least when I learned it, using pthreads) called a 'Conditional Variable'.

In your case case, your mutex would be the num_results variable and the conditional is that it is greater than 0.

tmarthal··on Writing an MMO with Web Actors
This looks super interesting, a great example of functional programming and an implementation of websockets.

However, whenever something like this comes up, I always wonder about the new Java implementation of Web Sockets / actors. Do you have a comment on the difference between using Akka for your web interface instead of comsat WebActors?

I guess if this is a proof of concept for the comsat classes, it works quite well.

tmarthal··on Facebook Fraud [video]
The Google engineers are also trying to release their methods. This is not detecting 'fraudulent' activity, but rather trying to classify adversarial advertisements: http://research.google.com/pubs/pub37195.html (pdf is linked there). You can assume that they are doing the same types of thing to classify fraud.
tmarthal··on A case for writing unit tests for your blog posts
Everything old is new again. The 'Five Paragraph Essay' is not a 'Decker Grid', nor is it new. http://en.wikipedia.org/wiki/Five_paragraph_essay
tmarthal··on oDesk to merge with Elance
Do you have a network on Twitter? I see people looking (and responding to) freelance inquiries all the time. All based on word-of-mouth and friend-of-friend networks.
tmarthal··on Django plugins you shouldn't start without
I would also suggest checking out the TwoScoops book, very well written best practices. Barring that, just using the template defined in the book to start new projects will get you on the right path to great project organization.

     django-admin.py startproject --template=https://github.com/twoscoops/django-twoscoops-project/zipball/master --extension=py,rst,html PROJECT_NAME
tmarthal··on Healthcare.gov technology stack (infographic)
http://www.techempower.com/benchmarks/

There are a ton of Java frameworks in the top performing benchmarks (See Lng:Jav). Whatever conclusion you've made from Java Swing applications running on top of WindowsXP machines in the early 00's are not applicable to modern web server performance.

tmarthal··on LinkedIn Intro: Doing the Impossible on iOS
If you look at one of LinkedIn's alternate applications, LinkedIn Contacts, http://contacts.linkedin.com/ it actually is a light-weight CRM application. The CRM meaning, it automatically connects to your email and calendar to your LinkedIn account to know when and how you are interfacing with people. I get a daily email with the meetings that I had the day before about who I met, as well as information on their LI profile about the last email conversation I had with them. This is super nice if you meet a bunch of people and need a way to take notes on who they are and what they are doing, independent of their business card.

The contacts application also sends things like reminders for your contacts work anniversaries or when they change positions (something that you can't access in the LI API).

I sometimes think that I shouldn't be giving LI all of this information, but this is a typical case where the benefit received is greater than my privacy concerns.

tmarthal··on Who exactly is crawling my site?
That was almost the original intent behind sitemap.xml, (http://sitemaps.org). Except of course, that they were being used erroneously and changes published when no content was changed, and so the spiders stopped using them as canon.
tmarthal··on Maven is broken by design
Even for strict java builds, I've been using 'sbt' (the Scala Build Tool). It really is the successor to gradle, in many regards. Not sure you've tried it as one of your alternatives, but I'd implore you to give it a try.
tmarthal··on Why have SVG images not yet replaced PNG on the web?
I think that the web is slowly starting to adopt SVGs where it makes sense.

d3.js is bringing svg charting to the web in a major way. Most, if not all, modern web dashboards are using it.

Google Maps has been using SVG icons for custom markers for awhile, embedded into their json configuration. This may not be strictly 'the web', but it is a major component of online GIS applications.

tmarthal··on Amazon Redshift - What You Need To Know
Not only does it use the same datatypes, but I know that it uses the same pg wire format such that the java (and I'm sure python) postgres drivers work out of the box.
tmarthal··on JetBrains releases open-source Python IDE
What entrypoint are you using to run your programs? REPL function calls / test runners / main methods from the CLI / local web server w/browser? I think understanding that matters the most.

If you do a lot of development in the REPL (i.e. exploration), then that is where iPython (in a notebook, or a shell) really is nice. If you don't, then you may not see the benefit.

tmarthal··on How IPython Notebook and Github have changed the way I teach Python
Have you see the Data Science course being presented at Harvard this semester? All of the lectures go through ipython notebooks and they turn in the assignments by submitting their notebooks. The url is http://cs109.org (with the relevant lectures at http://cs109.org/readings.php). Could be a source for additional information.
tmarthal··on Starting an Airline
I think the point is that at a certain level there are only 3 or so defense contractors that have the infrastructure to bid on and win those contracts. Given that they are not "subsidies", but when there are only 2 corporations that are expected to receive the money, then you must call a spade a spade.

If you look at things like the Boeing/Northrop Grumman Tanker contract, then it is hard to fathom that any of the $35B for 180 planes (that's 190M for each plane) wouldn't lead to significant development that could be shared with Boeing's commercial sector.

tmarthal··on Why I'm Getting a PhD
You can definitely get a job as a staff researcher at a national lab or small applied-science company (there are a ton of companies doing DARPA research). Note that due to the way the world is, you may not be able to be the prime scientist on a contract, but that does not stop you from doing basic research via collaboration.

If your contributions become noteworthy in the field, then along with the relationships that you've generated with your DARPA/NSF/? partner, you will eventually be able to be a prime.

To reiterate some of the other comments here, you may not be able to pick the 'basic research' that you want to work on in the above scenarios... usually the grant sponsors decide that.

tmarthal··on Show HN: Easy to use Wikipedia API for Python
I would go one step further and suggest people that need structured queries use the Google BigTable API to query their structured Wikipedia data. Granted, their public dataset is from 2010, so is slightly outdated, but you can write structured SQL against all of the wikipedia article metadata and then use the mediawiki api itself to grab only the article text that you're interested in.

The wikipedia data is hosted here: https://bigquery.cloud.google.com/table/publicdata:samples.w...

Here is a sample query, searching for all articles that start with Positive:

SELECT id,title FROM [publicdata:samples.wikipedia] WHERE (REGEXP_MATCH(title,r'^Positive*')) LIMIT 10

Query complete (2.0s elapsed, 9.13 GB processed

  1|	464347|	Positive airway pressure	 
  2|	10008223|	Positive behavior support	 
  3|	464347|	Positive airway pressure	 
  4|	1354851|	Positivism in Poland	 
  5|	1023857|	Positive set theory	 
  6|	5154273|	Positivism dispute	 
  7|	2871407|	Positivism	 
  8|	17179765|	Positive psychological capital	 
  9|	9033239|	Positive Action Group	 
  10|	4163012|	Positive K
Here is the python API documentation: https://developers.google.com/api-client-library/python/
tmarthal··on Scala School
People are doing all sorts of great things with Scala.

Scala Notebooks, which can be classified as a learning resource https://github.com/n8han/scala-notebook/ - It is an in-work conversion of iPython notebooks using d3 instead of matplotlib.

Another great analysis resource is Saddle, a scala data munging library (analogous to pandas, around the 0.8 revision). http://saddle.github.io/

← PreviousPage 2 of 5Next →