A corpus of our community's knowledge

by infotropism1 min read18th Mar 200914 comments

8

Site Meta
Personal Blog

The purpose of this site is to help building a rationalist community, and helping individuals to fulfill their potential in that domain.

We have a lot of discussions going on, and a lot of material is being, and going to be, generated. At some point it may become difficult for any single individual to follow all of it. Even taking the karma system into account, interesting contributions may be missed by any particular individual. Furthermore, the sum of what would be elaborated upon here would not be as concise or even easily available as it could be wished to be.

To the point : would it be a good idea to try to summarize the most important, relevant ideas upon which we will be building our edifice ? So that a future student of rationality can come upon a concise, easy to digest introduction to our results and ideas, so that less active members can still manage to follow this ongoing process too ?

If so, how would we proceed ? What is being discussed here may not have the quality we'd expect of, say, a scientific publication, though I think that such a quality would be necessary, if even sufficient, for what would eventually become our own corpus of knowledge. How would we elaborate, layer upon layer of work and discussions ? A starting point would be to refer to, or summarize the relevant, existing scientific results that we would lay our base upon. We'd then move on to summarizing our most important achievements, however that word is to be taken, seamlessly upon that foundation.

Any thought on how or whether to organize this?

14 comments, sorted by Highlighting new comments since Today at 12:55 AM
New Comment

Maybe there could be a "Canon" of posts with a recommended reading order. Posts could either have "Nominate for the Canon" buttons associated with them, or else the moderators could decide which were worthy, based on factors like score and referral frequency.

How about a wiki?

Edit: just saw that annoyance suggested it first

One problem is that changes to the codebase seem to be going very slowly. We need Python-literate volunteers in order to start taking on complicated jobs like, say, adding a Wiki element to each post where the discussion can be summarized.

It would be easy to set up a parallel MediaWiki, announce it in a post, and develop a habit to link there whenever appropriate, from posts and comments on this portal.

I wonder if it would be quicker to port the codebase from Python than to find people willing to code in Python.

ETA: I just realized this reads like a jab. I was being serious.

Hell no. Any decent OO/imperative programmer can quickly pick up any decent OO/imperative language, and the sort of code monkey who can't ought not to be let near Less Wrong.

There's a difference between not being able to work in Python, versus not being willing to work in Python. C, perl, ruby, php, even Java would be more appealing to me than touching Python with a ten-foot pole.

Also, presumably you want proficient developers working on the codebase. There's a difference between being able to pick up Python when you have to, versus understanding Python idioms and coding efficiently and bug-free.

ETA: this is getting WAY off-topic, and so I won't continue this thread. If you're interested in why someone wouldn't like Python, there are fires from the holy wars visible clearly from Google that you can go investigate.

Why do you dislike Python so much?

You may not be a statistically significant sample :)

I for one do all my personal projects in Python, so I would prefer it to most of the languages you mentioned. Python is totally awesome :D

This week I forked the lesswrong codebase, installed all the needed dependencies and tried to get it to run. I still haven't succeeded (it seems it can't find psycopg een though I installed it, this morning before going to work I was fiddling around with PYTHONPATH). I will get it working in the near future and start working on it.

(pjeby, another commenter here, is a major Python guru, though I don't know if he plans to contribute)

Your impression is very different from mine: if my primary goal on a project was to attract developers I would write it in Python.

Very unlikely. First, any solution that involves rewriting substantial amounts of code is a priori highly unlikely to be a "quicker" solution than most other sane options.

Second, Python is somewhere in the neighborhood of the 6th or 7th most popular programming language in current use, and many languages of comparable or greater popularity are either painfully lacking in expressive power (C, Java) or detrimental to competent programming (PHP, Visual Basic, Java).

Overall, Python is reasonably close to an ideal language for these purposes, with respect to the metrics of expressivity, sanity, and popularity.

A permanent feature, akin to a FAQ, may be useful.

We already have an "About" link for LessWrong - why not an "About" link for rationality as our community understands and applies it?

On the other hand, who would be responsible for writing the FAQ?

Everyone. Think of a quasi-Wiki system.