Allo à tous mon jeudi!

Bon comment changer le monde, ben faudrait que je puisse faire une conférence TED

Bon c’est pas vrai que je vais passer aux conférence TED… c’est un but… un rêve…..

Imaginez la Vilaine charrue qui donne une conférence au très select club TED.  D’abords faudrait que je trouve mon sujet et comme je l’ai pas encore trouvé et que je cherche, je vais me contenter de voir ce que les gens font, ce qu’ils pensent!  C’est vraiment magnifique ce que les gens ont a dire, comment à partir d’un simple concept, association d’idée ils présentent des idées révolutionnaires! Imaginer que je trouve l’idée, imaginez que cette idée fait que vous aussi vous pouvez la promouvoir, que cette idée changerait le monde, votre monde… Vous aimeriez que cette innovation soit bonne pour tous, quel soit exceptionnelle!   La Vilaine Charrue Innove en lançant l’idée de…. et tous les gens qui la suivent ont découvert une nouvelle façon de penser de faire de créer… Wow!..

Vous ne seriez plus les spectateurs mais les gens qui changent l’histoire parce que vous le faire, vous l’utilisez vous le créer… Sérieux, c’est possible? ben trouver l’idée c’est pas évident! Je suis certain que c’est possible! La créativité c’est la ressource que nous avons tous! comment l’utiliser, la mettre à profit de tous! pour améliorer le sors de tous?  C’esdt possible, mais faut chercher, cela n’apparait pas par miracle..

Je crois que les bonnes idées sont nées des interactions et de combinaisons de plusieurs variables dans un contexte précis.  La seule différence c’est d’avoir pris la décision de le faire, de partir quelque part…. Comme dans toutes choses, la décision est cruciale. Décider c’est la base du changement…

Changer quoi! tout est parfait autour de vous? Il faut être en mesure de définir, d’observer…..En tout cas, je continue et on verra!

Faites la même chose, changer quoi autour de vous… Y’a rien a changer? possible…..

Pour les curieux, avec Moto-Men, ça va bien et je continue… Je crois qu’on va se voir ce week-end…. Kool.

Pour Peggy, ben y’a pas la foule de gars sexy que vous m’envoyez… Si ojn pouvait en trouver au moins 5, ce serait cool, allez chercher dans facebook et laisser un commentaire avec la page du gars célibataire en question…

Soyez Heureux et gâtez-vous!

Tribal leaders and ordinary people across Afghanistan said they would cast aside dismay over the fraud that marred their presidential election to participate in a second round run-off.

US soldiers return fire on enemy positions as their platoon came under intense fire during an ambush at the village of Lanyal in the Korengal Valley, Kunar Province, eastern Afghanistan on October 20, 2009.

Afghans will go to the polls on November 7 to vote in a run-off between the two top contenders for president after fraud allegations cut the lead of President Hamid Karzai to below the 50 percent needed for victory.

Karzai, who took 49.67 percent of the vote in the first round after nearly 1.3 million ballots from 210 polling stations were discounted, will face his former foreign minister, Abdullah Abdullah, who took 30.59 percent.

Tribal elders in the troubled south of the country said they were angry that the risks taken to vote in the first round, in the midst of a vicious Taliban intimidation campaign, had been wasted.

Nevertheless, the elders in Kandahar — traditional stronghold of the Taliban and Karzai’s mainly-Pashtun powerbase — said they would vote again to make the risks worthwhile.

“We did vote, risking rocket attacks, threats and intimidation, and now that the election is going to a second round for whatever reason, we the elders of Kandahar, have decided to vote again,” Shah Aka told AFP.

“After the president’s appeal to the nation and the announcement of the IEC (Independent Election Commission) we will go and vote in the run-off and will support Karzai. We will strongly back him.”

Tribal elder Nor Wali Khan Shinwari added: “The run-off showed there is no fraud-free and transparent phenomenon in Afghanistan and now we see fraud even in a national process to determine the nation’s fate and future.”

Shinwari, from eastern Nangarhar province, said democracy was still in its infancy in the war-ravaged country and democratic values were being “played with”.

“Indeed we will go and vote again but who is going to guarantee that no fraud will take place again? People did vote but it was the election commission officials who commited the fraud, not the people,” he added.

Mohammad Sharif, 45, believed the run-off would strengthen new democratic practices.

“It also shows our politicians believe in democracy and are ready to accept the results of the election even if it is not the result of their choice,” Sharif said.

The second-round announcement has changed the belief among many Afghans who think Western powers — with billions of dollars and their own credibility at stake here — will decide who takes office in Afghanistan, he said.

“The announcement of the run-off puts an end to the belief that the future president is already chosen by foreigners and the election is only a practice to deceive us.

“Now we know it is for the nation to decide who will be our president.”

There have been sporadic protests in the past week from Karzai supporters across the country, calling for a boycott of the run-off and insisting their candidate won the vote.

But a Karzai campaigner in northern Kunduz province said his supporters would comply with the president’s call to vote.

“Now that our own leader, our own candidate, has accepted a run-off and called on us to vote, we will indeed get out and vote and show our candidate was and is the winner,” said Malim Akbar.

Mehbubullah, also from Kunduz and who, like many Afghans, has only one name, said he would vote again.

He said he hoped the poll would be fraud-free and the people’s choice would determine the next leader.

Bookmark & Share

So, the Github contest is over. To spoil the suspense, I didn’t win (I got crushed, actually) but with some Netflix Prize alumni and pro data mining consultants in the mix, my odds were pretty long to begin with. Given these realities, I viewed my entry in the contest mostly as a way to improve my ruby machine learning library, The Decider, so I actively avoided some Github-specific hacks that might have given me a much better score.

My approach to the contest was to find similar users or repos using a k nearest neighbors algorithm. This approach matches our intuition about making recommendations: if you liked Rails, you’ll love will_paginate! Similarly, we can think of finding similar users, as the Amazon recommendation system (apparently) does: “people with prototype in their cart also bought scriptaculous.”

Vectors, Distance Metrics, and Lots of Flops

To start with, we need a way to determine the similarity between users or repos. To do that, we start by representing them as vectors, which for our purposes are basically just arrays with lots of zeroes and a few ones (actually, there’s so many zeroes that it takes too much memory, so, in practice, optimizations need to be made to account for the sparseness of the data). To illustrate this, look at the table below which shows what a small part of the github data might look like:

In ruby, we’d probably represent that data like this:

github_users = {
	:alice    => [1,1,1,0,0,0,0],
	:bob      => [0,0,0,0,1,1,1],
	:chandra  => [1,0,0,1,0,0,1],
	:dave     => [1,1,1,1,0,0,0]
}

So, now that we have lots of zeroes and ones, how do we figure out how similar two users are? There’s several ways, but we’re limiting ourselves to distance measures and excluding similarity measures, for reasons I’ll explain shortly. What’s the difference? A similarity metric will be exactly 1 for two identical vectors, zero or -1 (depending on the measurement used) for completely dissimilar vectors, and somewhere in between for partially similar vectors. A distance measurement, on the other hand, will be zero for identical vectors, and something greater than zero for vectors that are different.

The simplest distance metric we could implement is probably the Hamming Distance. I won’t discuss it here, except to say that it worked much better than I expected for finding similar users, but, for finding similar repos, it was abysmal. One deficiency of the Hamming distance is that it doesn’t make any corrections for the magnitude of the vectors, that is, the number of repos followed by a user (or the number of followers a repo has). To understand what I mean by this, think about two users who have exactly the same preferences, but one has used github for a much longer time and follows many more projects. The Hamming distance in this case will be fairly large, which doesn’t reflect the two users’ true similarity. Cosine similarity, which is also called the Tanimoto coefficient in the case of binary attributes, has the desired property of taking the size of the vectors into account. We can see how this works for two dimensional vectors below.

In the diagram, the vectors represent individual github users, their lengths represent the number of projects the user is following, and the directions show the users’ preferences based on which projects they follow. Given this information, we can see intuitively that Bill and Charlie are the most similar in a way that is useful for making recommendations, even though Anna and Bill follow about the same number of projects and might have a smaller hamming distance.

Observant readers will have noticed that I mentioned using cosine similarity, but ruled out similarity measures. After all, cos(0) is 1, but that’s exactly what I said I didn’t want. Well, one way to solve this is to use the arccosine function to get the angle, which will give us a distance measure, but I simply used (1 – cos(θ)), which gives the desired properties.

To compute the Tanimoto coefficient for binary-valued vectors, we use the Jaccard index formula, which is given by the set intersection divided by the set union. If we compare Alice and Chandra from the table above, we see that the only project they have in common out of the seven total is rails. Taken together, at least one of them watches rails, prototype, scriptaculous, jQuery, and rip (5 total). Therefore, the Tanimoto coefficient for their vectors is 1 / 5 or 0.2. If we compare Alice and Dave, we see that they both watch rails, prototype, and scriptaculous (3 total); at least one of them watches rails, prototype, scriptaculous, and jQuery (4 total) giving them a Tanimoto coefficient of 3 / 4 or 0.75.

Need for Speed

Probably the biggest shortcoming of the k nearest neighbors approach is that these relatively slow distance measurements need to be made between every combination of users or repos, which is quite an undertaking with 50,000+ users and 120,000+ repos. Making an analogy to relational databases, the brute force approach is like tens of thousands of full table scans. In database-land, we use indexes, which create tree structures (ususally B+ Trees) of our data, to limit the number of comparisons needed for each query. So, the natural question to ask is, “can I use a tree structure to make my kNN search faster?”

BK Trees

The reason I limited comparisons between users to distance measurements is to take advantage of BK Trees. BK Trees are organized like this:

1. Make any arbitrary user’s vector the root node.
2. For the next user, compute the distance between his vector and the root node’s vector.
3. Attach the user to the root node along an edge, where the edge is simply the distance between the root node and the user.
4. Continue the process. If there is already a child node with the same distance, attach the user to the child node

This concept actually is quite simple, but it’s a bit difficult to explain, so an illustration will help.

In the illustration, we’ve arranged seven users into a BK Tree. We arbitrarily selected Alice to be the root node. We then attached Bob at distance==3, Chris at distance==5, Dee at distance==10, and Evelyn at distance==23. Frank and Ginger also have distance==5 from Alice, but we already added Chris at distance==5, so we attach them to Chris’ node. When we attached Frank to Chris’ node, we measured the distance between Frank and Chris, found that it is two, so we attached Frank’s node to Chris’ node at the edge for distance==2. Similarly, the distance between Chris and Ginger is three, so we attach Ginger’s node to Chris’ node at the edge for distance==3. To represent the tree in ruby, we might start with something like this:

# The BK Tree in Ruby
bob     = :leaf
dee     = :leaf
evelyn  = :leaf
frank   = :leaf
ginger  = :leaf
chris   = {2 => frank, 3 => ginger}
alice   = {3 => bob, 5 => chris, 10 => dee, 23 => evelyn}

Now that we have our users organized into a tree, how do we go about finding the k most similar users? We exploit the properties of Metric Spaces. Metric Spaces are sets where the objects in the set have a distance between each other. The distance function d(x,y) (a.k.a., the metric) must have the following properties:

The Triangle Inequality

• The distance beween a vector (or any object) and itself is 0. In math lingo:
  d(x,y) = 0 if and only if x = y
• d(x,y) ≥ 0 i.e., metrics are scalars, and the shortest possible distance between two objects is zero.
• d(x,y) = d(y,x) That is, the distance from x to y is the same as the distance from y to x. “Uphill both ways” is not allowed.
• d(x,z) ≤ d(x,y) + d(y,z) This is the triangle inequality.

Of all of these, the triangle inequality is probably the most interesting. Essentially, it says that if you go from point A to point C by going from A to B to C, you have to have gone at least as far as the distance from A to C.

So how does all that help us search the BK Tree and find the k Nearest Neighbors? Let’s start with a slightly easier problem. Say we now have a user HAL, and we want to find all of the users within a distance of 4 from HAL. We start by measuring the distance from HAL to Alice. Let’s say the result is 6. Using the triangle inequality, we know that we need to check every user with a distance ≤ 10 from Alice. How? I’ll leave the formal proof to you, but you can see for yourself in the diagram on the left. You can also see that many nodes that match our triangle inequality criteria aren’t within a distance of 4 from HAL, so each node needs to be checked before we can add it to the result set. To descend through the tree, we simply repeat the process for each non-leaf child node that matched our triangle inequality criteria.

To search through the tree for the k nearest neighbors, we simply set our initial “cutoff distance” to infinity. As soon as we have k results, we know that any results we’re interested in must, at minimum, be better than the worst result we have so far. So we continually adjust the cutoff distance as we descend the tree, hopefully converging on a small value for the cutoff distance that will allow us to search the smallest possible part of the tree.

The hyper-observant reader will have noticed that BK Trees will only be advantageous for integer distance functions, but I’ve used one that gives floating point values. Well, I cheated and multiplied my distance measure by a large constant factor and, for the purposes of the BK Tree edges, converted to an integer. In practice, the improved accuracy of the cosine similarity-based distance metric more than made up for any inaccuracies caused by this slight of hand.

Practical Concerns in the Github Contest

To generate the actual recommendations, I took two approaches. The first one I tried (which gave me a slighly better score by about 1%) was to group the data by user. Having a list of the k most similar users for each user in the test set, I then recommended the most popular repos watched by these “neighbors” that the test set user wasn’t already watching. The second was to find the k most similar repos to each repo in the entire dataset. For each user in the test set, I recommended repos similar to the ones the user was already watching.

As I mentioned above, I decided early on not to use any Github specific “hacks,” such as replacing forked repos with their parents, as I was interested in learning about recommendations in a more general case.

Computing the Recommendations

Even with the optimizations afforded by the BK Tree, generating the recommendations takes a lot of time. So I took the opportunity to get acquainted with Jruby. In fact, without it, it would have been impossible to use ruby for this. Some of the truly enormous computations, such a generating the 50 nearest neighbors of every repo in the dataset, required days of CPU time. With a trusty high CPU XL EC2 instance and Jruby, wall clock time was still measured in hours, but manageable.

Speaking of the BK Tree, because it has no balance constraints, the amount of the tree searched for a given kNN computation isn’t guaranteed to be anything less than the whole tree. On the several occasions that I inspected the efficiency of the BK Tree searches, I noticed wildly differing percentages of the tree were searched, between 5% and 90%. One interesting result was that as I increased the value of k, many search’s efficiency dive-bombed, that is, instead of increasing the amount of the tree searched from 10% to 20%, it would increase from 10% to 80% or 90%. I suspect that the users are naturally clustered, and when k exceeds the size of this natural cluster, there are many equally bad results. It would be interesting to verify this by analyzing the dataset with an edge detection-based clustering algorithm.

Improving the Recommendations

Although I tried using both users and repos as the basis for the kNN groupings, I didn’t get around to combining the results of the two. I speculate that this approach might help overcome the weaknesses inherent in each approach. For example, some users in the test set only watched one or two repos. For these users, picking a “similar user” isn’t likely to be a good approach since we have so little information about them. Conversely, for users watching many projects, picking repos that are similar to the ones they already watch ignores the information about their preferences that we could take advantage of by basing the recommendations on similar users.

Aside from being sloooow, another problem with the pure kNN approach is that no effort is made to reduce noise in the dataset. One could imagine that users might have decided to follow a repo outside of their normal preferences on a whim, or they might not follow a big project, such as rails, even if they’re interested in it, because they hear about what’s going on with that project from other sources (the developers’ blogs and twitter accounts, etc.). Smoothing out that noise is likely to give better predictions. One promising way to do this is via Singular Value Decomposition. Be sure to also have a look at Ilya Grigorik’s post on the topic, if you haven’t already. This is the next item on my list to add to the Decider library; in fact, I’ve already found a suitable Java implementation of the whiz-bang linear algebra to use with Jruby.

Github Specific Approaches?

(This paragraph added on 30 Aug, 2009) Scott Chacon, in a post on the Github blog, expressed some dismay that it was possible to get quite a few correct answers (I’ve read that it’s between 15%–20%) by guessing that users are watching projects they’ve forked. I can agree with his sentiment to some extent, but another way of looking at it is that the project/fork tree is simply another “social graph” structure full of interesting and useful information. In an IRC conversation, careo discussed the idea of putting the data in a Neo4j graph database and generating recommendations using queries against the graph. One interesting avenue of research would be to augment the explicit relations (users following projects, projects forked from projects, etc.) with the implicit relationships found by the distance metrics and use queries against that graph to generate recommendations. The obvious issue with this approach is that different types of measurements are being treated as equal (or at least interchangeable), so I anticipate a good deal of tuning would be required to correctly balance the relative weights of each type of relationship.

Fin

The contest was pretty awesome, and I learned a whole lot in the process, even getting reacquainted with some math I haven’t used in a while. So, thanks are definitely in order for the Github crew for putting their dataset in the open and generating open source activity in a realm of computing that’s all too often sequestered in stuffy journals and corporate trade secrets. I’ll be updating this blog with more information and code examples as I continue to improve the decider, though hopefully it won’t be anything quite so long-winded.

Further Reading

• Algorithms of the Intelligent Web, Haralambos Marmanis and Dmitry Babenko. ISBN: 978-1-933988-66-5
• Matrix Methods in Data Mining and Pattern Recognition, Lars Eldén. ISBN: 978-0-898716-26-9
• PDF The Unreasonable Effectiveness of Data: Some of the Google team get you pumped about enormous scale machine learning.
• Ilya Grigorik’s Series of Machine Learning Posts (in Ruby)
• The KDD Cup 2007: Academics drunk off of the Netflix data.
• Another introduction to SVD, with a demonstration of noise reduction and compression of images.

Mea Culpa: I messed up the Tanimoto Coefficient calculations (for Alice & Chandra and Alice & Dave) in earlier revisions of this post. They should be correct now. My Bad.

Créer un site internet ou une application web à partir de zéro, “from scratch”, n’est absolument pas trivial. Cela représente de nombreux points à étudier, analyser puis décider. De plus il y a une forte corrélation entre chacun des éléments ce qui ajoute à la complexité du choix et à l’obligation de faire des compromis. Par exemple ajouter des fonctionnalités, va augmenter le temps nécessaire à réaliser l’application et donc aussi le budget (en général).

Ici j’essaie de faire une sorte de pense-intelligent qui reprend un peu tous les points clés qu’il faut traiter pour arriver à une application bien pensée : fiable, efficace, et évolutive. Cela va de la définition des besoins en allant à la vie de l’application en passant par les ressources disponibles (budget, temps, hommes) et les aspects juridiques et commerciaux. De quoi faire donc.

Faites chauffer vos neurones ! On y va …

1. Le besoin  / les fonctionnalités
   1. Internationalisation (côté client, côté serveur)
   2. Quantité de données techniques / utilisateurs (base de donnée, cache)
   3. Expérience utilisateur : statique / dynamique (gestion de l’historique)
   4. Communication avec d’autres applications/sites (Rss, web service, cUrl, widgets)
   5. Gestion de  données privées / sensibles (authentification, cryptage, côté client, côté serveur)
   6. Importance du référencement (url lisibles et uniques, génération des meta, génération de liens internes)
   7. Statistiques de fréquentation incluses (stockage IP, gestion cookie)
2. L’existant technique
   1. Plateformes complètes à paramétrer et/ou modifier (ex: ERP, CRM, gestionnaire de documents, forum, outil de collaboration, service de partage, etc.)
   2. APIs (Google, Yahoo, etc.)
   3. Frameworks
   4. Librairies
   5. Classes uniques prêtes à utiliser
3. Le contexte d’intégration, d’architecture
   1. Serveur d’application : (Java, Php, .Net)
4. Les langages de programmation
   1. Web côté serveur : Php, Ruby, Python, Java, .Net
   2. Web côté client : Javascript, Flash/ActionScript
   3. Design : HTML/CSS, XML/XSS
5. Les ressources disponibles
   1. Les connaissances (expertise interne/externe, documentation des outils, communauté, etc.)
   2. Le budget (global, par fonctionnalités, par itérations, etc.)
   3. Le temps (délais, disponibilité/participation du client)
   4. Les moyens humains (responsabilités, prestation interne/externe, disponibilité des équipes, etc.)
6. La méthodologie de gestion de projet
   1. eXtrem Programming
   2. Scrum
   3. GTD
7. Certifications requises
   1. CMMI
   2. ITIL
   3. CISQ
   4. Langage de programmation (Zend, Java, etc.)
8. L’utilisation
   1. Documentation
   2. Formation
9. Juridique
   1. Déclaration CNIL : gestion de  données privées, durée de conservation, objectifs recherchés, etc.
   2. Modalité de signalement d’un abus
   3. Utilisation Opt-in / Opt-out
   4. Mentions légales
10. Commercial
    1. Suivi de popularité (PageRank, Tweets, etc.)
11. La vie de l’application
    1. Maintenance corrective (gestion des bugs, feedback)
    2. Maintenance évolutive (ajout de fonctionnalités, veille technologique)

Je tiens a éclaircir plusieurs notions :

• Le classement adopté n’est pas un classement temporel car tout est lié et tout doit être pensé en même temps (si je puis dire)
• J’ai essayé de ne pas adopter un point de vue trop technique (bien que les choix techniques soient très importants) tout en restant précis
• Je développerais surement, dans quelques temps, chacun des points cités
• Cet article est à vocation évolutive et je le mettrais à jour au fur et à mesure de mes connaissances / retours

J’ai sûrement oublié certains aspects et je compte justement sur vous pour me les signaler. De même si vous pensez à une organisation plus claire ou plus logique.

Six months ago I started this blog with President Obama’s Inauguration. That was a great day and I was very proud, once again, to be an American.

So how do I feel now, six months later?

Slightly less enthused.

I have to be honest. I feel that the hope from the red, white and blue poster of his face has tarnished a bit and the reason is a growing pit of fear in my belly that we’re still enduring business as usual with a new wolf in the sheep’s clothing. This sheep outfit just doesn’t yet have visible zippers.

Now before you go declaring me unAmerican and anti-change and some sort of sudden racist, read just one reason why (For the purpose of this blog I’m only going to focus on one of the things that has set me off):

The Goldman Sachs Connection:

2 weeks ago Goldman Sachs posted their earnings and showed quite a profit for the last quarter. They’ve paid back their TARP money in full. At first glance it seems like good news, but we should all be media savvy enough to know that the news can be spun in more ways than a spiderweb. I’m going to post links to a bunch of stories about this relationship and the situation at hand but before I do that here’s a summary:

- Goldman Sachs was one of the largest contributors to Obama’s Presidential campaign.

- A Goldman Sachs exec has ranked high in every presidential Administration since Clinton’s.

- When the financial crisis hit Goldman Sachs benefited by having their assets in other companies bailed out by the US Taxpayer (to the tune of almost 100% on the dollar). Without that help they’d have lost billions.

- The US taxpayer was not given any stake in Goldman Sachs. Now that they are profitable again by returning to their old ways of doing business (aka the very ways of doing business that led them to require a bailout in the first place) they are not sharing any of those profits with us, the taxpayer. Should they need a bailout again no doubt we’ll be there handing them cash but our hands will always remain empty. And who is making that decision for us? A whole new decider.

Angry yet? Scared? Read some of this then do your own searches. Share what you find out.

From CBSnews.com blogs:

“[Taxpayers] didn’t get stock in those banks, they didn’t ask what was going on – this begs and cries out for hard, tough examination,” he continued, complaining of “a Ponzi scheme, an inside job.”

…All this has led to more scrutiny of a company that has long been a mystery to most Americans. Because Goldman operates as both a trading house and “a fee-based adviser to the companies being traded,” as New York Magazine puts it, some have begun characterizing the company as a player that has essentially gamed the system – an entity that knows “every hand at the table and using that information to enrich itself at the expense of others.”

From Crime & Federalism:

July 17 (Bloomberg) — President Barack Obama announced today he was nominating Robert Hormats, a vice chairman of Goldman Sachs International, to a top economic position at the State Department.

… the media has failed to hold Obama responsible for Goldman’s actions. Every story on Goldman’s influence in government leads with the line that George W. Bush appointed former Treasury Secretary Henry Paulson. Most fail to note that one of Paulson’s predecessors, Robert Rubin, spent 26 years at Goldman Sachs: Rubin was a Bill Clinton appointee.

Goldman Sachs’ power extends over the Republican and Democratic fiefdoms alike.

From The Huffington Post:

I noted that, by the end of June, Wall Street had already given Obama $9.5 million, that four out of his top five contributors are employees of financial industry giants, with Goldman Sachs at the top of the list. Even conservative New York Times columnist David Brooks was appalled: “Over the past few years, people from Goldman Sachs have assumed control over large parts of the federal government. Over the next few they might just take over the whole darn thing.”

PRLog: Shows you how the banks fared in terms of costs of shares before during and after Obama’s election. Goldman’s done very well.

Decider: You once said your main goal was to get Bush out of office. What’s left for you to do?

Fat Mike: Kick back and see what happens! My angst is gone right now, and I’m gonna see what comes of it all.

D: A lot of people are celebrating Obama’s victory, but some people in the punk community, including Jello Biafra, remain suspicious. What’s your take on our new president?

FM: Actually, I’m not suspicious at all. I think he’s a really good man. Still, the chances of surviving this are pretty low. I don’t see a big future for this country. I do think the America we know will become a much more dangerous country, much less wealthy. I think the U.S.A. will become more like Argentina or South Africa.

D: Bill Stevenson, who’s played with the Descendents, Black Flag, and All, produced Coaster. What did he add to the band’s sound?

FM: What’s great about Bill Stevenson is his ability to record a song. When I tell him we want to sound like X, he makes us sound like X. He gave us inspiration. Seventy-five percent of the guitars on this album were played on a Jazzmaster Fender through a Silvertone amp. We haven’t recorded like this in a long time—it’s a real classic sound. It’s our most old-school-sounding record. It comes from a time before we started playing. It has the L.A. punk sound. We nailed the Germs guitar tone.

D: This is your 12th album. How will you avoid becoming a novelty or nostalgia act?

FM: Well, I think we’ve made it this far, so we have to just keep following our own footsteps. Never writing for your own fans keeps it real. Keep changing, never trying to outthink your fans. You have to have a contempt for most of American society to do what we are doing.

D: Do you worry that your comedic side overshadows your politics?

FM: It does. People don’t take me as seriously as other people. We’re more of a joke band. We’re always lighthearted as a band—but that helps Michael Moore. Who are people gonna take more seriously, Noam Chomsky or Michael Moore? I’ll take what we can get.

D: In an interview from the mid-’90s, you said that NOFX had gotten better since it started. Are you guys better now than you were then?

FM: I believe so. We’re also lazier. We’re not as hungry now. We kind of take it easy. Around 1994, we reached a high point in our music. But we haven’t gone down from that level, live or on our records.

D: The members of NOFX are known for drinking and doing drugs, and people talk about it all the time. Does this bother you?

FM: I have a problem. I have a diarrhea-of-the-mouth problem. I don’t think anything is too private. I am really sickened by the state of the music industry today. No one wants to say anything specific these days. Other bands don’t want to lose the Christian or religious fans. Well, I like singing about stuff that is offensive. I don’t limit my lyrics. I get right to the point. So no, it doesn’t bother me at all!

From: San Francisco Decider

You know, it’s so fashionable to be anti-Bush right now. I wonder how many people really know the factual reasons why people are anti-Bush.

I look at movies like “W.” and if you ask me it’s just a matter of time until we’ve forgiven him for everything he’s done. You look at that movie and most of the time he’s stuffing his face with sandwiches or drinking a beer, just looking a lot like your average guy. It’s an incredibly sympathetic depiction of Bush as just an average guy, maybe not too bright, and in over his head.

I think it’s a tribute to the fact that we’re not really emotionally angry with Bush. We’re rationally angry with Bush, which tends not to work. People often don’t seem to make import decisions based on reason.

I have to wonder if in a couple of years people will just move on and forget about the whole thing as more pressing problems take the forefront in our minds. At the end of the day he’s just a real likeable guy, the Bart Simpson of presidents if you will.

If people don’t have it very clear in their heads right now exactly why it is that we all don’t like Bush then I gaurantee you in a couple of years all will be forgiven or at least forgotten simply because at the end of the day he’s just a real likeable guy.

Best moment: “I spent eight years with my face out there saying things I barely understood while you were nowhere to be found,” Dubya says to Darth.  “I got the one guy who scares me more than my dad.”

Thank to DIYDanna.

First part is here.
We discuss feminism, right vs. privilege, The Reader parody, the election, and more!

That title is one way to describe Austin.

Here are a few of the first photos I took there when I performed at the LAFF Festival this weekend.

So my friend from my days at Emerson College, Anna, lives in Austin now and works at UT. Therefore, she was an awesome guide to things and places. The last travel adventure she and I had together was when we lived at a
castle in the Netherlands for four months in 2000. I lived in the servants’ quarters and she lived in the prison tower (both had been renovated several times in the previous 700 years, thankfully).

me and anna and my beyonce claw at lucy in disguise on south congress

Across the street from Lucy in Disguise, there’s a row of trailers that sells such delicious items as cupcakes and crepes.

The cupcake trailer is on the far right.

Then there was this rad candy shop.

Austin will, in fact, let you lick its lollipop.

It had a wacky circus sideshow theme, plus Jesus.

Wheeee!

Such as this one.

Um. And also. There is a real School of Rock in Austin. It was founded in 1998 by a dude named Paul Green. And this weekend, they played an ’80s prom at a joint called Mohawk.

This kid was like 10. She won the title of Prom Queen.

Some ninth graders (probs) at the rock show.

It was pretty rad. We didn’t vote for prom king and queen, even though we had a ballot and they told us to, but Travis and Sairah (NOT “Sarah”) totally won.

Also we went to see “Star Trek,” and Young Spock (NOT Baby Spock, who is also in the movie) gave me a total girlboner. I will probably need to talk about this on the radio this week.

Ces and I invited my moms to come out with us, but she was tired after we hung out at Barton Springs and kicked it with our cousins (7 of them, from 4 to 70, were there! They drove from Houston and Dallas!) so she just chilled at the hotel.

And then Ces and I decided it was time for Star Trek at the legendary Alamo Drafthouse. They make their own kickass trailers for their own kickass events and run them before the movie trailers. They’re having a pirate party on a fucking boat next week. NO BIG DEAL. JUST AWESOME.

As. Was. The. Movie.

i can haz Vulcan dick now?

Okay, look, I know you not might be into “Star Trek” the series? Or even the previous movies? Or the cartoon? Or the Trekkie documentary? Or space, the final frontier? But seriously, this movie is good even if you think Neil Armstrong should have stayed at home and jacked off to “Howdy Doody” reruns because space is gross but puppets are awesome. Even if you are one of the zero people who thinks this way, you will still love this movie.

And Young Spock will give you a huge, and I mean fucking HUGE, nerd girl boner. Like, HUGE. I wanted to hump his weird face. They cast a pretty boy as Kirk, and he does a good job and actually manages to be quite charming and rakish and enjoyable by the end, but Young Spock steals the movie, which isn’t hard because so much of it is actually about him. It’s really interesting and Winona Ryder doesn’t even ruin it at all.

Now here is a photo of a baby’s tombstone. You can buy it at Uncommon Objects on South Congress in Austin. You’re welcome.

Probably the best thing I've ever seen for sale.

NEW YORK (Reuters) – La querelle legale sulla prossima America’s Cup di vela è di nuovo nelle mani di un giudice di New York, dopo che le discussioni tra i due rivali sono, nuovamente, naufragate.

Il Golden Gate Yacht Club, sponsor del team Bmw Oracle, ha presentato una denuncia ad un giudice di New York, contro i campioni in carica di Alinghi, che corrono per i colori della Societe Nautique de Geneve, per aver ignorato la sentenza del Tribunale di New York e per il tentativo, secondo il Ggyc, di ritardare l’inizio della più importante competizione di vela al mondo di altri tre mesi.

All’inizio del mese di aprile, la corte d’appello di New York aveva giudicato irregolare la decisione di Alinghi di accettare la sfida del Club Nautico Espanol de Vela, team spagnolo creato proprio pochi giorni prima di lanciare la sfida ai campioni.

Thomson Reuters is the world’s largest international multimedia news agency, providing investing news, world news, business news, technology news, headline news, small business news, news alerts, personal finance, stock market, and mutual funds information available on Reuters.com, video, mobile, and interactive television platforms.

Fonte:
http://it.reuters.com/article/entertainmentNews/idITMIE53S0F420090429

Authoress and highly magical Texan Danna Williams kindly did an interview with me for The Onion’s Decider Austin. My favorite part is the pullquote above my photo. (For the purposes of this quote, “stupid person” = “person who thinks most women are not funny.” Also, “my comedy” = “poopy poopy fart fart.”)

Thanks Danna!

My efforts to post my thoughts to all the movies I attened at the 2009 Wisconsin Film Festival fell short. I’ve been sick since Sunday with some weird sinus infection/cold. In the mean time, ultimate spring season and baseball began. And between fighting infection and other demands for my time, little has been left in the world of writing about the film fest. Here are my non-film specific thoughts, and hopefully, by the end of this week I will get all of my reactions written to the shows I saw.

In the end, I went to nine films. A good amount, but not nearly as hardcore as some of the attendees. (Perhaps, we could make a documentary about the people who go to see a record number of films?) I saw, in chronological order: Anvil: the Story of Anvil, Milk in the Land: Ballad of an American Drink, Illegal Use of Joe Zopp, British Television Ad Awards, Between the Folds (and two accompanying animation shorts), Chef’s Special, Paper or Plastic?, Visual Acoustics: The Modernism of Julius Schulman, and Mermaid. This year’s experience for me was much better compared to last year’s. Last year, I saw seven films, but four of those came on Saturday. This year, I wasn’t trying to do too much on any given day and I felt like I saw a lot of shows. I am more of a film – drink – film – coffee – drink – film – drink person. Being social in between shows makes for a much more enjoyable festival experience. One needs a little socialization to spruce up the individual experience of watching film. I’m a bit disappointed that I didn’t gut out the rain on Thursday night to get rush tickets to Rockafire. Regardless, I’m happy the way things worked out.

Meg Hammel and the volunteers deserve a lot of credit. They gracefully handled issues that would have made the rest us panic in a heart beat. A film not in Madison on the day it’s supposed to show up? Not a problem – we have a driver intercepting it on the route back from the Chicago airport. A show canceling the week of the festival? Never fear, we found a backup on an incredibly short notice. On top of it, everyone was extremely polite and helpful.

As soon as the festival ended – like many of the Madison blogs – my attention quickly turned to the next thing. Which is, apparently, baseball. And with the beginning of baseball season, come the usual questions about why I’m a Cubs fan. The Brewers-Cubs rivalry has gotten a bit intense over the last few years, mainly because the Brewers have slowly built a good team. And so, I get some hatred from people who have recently discovered they are baseball fans. For those of you who aren’t sports fans, think of how recently converted evangelicals feel about atheists/Muslims/Jews/liberals/city-folk/gay-people/and well anyone not like them; it’s irrational and makes little sense.

A couple of days ago, Madison’s Decider decided to fan the flames. I found a couple of responses to be extremely identical to how I feel about this kind of garbage.

User Wallrock wrote:

I’ve been a life-long Cubs fan from the Madison area and I’ve never quite understood the rise of the virulent hate directed my way in the last five years. Three reasons:
1) when I grew up I was kind of a secondary Brewers fan, and since the likelihood of both teams meeting in the World Series was minuscule this went unchecked until ‘98 when the Brewers switched over from the Junior Circuit. Now I obviously can’t really support two teams in the same division but 18 years of Brewers-like doesn’t just go away despite the fact that everyone seems to hate me now.

2) When the switch first happened I was actually kind of happy since it meant an easier trip to Milwaukee to see the Cubs. But back in the seasons following the ‘03 collapse both teams were fucking terrible, so it was a bragging rights 5th place in the Central-type game. You’d sit out in the parking lot and maybe talk a little shit but it was an overall friendly affair punctuated by the sharing of many beers. Of course at that time it was like 3/4 Cub fans in the place, but it’s not exactly like those tickets were stolen from Brewer fans – they weren’t showing up anyways. But I got used to the easy-going nature of those kind of games at Miller Park, so I was not expecting what I experienced the last two years. And let’s be honest – douchebags are douchebags no matter who they’re supporting. My experience with the Marquette fratboys last year may have been due to either my blue hat or red hoodie but it doesn’t change how I saw the Brewer fans afterwards.

3) It pains me to make this comparison, but the whole rivalry comes off like the UW-Michigan one, in that beating Michigan (and hating Michigan fans) is a big part of the Badger psyche, but Michigan is always going to have bigger rivals than Wisconsin. Last year was definitely heated between the Cubs and the Brewers, but I can attest to the fact that as a Cubs fan my sports-hate is directed to the Cardinals and the White Sox more so than the Brewers. And I’m guessing that knowing that your rival doesn’t hate you as much as you hate him, well it probably makes you hate him even more.

Oh yeah, 3.5 – I’m from Madison. I don’t hate Chicago, I hate Milwaukee. I can’t stand anything about that goddamn hole. But that’s the way it works – you hate upward. Madison hates Milwaukee, Milwaukee hates Chicago, and Chicago NY or LA.

And user Klik writes:

Let me preface the following by stating that I am, like Wallrock, a lifelong Cubs fan (also currently residing in Madison), so I grant that a certain amount of my objectivity was lost after reading the article title and certainly after discovering that this laughable attempt at sportswriting was crafted by a hardcore (aka post-2005) Brewers “fan”.

I’m quite glad that you mentioned that your magnificent insights (thanks Rmarshall) are a couple of years old, as this highlights the timeline on the growth of your Cubs hatred. You’ve only been a Brewers fan since 2006, yet you developed an intense hatred for the Cubs by 2007? Despite the hilarious bandwagoning aspect of this admission, rest assured, you’re not the only one. All of the other Brewers fans which I’ve met seemed to share this same sentiment…starting in 2007. Now it’s easy to see why – as a perennial bottom dweller, Milwaukee gained boatloads of new fans by building a team worthy of walking onto the field, and they’ve definitely been an exciting team to watch for the past couple of years. The fact that this anger was born towards the past two years’ division winner is no surprise, and Chicago’s close proximity to Milwaukee doesn’t hurt either.

Klik then takes to task all of the reasoning in the article. It’s almost more amusing than the article itself. You can check out the article and all of the other comments at the link above.

Root for your team. Who ever that is. Rag on the opponent when appropriate, but don’t provide reasons or additional opportunities to hate on others without thinking about why you are doing so. It’d be like if I found reasons to rag on films that didn’t get the same success/participation/acclaim as others. It does little good and doesn’t contribute.

Oh.

And Go Cubs.

(FYI: I’m going to Saturday’s Cubs game @ Wrigley Field North)

Today is Opening Day for MLB. It is one of my favorite days of the year. I’m a huge Milwaukee Brewers fan and can’t wait for this season to unfold after getting to witness some of the most exciting moments in club history last year.

Yet, as any loyal Brewer fan will tell you, there is quite a distaste for our NL Central rivals 90 miles south of us – the Cubs.

For a proper breakdown of why this is, check out this wonderful article on Decider Milwaukee.

Now onto the first pitch… GO BREWERS and PLAY BALL!

I thought we were voting for hope and change.  I didn’t realize that we were voting for a total dict-hater.

According to Gawker.com, the Department of Homeland Security has sent a cease-and-desist letter to Sixpoint Craft Ales, a Brooklyn based brewery, to stop production of Hop Obama beer.  DOHS goes so far as to threaten to seize the brewery if the hoppy brew featuring the President’s name is not discontinued.

Honestly, how can he talk about economic stimulus when he’s basically shutting down the fastest growing segment of our economy?  As this blog has documented, people will sell (and buy) anything with Obama’s name or likeness on it.

How will I be able to continue this blog if there’s no more Obama dildos, no more Obama plungers, no more quarters with the word “Obama” written with a sharpie?

Seriously Obama, would you like some Hateraide with those hater-tots?  Who are you, Ruth Hater Ginsburg all of a sudden?  Do you shop at Hater Joe’s?  If Dubbya was the decider-in-chief, Obama is surely the hater-in-chief.

So Obama.  Do us a favor.  Times are rough enough.  Let us keep our Obama condoms, Obama pornos, and Obama soap-on-a-rope.  You stick to the economy, the wars in Iraq and Afghanistan, your next Commerce Secretary and what type of dog to get Sasha and Malia, mmmmk?

Here’s to 2009–and a new administration!

Jon Stewart’s Bush Impression::

See more Best of Bush videos (including a Farewell and the epic Iraq War story)

~ XOX

Yeah, the buck stops at the top and the self-styled “Decider” had all the wrong ideas at all the wrong times for all the wrong reasons. Here’s how the NYT explains it in a beautiful analysis today:

From his earliest days in office, Mr. Bush paired his belief that Americans do best when they own their own home with his conviction that markets do best when let alone.

He pushed hard to expand homeownership, especially among minorities, an initiative that dovetailed with his ambition to expand the Republican tent — and with the business interests of some of his biggest donors. But his housing policies and hands-off approach to regulation encouraged lax lending standards.

Mr. Bush did foresee the danger posed by Fannie Mae and Freddie Mac, the government-sponsored mortgage finance giants. The president spent years pushing a recalcitrant Congress to toughen regulation of the companies, but was unwilling to compromise when his former Treasury secretary wanted to cut a deal. And the regulator Mr. Bush chose to oversee them — an old prep school buddy — pronounced the companies sound even as they headed toward insolvency.

How could one affable, good ol’ boy from Texas screw up so many things so quickly? Good question. But he didn’t do it alone. He had a whole host of enablers who elected him to office twice without ever looking at his resume. If they had, they would have seen that what George Bush did as president is exactly what he has done all his life: Time and again he was elevated to a position he didn’t deserve by his father’s influence and money, screwed things up royally, then turned to his father’s buddies – including the bin Laden family – to bail him out. Only difference this time, is we’re the folks doing the bailing!

Hey – let’s look what else Texas has given us – I mean,  how about them Cowboys!? America’s team playing it’s last game in hole-in-the-roof stadium. (Yeah, so god can watch his team!)  Yeah – and in the last quarter they give up  back-to-back touchdown scampers of 77 and 82 yards from scrimmage?

Hey, how about them Lions?! Wonder if the Big Three Idiots fly in to watch their team? Does nothing good come out of Texas these days? Or Detroit? As the NFL teams go, so goes their states? Hmmmm . . . Maybe we need to root harder for the Pats.

See why I seek solace in the night sky, coming back to “reality” only  once in a while to look at a football game? Every time I look at the “real” news, I see something like the preceding. Yeah, I saw a lot of this coming. But there’s little comfort in knowing that the nightmares you had in 2000 were an accurate preview of the country in 2008!

Believe me – I hope that the Miracle Man can perform miracles. But he’s the ultimate version of the kid who is given a barn full of manure for Christmas and immediately grabs shovel and with a smile says – “There must be a pony here somewhere!”

Arghhhhhhh . . . and speaking of shovelling crap, I’m guessing his new chief of staff didn’t emerge from the Chicago sewer system without some of the smelly governor’s slime sticking to him. Damn!

ooops -  I forgot - major mood shift time – Merry Christmas . . .

I know, I know . . . some folks don't like snow. Sorry. I love it! Gives the world a major do over! Hmmm. . . maybe that's the derivation of "snow job?" Ah what the heck - grin and bear it

The scramble to define the Bush Presidency is in full swing. Surely, it will continue for years to come, but being a grammar-nerd, I find this one particularly endearing.

Enjoy