There is a reason why conferences are so hugely important. Often, what will be published as an article is presented beforehand on conferences, among knowledgable and interested peers. And while some conferences are invite-only or have to be paid for, most universities have a lot of conferences either completely open to the public or at least open to students. Often, just asking someone of one is allowed to just audit the talks without going to all he social events aimed at invited people will be enough in order to be "let it".
Overall, what bothers me most about this blog post is this:
>> Papers are, intentionally, written for an audience of specialists rather than a general interest group, which reduces both the tendency and ability of non-specialists to read them when asked
This repeats the same bias that everyone can say something about everything. Nobody here would complain that mailing list discussions about bugs, features or programming questions are hard to understand for non-programmers. However, a scientific discussion in a paper is supposed to cater to everyone. In reality, especially within the confined spaces of an article, one has to assume some baseline of knowledge in the reader, because one can't afford to explain all the basics one implicitly builds upon. To give an example: I do literary studies, and often people assume that they know just as much without having properly dealt with it, because hey, everyone can read a book, right? right? While that part might be true, every scientific discipline over time builds up a plethora of instruments for analysis of their respective topics, an understanding of the matter's and the subject's history, its pitfalls, shortcomings and strengths.
Also, I do not understand his paragraph regarding the "bottom line", since I know noone who writes the abstract first. The abstract is the very last bit to be written, after the paper's structure, arguments, theses, supporting evidence and rebuttals are decided, fixed and usually written. What both the cited Paul Graham essay and his text fail to see is the process of writing a paper. It is true that the actual writing process usually doesn't start with a thesis statement and a chain of arguments. But structuring a paper is just as much part of writing as finding words for what one wants to express.
The author hints at being an undergrad in maths, having trouble with the understanding of state of the art research. A professor I know once said this about degrees: undergrad & grad work are a necessity, they teach how to start a car, how to accelerate, break and also drive on basic roads (research methods/skills). A PhD then allows (and teaches) you to drive on the autobahn (research), where you can start to close the gap somewhat to all the people already on it going faster than you do. The actual entry into novel research doesn't happen until after the PhD. That's where journals and conferences, workshops, groups, funding etc. become interesting and important. First then are you within the actual target group for academic papers, since you are seen as an equal, equal also in terms of discussing on eye level.
Nobody expects complex technology etc. to be understandable by everyone, but one could say: "Everyone thinks they are a writer, much like many people stand in front of contemporary art saying they could produce the same thing." What they fail to understand is the complexity involved in the writing process. What sometimes looks easy or even "trivial" in a finished paper often has a highly complex, dense and far-reaching background.
[edit] Mhm...I apologise for the length of this comment, it grew larger than I originally thought it would... [/edit]
I disagree about the value of conferences, especially large ones, for actual discussion.
If you go to a large conference (e.g. the International Studies Association annual conference), you are almost guaranteed to be fielding questions from an audience that has not read your paper.
You would get more valuable feedback on your actual work (rather than your academic powerpoint/presentation notes) from a small colloquium than from a large academic conference. The only person is guaranteed to have read your work is the discussant, and if you're lucky the panel chair and other presenters.
Conferences are a good way to network and get a feel for the overall state of the field, but for having actual discussion of a problem or issue area they're pretty ineffectual.
I suppose I should have defined "conferences" more closely, which for me includes workshops, workgroups etc. Right now, my subject isn't the biggest, so people usually know each other, each other's interests and what most other people are currently working on.
So yes, different experiences with conferences here. However, you are obviously right that the bigger the conference, the less constructive the overall discussion. This in a way corresponds to the aforementioned amount of readers targeted. If you aim to reach a broader audience, discussion will be less in-depth, if you specialise and focus on a smaller amount, discussion is more likely to be more relevant.
I totally agree that conferences are important, and lack some of the "bugs" I identified. However, that wasn't the point I was trying to argue.
Nowhere did I argue that "everyone can say something about everything". That's a strawman. In fact, in the last paragraph I said the opposite: "All forums need some sort of moderation to avoid degenerating."
I do think that being readable by a larger audience is an advantage, all else being equal, to being readable by a smaller audience. Especially since, as explained in http://lesswrong.com/lw/kh/explainers_shoot_high_aim_low/, we tend to vastly overestimate the size of the audience who can understand us. However, there's a huge difference between "can read" and "can contribute equally to the author".
In an experimental paper, sure, the abstract is written last. However, as one of the other comments noted, my post is (by virtue of being originally written for Less Wrong) targeted more at philosophy papers, like in point #4. In experimental and mathematical papers, you're right that this is less of a concern.
For the record, I have already graduated college (with a major in math and economics).
>> For the record, I have already graduated college (with a major in math and economics).
Oh, the undergrad bit wasn't in any way meant as an ad hominem or the like. Apologies if it came across as such.
>> I do think that being readable by a larger audience is an advantage, all else being equal, to being readable by a smaller audience.
I don't necessarily disagree with this statement, all I am pointing out is that there are limits to how wide an audience a given paper can address. Specifically in more abstract papers (e.g. philosophical papers) is a profound knowledge important, especially in the history of thought and the foundations of our modern philosophy, together with the objections raised along the way. To give just one example, if you want to actually talk about something as abstract as postmodernism (an internet favourite, as shown also by xkcd etc.), you will need to know about its genesis, about its criticism of structuralism, modernism, idealism, society, art as an institution, the avant-garde movements etc. All I am saying is while addressing a wider audience is an admirable goal, it becomes harder and harder to achieve the higher the paper's complexity.
>> In an experimental paper, sure, the abstract is written last.
Coming from a non-experimental background, I still don't know anyone writing their abstract first. First comes an interest, an idea, something one doesn't understand, believes to understand, something one finds irritating, interesting etc. Then one keeps digging deeper, looking at other people's work and interpretation while simultaneously structuring one's ideas, arguments and problems. The abstract of a paper -- in my opinion -- can only be given after this process of sorting and structuring, of testing, evaluating and discarding is over. So I would like to disagree that a philosophical paper is the result of a vastly different thought process in terms of structure and argumentation.
Overall, what bothers me most about this blog post is this:
>> Papers are, intentionally, written for an audience of specialists rather than a general interest group, which reduces both the tendency and ability of non-specialists to read them when asked
This repeats the same bias that everyone can say something about everything. Nobody here would complain that mailing list discussions about bugs, features or programming questions are hard to understand for non-programmers. However, a scientific discussion in a paper is supposed to cater to everyone. In reality, especially within the confined spaces of an article, one has to assume some baseline of knowledge in the reader, because one can't afford to explain all the basics one implicitly builds upon. To give an example: I do literary studies, and often people assume that they know just as much without having properly dealt with it, because hey, everyone can read a book, right? right? While that part might be true, every scientific discipline over time builds up a plethora of instruments for analysis of their respective topics, an understanding of the matter's and the subject's history, its pitfalls, shortcomings and strengths.
Also, I do not understand his paragraph regarding the "bottom line", since I know noone who writes the abstract first. The abstract is the very last bit to be written, after the paper's structure, arguments, theses, supporting evidence and rebuttals are decided, fixed and usually written. What both the cited Paul Graham essay and his text fail to see is the process of writing a paper. It is true that the actual writing process usually doesn't start with a thesis statement and a chain of arguments. But structuring a paper is just as much part of writing as finding words for what one wants to express.
The author hints at being an undergrad in maths, having trouble with the understanding of state of the art research. A professor I know once said this about degrees: undergrad & grad work are a necessity, they teach how to start a car, how to accelerate, break and also drive on basic roads (research methods/skills). A PhD then allows (and teaches) you to drive on the autobahn (research), where you can start to close the gap somewhat to all the people already on it going faster than you do. The actual entry into novel research doesn't happen until after the PhD. That's where journals and conferences, workshops, groups, funding etc. become interesting and important. First then are you within the actual target group for academic papers, since you are seen as an equal, equal also in terms of discussing on eye level.
Nobody expects complex technology etc. to be understandable by everyone, but one could say: "Everyone thinks they are a writer, much like many people stand in front of contemporary art saying they could produce the same thing." What they fail to understand is the complexity involved in the writing process. What sometimes looks easy or even "trivial" in a finished paper often has a highly complex, dense and far-reaching background.
[edit] Mhm...I apologise for the length of this comment, it grew larger than I originally thought it would... [/edit]