a companion discussion area for blog.codinghorror.com

Because Reading is Fundamental


I think we’re actually on the same page in terms of being frustrated with the common commenting experience, without diving further into contentious issues.

What then, would you consider as solutions to the problems you mention? Is it completely hopeless, or are there features that could be implemented to help make discussing even contentious issues more palatable?

ETA: For what is worth, I’m still not convinced there is a great platform for having debates on the Internet, and I think that’s partly what is missing, the ability for a community to refute unfactual claims, to build real arguments in a rigorous manner, and to back them up with verified science and journalism. The Stack Exchange model works well for simple single-faceted questions, which I find promising, but expanding that model to more complicated questions isn’t in any way trivial.


I think it is simple values. You get more of what you reward, less of what you punish. This is hard when the majority has turned. Originally “honor” was sufficient. If you could show someone being cowardly or shallow, they would be embarrassed enough to at least reconsider. Now the response is typically to allege abuse - but the truth hurts, or to obfuscate. Or worse, just more “Bulverism”.

Other than continually trying to get someone to be embarrassed that they have no idea how to argue (in the sense of rhetoric or dialectic, not quarrel) - much like you sometimes have to keep pointing out to an alcoholic when they are drunk that something is wrong, I don’t know how to get people to value reason. But there are some that just don’t know, and they want to be intellectual, but have tried the quick way that only results in the facade. They might be persuaded the real thing is preferable.

This will be difficult. One book I have is basically coverage of creationist-evolutionist debates. The striking thing about is that the creationists were the only ones to talk actual science and reasoning. Mathematics, complexity, probability, entropy. The Atheists were derisive and talked about imposing religion. Yet almost everyone says that the evolutionists are the more rational. Occasionally there will be a debate with the actual science talked about. But more typically, the actual structure of the debates are ignored. The first question should be “who is making the better arguments”? Often the audience would side with the creationists to the surprise of the evolutionists.

The basic thing would be to teach everyone from the earliest what proper arguments are and what they are not. You may like, tolerate, or be friends (or the reverse) with people you agree with or disagree with, but that does not change whether their arguments are clear, reasonable, and consistent. Dialectic literacy, for the lack of a better word. Let those who can only dispense talking points instead of thinking through a subject be treated like those who cannot do sums and make proper change without a calculator (Oh, never mind).

There is also the can’t v.s. won’t. Some of the worst offenders have the mental muscle to do the heavy lifting, but instead shout “Aristotle! A is A, the law of non-contradiction” and use rhetoric, turning reason itself into a cult, and techniques of dialectic into talking points.

Many on this board are coders. Logic should be second nature, and many want more people to get into the STEM fields. But I think this should be turned around in that reason needs to be a way of life. Look at nature, at politics, at everything, and see if the results are consistent with the experiment. And if not, admit the theory is wrong.

An example - the central banks are holding interest low, but unemployment is stuck at high levels, there is inflation (at least with food, and until recently energy), and the top of the 1% are getting richer. If this is the desired outcome, the politicians are knaves. If they desire to persist with something which fails, they are fools. In either case, the policy or the politicians need changing or the outcome won’t change. (See “Japan’s lost decade” which we are repeating). We can talk about the alternatives, but there is no point until we admit the failure of the status quo.

Another example, sometimes we need to check our premises - “Gay Marriage” is controversial, but few ask “Why is marriage a state matter instead of a church matter?”. If we the assumption of the second question is wrong, then the whole controversy is wrong. I can’t think of one public figure (other than Ron Paul) who as asked the question. My point is the question should be asked and answered satisfactorily.

Finally, I think blogs and comments can be a great platform. Check the discussions at http://voxday.blogspot.com/ - even if you disagree with them. They are discussed and many commenters come from opposite sides and make very good points.

Or as I pointed out above, it is easy to be labeled a troll for being extremely shallow and just hurtling insults, but when regurgitating talking points and ignoring direct questions (one of the rules of voxpopoli), is also considered the behavior of a troll, the proper discussions can and will take place.

In our quest to be “nice”, we have turned tolerance of ignorance, foolishness, and stupidity into acceptance if not celebration.

One thing the internet has going for it is that the accidents of personhood are generally hidden. If I can’t tell if the person making the point is young or old, male or female, gay, straight, or whatever, disabled, etc. I can only see their words. And if they make sense. Because “Bulverism” can be made far more difficult. Or impossible - often the “you’re just saying that because you’re a straight-white-cis-male” is met with #NotYourShield.

This ought to be true of coding as well - I may have to tolerate profane v.s. strait-laced, but either the code is bug-free and clear and of high quality, or it is not. And to quote Martin Luther King Jr, we should judge people on the content of their character. But that equally means defining character up, not lowering the standard so that anyone can meet it.

We are reaping the results of lowering the standard of rational discourse to include the irrational. In this the answer is to raise the standards to their old level, and then only let those remain who will abide by it.

When places are created where “You can’t be a troll - overt or subtle - even if you agree with me” is enforced, then things will work. When sycophants are hated more than (reasonable) dissenters.


Two thoughts:
First, I hate infinite scrolling. I want to know how big the thing I’m looking at is. The browser scrollbar gives me this, except when someone decides to f–k it up and implement infinite scroll. Just put it all on one page and let that be OK.

Second, I do a LOT of tablet based browsing. And I have a habit of opening a page only to come back and read it MUCH later (hours or days). So I really wonder about your read time idea.


The problem with the “reading time” metric is that it’s wrong. Not just wrong, but it’s misleading and deceptive. It assumes a reader has come to the webpage, read the article, read through the comments and decided to comment. But that’s not necessarily true and, worse, there’s no way for the site owner to know if it’s a trustworthy metric or not.

In my case, for example, I first read the article in my RSS reader* on my phone, then later thought it was interesting and went to the website on my phone and checked out the comments. Now, I’m on my desktop actually writing this comment. That means the reading time associated with this comment will be somewhere around 30 seconds, which is tens of minutes shorter than the correct value.

The only thing worse than no data is bad data.

(*) Well, actually, I first let it sit there for several days before getting around to it, which is why this comment is four days after the original post and no one will probably read it.


That’s amazing. You read the article for several days! Clearly, you are one of the Coding Horror community’s most valuable members.



I would like to mention bananas.


I jumped out of your blog post to read ‘You Won’t Finish This Article’ in a new tab to get context.

and then decided to come back and post a comment here about how that article mentions how people jump to comment sections before finishing the article… before finishing that article.

(update: before the ‘nutgraph’ bits even!)

Now I’m going to head back to that article and see if they mention how Medium has in-line commenting to prevent this from derailing the reader and also doing some tracking of reading time. (they really do a lot to make reading first class!)

And I’ll probably go post this on twitter: http://ashow.zefrank.com/episodes/28 - which discusses how the impetus to comment is largely out of emotional appeal.

Then come back to your article to finish it.

Finally I’ll read all the comments here before realizing how out of context my post was.

I wish we had metrics that could track user behavior like that. I really wish they had hired me at Amazon on their page metrics team because this is the kind of analysis keeps me up at night. I’m still kicking myself for forgetting 'Markov Chains".


Online we tend to read these conversations as they’re being written, as people are engaging in live conversations.

I think that Online we tend to read these conversations as they WERE written. Starting at the top. Wading through the middle. Skipping to the end.

If we were joining a conversation, we would join at the (current) end, then try to figure back to get some context, then join with a comment relevent to the (current) state of the conversation.

Which is to say, conversations should be “top posted”. Which might excite an avalanche of criticism, but would be an interesting excercise.


Area of the trouble is seasoned visitors. Countless articles you get related to on-line are just waste. You wind up skipping several articles, after which it people produce a routine connected with not really reading through the whole thing, regardless if this content is good.


You’ve touched on one of my main problems with stack exchange - Reputation.

It’s totally normal to want to encourage good answers by rewarding people who help others, but as you say, people will do anything to get that number up.

So, reputation is tied to a user, and the more rep someone has the more they have contributed to the community. Make sense… except… this isn’t really a good representative of the quality of the contributions - Just the number of the contributions themselves.

Yes, answers that are better tend to get more upvotes, but the easiest way to get that number up is to flood the board and collect where you can.

So, as a new user I see this every day. Regulars that post one paragraph answers and are assumed to be the expert because they have more rep then me. I put effort and research into an answer just to be overshadowed by someone who has been around longer because they hopped on it quicker then I did.

I’ve almost quit the site because of the system. It’s a good idea, but I just can’t get around the “my number is bigger then yours” syndrome.


This comment is quite revealing, but likely not in the way you intended: #GamerGate was intentionally created by 4chan as a harassment campaign, using allegations of corruption to provide a shield for attacks on Zoe Quinn and, later, other women (this is exhaustively covered on Wikipedia). This was well understood by anyone who actually read any of the articles or primary sources being attacked; the people who lobbed words like grenades, however, never cared to do so because the point was spamming attacks, not understanding.

This is amusingly meta: you’re commenting on an article about people not reading before commenting to defend the practice by repeating propaganda which depends on people not actually reading something before repeating it. I guess you’re at least consistent in preferring conspiracist fantasies.


One gamergate editorial discusses the problems with “Real ID”.


But it has a lot to say about the quality of discussion and how to preserve it.

It opens thus:

Arguing on the internet is like trying to give a cat a bath. Whenever you finally seem to have someone cornered with evidence and superior reasoning, they somehow squirm out and reset the stand-off as if nothing had just occurred. We’ve all experienced that frustration of “winning” an e-debate only to realize that, since it’s on the internet, no one ever has to admit they were wrong. They simply move on, more often than not restarting the same argument hours or days later. It’s intensely frustrating for people who don’t understand that there’s no winning beyond appreciating the rush of the in-the-moment zing against some nasty internet troll. With that in mind, I understand the temptation to demand a way to hold people accountable for what they do and say online.


Banana++ Great read!

I wonder if not just recording how long you’re on the page but also how long you’ve spend at a certain scroll position, you might be able to tell which paragraphs were read and which ones weren’t read.

This depends on the speed and comprehension of the reader. But these are things your system could learn over time to be able to tell that Bob reads really fast and makes high quality posts etc.

I just now created an account with facebook, have been a loongtime lurker. This is my favorite philosophy website. Stack Exchange and Coding Horror all day err day. :stuck_out_tongue:

@shapira_yehuda So not even the professor read the whole thing? :pensive:


I wonder if scrollbars might be too primitive for reading. We’re certainly used to them, but but they’re in the context of document height or section height.

Maybe a pagination + scrollbar combo that isn’t necessarily tied to the content of the webpage but the content within the article. I probably should draw a picture, but I mean something similar to what discourse has for comments, except pagination. When the scrollbar reaches the bottom of the page, the next page automatically loads and the previous page automatically closes.

Its hard to describe but the idea being that you’ve always got two pages loaded, the scrollbar shows you where you are in the context of the later of the two pages that you have loaded, and the pagination keeping track of which page of the article you are on.

If you want to scroll beyond the article, that could perhaps be tied to scrolling outside the visual boundary of that article.


One more thing @codinghorror I’d like to point out about people not finishing an article, and I’m not sure that you’ve covered this before, are that some authors are not considerate to their reader’s time.

To the authors that do a hasty mind-dump to their keyboard and don’t proofread or get to the bloody point and instead fumble around for 10 pages of ICANTEVEN, they sometimes are deserving of the tl;dr.

And honestly, you aren’t the average blogger. Yours are courteous to your readers’ time, thought provoking, and worthy of having been written.

There are so many mind-dump blogs out there that, we may have been conditioned to scan. We don’t have all damn day to read through 10 pages that could have been condensed down into 2 1/2. :unamused:


If readers only read 50% of the article does that suggest that the article is too long for the web? How about writing shorter articles and get to the point? Personally, I don’t read web pages like a book – I scan through it. You need to visually grab my attention with stop words or get to the point with the first sentence in a paragraph.


This is one of the best articles i’ve read on your website.



I think it is interesting that in the gamification paragraph there’s a typo - “but if you’re going to reward someome” - is the incorrect spelling of “someone” there a game that nobody has picked up on yet or just an honest typo?


Congratulations, you’ve won this t-shirt:


I’m curious if you think there’s a strong difference between forums and stack overflow sites. You obviously felt strongly that both had a place in the world, since you worked two create two great versions of the two technologies, but why not just combine the two, ensuring that there is a place for people to discuss random things, or get specific answer if they want them. The most problematic aspect then becomes the voting functionality disorganizing comments that might need to appear in chronological order, but I can think of ways to solve this. I think in many cases, one or the other might be a better fit, but I’m starting to see a lot of companies create a space for both, and this seems like a waste of resources and space, and will make it more difficult for users to find the answer they’re looking for in many cases.

Do you have any other reasons why you feel strongly that these should be two distinct/separate technologies?