Society is running up against the truth of what ‘Peer Review’ actually means and finding it disconcerting. The notion that truth by committee as practiced on wikipedia is something different than truth by committee as practiced by a journal is getting tiresome[5]. When society actually uses this idea of Peer Review to create[1] the most powerful and easy to use system of edit and review so far, articles like this bemoan that we have somehow let the Golden Days Of Research[4] slip away. That article was decent but when they complain that Google is linking to wikipedia instead of Peer Reviewed research I say it is working perfectly. A key point: google operates by Peer Review; that is the fundamental basis for their algorithm– someone linking to a website is a vote of confidence[2]. The fact that Wikipedia has risen to the top in this process shows how fundamentally useful it is.

The truth[3] remains though: research is hard, it always has been and probably always will be.  Maybe what these authors are really bemoaning is more people think they are good at research because it has actually gotten easier for once.

Maybe I just need to follow their advice:
Wikipedia is awesome.
Wikipedia is the best source online.
Didn’t you hear how great Wikipedia is?
There is nothing better than Wikipedia.
O O
<       “Wikipedia works for me.”
\__/
And for fun, here are my contributions to this great endeavor, anyone care to review them? What about yours?

[1] By Clay Shirky’s estimate, this–all of wikipedia– was created in less time than American’s spend watching TV every year. Just wait till we put a couple more years into it.
[2] The more people that link to a website, the more weight an outgoing link from that website counts for. So an expert– someone with lots of incoming links– will be counted more than one without. http://www.rose-hulman.edu/~bryan/googleFinalVersionFixed.pdf
[3] In my eyes, do you agree?
[4] This may be a bit of a straw man argument– they never said as much in that article– but it is a common theme I’ve seen about.
[5] The main difference, one side is much less about the reviewing and more about designated authority.

Even If you aren’t looking to learn Common Lisp, I would recommend Object Reorientation: Generic Functions for a non-Java-like look at how object orientation can be done nicely. Multimethods rock.

But after renting for a week I don’t intend on buying even though there is so much of interest left in the game. I (and another friend) found the controls to be sloppy and tiresome. Shaking the remotes to recover from a fall gets old really quickly. It was hard to repeat even basic tricks exactly. I try to shake or flip the controller in the right way and find it doing a different trick, or not registering at all. Ubertricks now are complicated shapes to draw. Even on the practicing screen trying over and over again I could never top a 2/3 success rate on even the simplest tricks. When you are going for the big tricks in events you don’t have time to try drawing the shape several times… you need it to work. My arms just hurt after an hour of playing this game, it isn’t exercise it is just bad jerky movements to try to get it to respond.

The one aspect of the motion sensing I liked was the use of the nunchuck to steer. A combination of tilting from side to side and using the analog stick (each with a slightly different effect) is used to control the riders side to side movement. This worked nicely and naturally (after a few minutes to get used to it).

It might be worth it if you think motion sensing is the greatest thing evar, but I found it clumsy and would recommend SSX 3 instead. Get the GameCube version, you can play it on the Wii with the Wavebird and everything is happy again.

Overall I give it 2 stars.

I’m asking for something better, the following is excerpted from my email to the TA:

Can I *please* do project 3 in erlang? http://www.erlang.org/index.html From the FAQ: “Erlang is a general-purpose programming language and runtime environment. Erlang has built-in support for concurrency, distribution and fault tolerance. Erlang is used in several large telecommunication systems from Ericsson. The most popular implementation of Erlang is available as open source from the open source erlang site.” The entire language is centered around the idea of lots of independent lightweight processes– not threads, there is no shared memory. The processes communicate entirely through message passing.

I’m willing to take a 10% penalty similar to Java. Which having done the first two in C++ I think might be too small and my wrists don’t care anyways.

All the required protocol functionality for the project would be implemented in my code, but as this is a higher level language I can do it in symbols and pattern matching constructs not the tedious mucking about in string splicing and parsing (which I’ve done plenty of and am tired of re-implementing, and will still need to do somewhat for the user’s REPL). I will be dealing with the specifics of sockets and the old BSD interface less directly, but I already have that mostly factored out in my C++ code anyways.

The Rant (you can feel free to skip this, but here is the why):
http://www.slideshare.net/Arbow/comparing-cpp-and-erlang-for-motorola-telecoms-software/ Compares C++ with erlang for use in motorola telco. Found erlang programs are shorter, faster and more fault tolerant, but user more memory (5MB runtime, only really an issue for really small embedded devices).

I’ve been looking for an excuse of a project to learn erlang on for a little bit. I’d rather learn this new language to implement the project rather than Java, which I already know and have done similar projects in before. I was hoping to get some protocol design knowledge from this class, not bit twiddling and pointer chasing. In erlang I can focus on that (and have fun with the erasure codes).

Seriously, after the last project in C++, my forearms ached severely– the language is unnecessarily verbose and inconsistent.

I have been doing web programming professionally for 7 years now while at school and this has led to me doing networking code from time to time. I’ve implemented application layer protocols over ssl sockets for charging credit cards through at least 4 different merchant gateways in C#. My first Ruby program was an http client for an octet-stream of random numbers from random.org. I did network programming in Java all last semester for Nemo’s distributed OS class, and previous undergraduate classes.

One open source thing I’m somewhat proud of is hacking the UCW (Common Lisp web framework) multithreaded backend to use a pool of worker threads dealing out connections to them and having the threads return to the pool when done rather than assign connections in round robbin fashion that could lead to starvation and idling at the same time.

http://common-lisp.net/cgi-bin/darcsweb/darcsweb.cgi?r=ucw-ucw_dev;a=filediff;h=20051111160037-783f4-d10d54dac7e3b5560204e4a880239472b3d1f263.gz;f=src/backend/httpd.lisp
http://common-lisp.net/cgi-bin/darcsweb/darcsweb.cgi?r=ucw-ucw_dev;a=commit;h=20051111162051-783f4-8fb8c299cd4f4435f91578c0cc5b10984e1fc5a4.gz

I rejiggered someone else’s code to connect emacs to a javascript server embedded in a local firefox to help debug javascript live from the editor. https://trac.acceleration.net/JsShellServer Luckily didn’t have to do to much here– XPCOM nightmare… most of what I did was application and presentation layers.

Why do I say all of this (other than to toot my own horn just a little bit :-)? Of this work in implementing application layer protocols and dealing with sockets in different languages and interfaces; the network interfaces are familiar (boost::asio is nicer than some), but I’ve found nothing so difficult about it as everything else in C++ to get this done. The STL is just a shell of a library. Boost helps get it halfway to being useful. boost::shared_ptr is almost like having garbage collection except not. C++ is a world of special cases with little consistency. I’ve seen some absolutely abominable template error messages (ending with unnesting 7 layers deep completely unreadable and because i forgot to link in the proper libraries). This isn’t an “Enterprise language” it is a tarpit– I don’t care what Microsoft tells you. How do you think they got their reputation for OS and application security? I am a smug lisp weenie ;-) This was not my first application in C++ (although first in a while), but my opinion of the language has been cemented by this latest project. The feeling is “Should I be using punch cards as well?”

sorry for the length… C++ really gets to me. erlang?

Did your senator vote us (further) into a dictatorship?
At least there are some independent candidates. I’m not positive, but this guy currently has my vote: http://www.votebrianmoore.com/index.htm

My sister gave me his book last Christmas, and I often read snippets before going to bed. I’ve had a hard time getting through it as every time I go to read it I only get a page or two in before something strikes me as genius and I put it down sated.

He left Robert Siegel speechless on NPR recently– it also includes “The Secret Life of Public Radio”. If you are interested in buying his book you can do it through NPR and they benefit as well.

How to generate content? Well, actually think about what you are trying to sell and write about it. Most clients don’t want to hear that. It is not something that the web designer can do well (he likely isn’t intimately familiar with the product), and it is not something he can just pay $100 to be done with. It requires work from the client.

In lieu of that, reviews are one way to help generate content, although if you are selling schlock then that may not work out in your favor. Post purchase information such as online instruction manuals, references, faqs, stories, will also kee people coming back, and maybe more likely to get you into their del.icio.us bookmarks.