Why I Like Foreign Films

Ever since I got my NetFlix subscription, I've intentionally been looking for films that I probably wouldn't see on TV from day to day. A lot of these movies have been foreign films, such as Amelie, The Motorcycle Diaries, Pans Labyrinth and some documentaries: Zen Buddhism: In Search of Self and Ayurveda: The Art of Being

What I've really enjoyed about these films is not necessarily the raw content being presented. Instead, what I've really appreciated is the sort of cultural window they expose. Coming from the US, I won't for a moment pretend to think that motion picture paints an accurate picture of society, but instead, it's really like looking at a culture through a kaleidoscope.

Because of this, it makes it entirely possible for a film to still be enjoyable to the foreign viewer even if it exhibits the same old boring themes, type-casted actors, and pop culture references that are alarmingly obvious to domestic viewers. With the freedom to bounce around from culture to culture, I feel like the artistic value of films keep their novelty. It's true that many elements of film are transcendent of culture, but I've been consistently surprised about certain little things in each foreign film I've seen.

Beyond the cultural exploration, I've also been enjoying the requirement of reading subtitles. It turns an audio-visual experience into something that is somehow literary. It's easy to let the dialog slip by in a film in your native language, but when experiencing a foreign film, reading the subtitles is often a requirement to understand what is going on, and at least for me, reading makes the lines register more strongly than simply hearing them. To some extent, this might be similar to the experience folks had with silent films back in the early days of motion pictures. I think there is something a little deeper to it though, in which you get to hear and see the emotional expressions of an actor, but also need to read their words to tie it all together.

All in all, watching a foreign film every now and again has sort of refreshed my interest in movies. It's a whole lot harder to guess the entire plot based on what actor is in the movie, or who produced the film. Of course, spend enough time viewing any one culture, and you'll pick up on that pretty quick. Luckily, there are a whole lot of different movies out there!

ICFP Day 2

Yesterday when I wrote my wrap-up of Day 1, we were performance tuning and waiting to get back results from various runs, all to get stuck on a riddle involving 1337 and a cheesy reference to GhostBusters. So although yesterday yielded faster implementations, we were not able to make use of them, and hit an absolute dead end.

Here's some statistics: 869 registered, 357 submitted, we came in at 116 with the same risk and survival chance (1160657,1.2719%) as those from rank 91-164.

This puts us in the top 1/3 of those who submitted, and top 13% of those who registered. Ironically, this is the best our team has ever done, though this sure does not feel like a victory.

For those interested in our implementation, feel free to browse it on the web. Our code may still have room for optimization, but we tuned the hell out of it, and though it is slow, it was working well enough so long as we batched tasks out to many machines.

Anyway, another year passed, another insane challenge attempted with only a modicum of success, but definitely a lot learned and plenty of fun along the way. Warm regards to all those who were involved in the competition, it was a blast yet again.

UPDATE (2007.07.24):

James has put up a short team summary page. I've also fixed permissions on our Trac so you should be able to view the source now...

O'Reilly Articles Released To The Community

I've now released the following four articles under the Creative Commons Attribution-Sharealike 3.0 License:

• Understanding ActiveRecord, Part I
  

• Understanding ActiveRecord, Part II
  

• Rails Testing: Not Just For The Paranoid
  

• How To Build Simple Console Applications With Ruby and ActiveRecord
  

My draft copies are at home, so I'll post them some time later this week, but you're now officially free to do as you wish with these works, providing you follow the very permissive terms of the CC/SA license.

All code samples should be considered released into the Public Domain, which means you may do whatever you wish with them, though attribution would be nice.

I will try to release whatever articles I write under CC/SA when I can. Please keep in mind that O'Reilly contributors sign 30 day exclusive publishing agreements, so if in the future you'd like to do something with one of my articles before 30 days from publishing, I'll have to get permission from my editor (who may need to go get permission from someone else, I don't know). However, Once 30 days are up, I can make decisions about the licensing of the work, and will always err to the side of free documentation unless there is some reason I can't.

I hope that folks find this helpful, and please do share with me any interesting changes you make (translations, additions, code modifications, etc). Enjoy!

ICFP Day 1, Progress!

After an entire morning wrestling with bugs and performance issues all around, the evening brought progress in the form of successfully running dna2rna and rna2ppm on the prefix provided in the contest specifications. This would not have been possible without an execution trace of the Endo

Shortly after getting the prefix provided by the spec working, we got one step farther and uncovered another image that will allow us to continue on. However, our implementations give us estimated times that exceed the competition deadline, so we'll be working today to tune them as much as possible. That means we'll be having fun with ruby-prof and scratching our heads thinking of better algorithms.

I did manage to cut our image rendering time down significantly through some memory-eating optimizations, but it may need to be tuned more before it is ready for real loads.

Anyway, I think any more information than that would be spoilers. We're now on the scoreboard and tied with everyone from position 21 on down (and possibly more, they hide the top 20 scores). Our current rank (team: Side-Effects) is #80. Much better than last year.

ICFP Contest Day -2,-1,0

I'm out in Oklahoma City working on the ICFP Contest. This years task is amazing, and as usual, massive and complex.

So, if you've been talking to me the last couple days, you'll know that was stuck in the Chicago Airport from Wednesday 3pm to Friday 11am due to bad weather and even worse management by American Airlines.

That having been said, aside from being delayed on arrival and travel fatigued, I was still fairly productive yesterday. I now have a mostly working RNA interpreter that is drawing primitives from the RNA, and actually does a few other neat things like do ruby2rna and rna2ppm

The problem of course, is that I lack a complete RNA dump to shake any bugs out of the system and start seeing if we're going to have a performance nightmare. Caleb, Ed, and James have the dna2rna interpreter built, but are optimizing it because our language of choice, Ruby, is once again, far too slow to allow us to be naive.

This morning I'll be testing and optimizing the image processing code from the partial RNA dumps we have and trying to build some more simple tests for it. Hopefully, we'll figure out what's wrong in the dna2rna code, and get a dump later this afternoon. If that happens, we can start saving Endo!

hai friend

actually i am new in ruby on rails
and i involved in that one

my problem is when i installed
gem install libxsl-ruby on windows i got some errors like it need zlib1.lib,z.lib etc lib file

actually my dead line is over pls can u help me asap

Sadly, this is a real email I received this morning. Maybe it's a joke. I hope so for the poster's sake.

On WebTwoOh

It’s a framework that allows a proactive paradigm shift from a synergy of leveraging the long tail to an immersion-enterprise scheme of next gen usability and AJAX compliancy. And a buzzword.

- Ten answers to the un-answerable question: "What is Web 2.0"

PayR getting closer to production

The payroll application I've been developing for BTree for the last couple months is now getting closer to being ready for actual use. It now boils down to bolstering the manager interface a little, reintegrating reports that were part of an earlier iteration, and taking care of some final first wave feature requests.

This isn't the first internal application I've built based on Camping, but it is definitely closer to being a real application than a quick interface for a command line tool. I might be able to get a beta release out on RubyForge by the middle of August if we get this system in production by the end of the month.

It is a bit specialized, but might benefit from some community insight. We're also going to be using this as the core application for fleshing out the Ruport book, so people will have a nice guide to its internals. Especially interesting is that this application contains most of the camping hacks I've listed on O'Reilly, as well as some interesting report running logic.

Anyway, more to come. The above screenshot is of the current main interface, intentionally Spartan. I'm happy that Mike Milner helped me get it to fit in 800x600, though I thought I'd never have to do that again. Turns out not everyone has quite left the "Giant World" from Super Mario Brothers 3 in terms of screen resolution.

Seriously Disturbing Google Security Issue

Okay, so this is really ridiculous. If you saw my last post, it might have been clear that I was investigating a potential security breach, but I thought it was an isolated occurrence.

Consider the following, and then read on:

1. Google is single sign on. Web History, Gmail, Blogger, Groups, all use the same account for verification
2. You can't just force all of your sessions to log out. The individual session stay logged in seemingly as long as needed, depending on browser setting. This means that when you sign out once, it doesn't sign you out everywhere.
3. Changing your password does not force existing sessions to log out.

Turns out, there have already been several reports of the issue that I had. I was using Google Groups on the 29th, and all of a sudden, my email address swapped to someone elses. I can't remember the name exactly, khalamas or something like that.

I had to log out, and by the time I logged back in, I was told that my invitation to "Patriots For Conservative Values" had been accepted. I was added to the group as 'Redhawk', instead of Gregory Brown. I did not sign up for this group, and it wasn't the same email notification as when you are added directly by a group manager.

After several rounds of confusing conversation, the group there seemed to not know what the hell I was talking about, saying that they just thought that I was some well known person that goes under that handle from gop.com

They were fairly helpful all in all, but I didn't know what to believe seeing as this seemed incredulous to me.

I've gone and changed a ton of passwords, made full backups of my email, offshored many of my services to a different email address, and done as much damage control as I could.

Le sigh. What is worse, is that google doesn't log you out of all systems when you log out. This is usually a feature of sorts, but means that if a session is highjacked, who knows how long it will stay alive. :-/

If it's not clear automatically, the fact that google is single sign on means that depending on how long this lasted, email, web search, and all of that other stuff may have been compromised. That's really what I'm worried about, and will post details if I find out more.

UPDATE: A blog entry from a user with a similar experience, who got in touch with Google. They said they recently made some changes to fix the issue, but no more details are there.

UPDATE2: Looks like we have even more reports of this issue happening that I overlooked. Sighs.

UPDATE3: I've received word from Google employees that this was a known Google Groups issue, and has since been reverted. I've asked a few additional questions for risk assessment, such as how long the sessions were live, but haven't heard back yet. At the very least, it doesn't appear to have been an attack.

Friends with security experience or interests, please contact me

Please get in touch with me if you'd like to help me explore a security issue I ran into. Use something like Freenode though, don't email me.

Sorry to be clandestine, I need some advice first. Most of my Ruby friends would be ideal candidates.