Show HN: Top books mentioned in comments on Hacker News
hackernewsbooks.com
hackernewsbooks.com
I’ve added a newsletter in case people want to keep up-to-date and I plan to maintain the service as I’ve found some unique books while I was coding the website. There is also search that searches only the submitted books, so this is another good way to browse the collections. Currently there is a newsletter popup, which may be annoying to some users, please click away if that is the case. I will rework it to be less intrusive.
Any recommendations and feedback are welcome. Disclaimer: For amazon links there are referrals, the other links don't have those.
- I want this & will use it, it's a great idea!
- I didn't understand the rankings based on visiting the site, I assumed it was number of mentions, but so far after some browsing, I've only seen a single comment listed under any book. Is karma used only as a tie breaker, or can a high karma with low mentions rise above a book with high mentions?
- I'd love to see a top 10 monthly or yearly. Weekly is just so much volume. And it looks like weekly lists usually top out with a single mention? I want to see which books get many mentions, whatever time window that takes. Also I can't read a book a week, so I don't shop for books weekly. This means I'm afraid with a weekly list I'd probably miss all the good stuff.
- The signup popup came up too quickly. I see other people complaining about it and saying remove it. But I would probably have signed up if I understood what I was getting. Seeing it before I even have a chance to see the value or understand who and what this is made me dismiss it immediately. But I'd gladly give my email address to get a monthly email with a top 10 list, and speaking personally it'd help me to have a sense for whether this is strictly a side project, or you'll be doing some marketing. The aggressive popup w/ little info just makes it feel like I'm getting trapped into something and I try to escape, when your intent may have been 100% to generously give people a no-nag opt-in reminder.
- The rankings use karma and number of mentions, which for a week basically bring books by top karma users to the top. The actual formula is:
book_entry['rank'] = (avg_karma_count * avg_karma + book_entry['karma'] * book_entry['karma_count']) / (book_entry['karma_count'] + avg_karma_count)
- I plan to make Top 10 monthly or yearly, just wanted to see if people would like the idea at all. Seems like people like it. As a side remark I also did http://hnbuzz.com/ where you can filter, segment HN posts daily, weekly, monthly etc. It is not about books though.- Yep the signup seems too aggressive, it will get changed :) After I clicked it away the first time it didn't bother me anymore while coding the website so I got carried away and left it as is. But it is also a learning experience to see what people think about those.
The project is totally side project, I actually do that often - build something I like to have and use a new tech to do it to learn. Drop me a mail if you want to know more.
Second this. Additionally, an "all time" list would be nice. Seems to be a great way to pick out the "timeless" books.
The modal dialog to subscribe to the mailing list is annoying; why is every site doing this now.
I too usually don't put a link to Amazon when mentioning a book, but maybe I should.
I'll be picking up that Palm OS guide as soon as I can find a copy in print somewhere, thanks!
I think the idea of this is great, but I suspect it'll become far more useful given larger aggregate ranges. Last week, last month, last six months - etc. As it stands, many books (like the Palm OS one) may only get mentioned once by someone with a lot of karma - which leads to another issue, since the algorithm is known it's now open to abuse by the HN karma-rich.
I will often mention books without posting an Amazon/etc. link figuring that anyone interested can bing with Duck Duck Google to find it. Other times, I will link to the author/publisher page.
Anyway, there's a natural language processing cum machine learning project in their somewhere. One interesting possibility it would open up is surfacing books in electronic form that pass across Hacker News as links to PDF's, eBooks, and HTML.
Good luck.
At least 12 of the books in the top 100 for 2015 are from my posts.
Other than that this already looks pretty useful :)
sum([u.karma for u in users_who_mentioned])
Or something more complicated? book_entry['rank'] = (avg_karma_count * avg_karma + book_entry['karma'] * book_entry['karma_count']) / (book_entry['karma_count'] + avg_karma_count)
Where avg_karma_count and avg_karma are the weekly values.Here's 2015: http://hackernewsbooks.com/year/2015
So just commercial books then? Plenty of excellent books are linked directly from HN, books with their entire contents online.
how are those accounted for?
Edit: you don't have to mention that in the site, but mention it in your description above.
This posting, as it was (before he added the disclaimer in his description of the site in this thread), was basically the same as posts to blogs that are not much more than thinly veiled sales pages.
I don't post/advertise my profile page link, people visit it on their own without me asking them to [check it out] (just like I don't expect him to tell people on his site that they are affiliate links).
Edit: It appears the mods updated the title to be a ShowHN. Thank you Mods.
In fact track the monthly revenue for a year and do monthly updates.
"Winning Through Intimidation" is not about what you think it is. It's actually a fascinating story about a fellow in [commercial] real estate from the 70's? and the things he learned. Somethings include "marketing" and projection and how he went about getting big deals done. One thing I found fascinating, and perhaps where the title came from, was after he got burned a few times he started thinking "If the seller has a lawyer and the buyer has a lawyer (to get through the paper work and finalize the deal) in the room (to which he wasn't in) then maybe he, the broker, should have a lawyer in the room too since he too was a party to the transaction but didn't have fair representation. So he started hiring his own lawyer to sit with the other lawyers to hash things out and soon learned that the buyer and seller lawyers didn't seem so work so hard to find a way to remove him from the transaction and not pay his fees.
As closings go, this was as good as it gets, because all three parties involved-the buyer, the seller, and the real estate agent-were working hard to accomplish the same objective. Then, as we progressed toward a closing, I observed a phenomenon I was later to discover occurs prior to most real estate closings. My professor [1] "sharpened his pencil" (his words) and continued to find one cost after another that he either had not previously considered or hadn't known about.
...
The closer we got to the closing, the more my professor sharpened his pencil. Being a Type Number Three [2], he was very nice about it and never came right out and said that he did not intend to pay my fee. He just mumbled a lot of negatives, and the more figuring he did, the more he mumbled-and the more concerned I became.
In fact, my concern became so great that I dared to ask myself a question that only an inexperienced reptile [3] like me would dare ask: "If the buyer and seller in a real estate deal are represented by attorneys as a closing, why shouldn't the real estate agent also be represented by an attorney?"
...
After all, didn't I have a vested interest in the deal, too? Nonetheless, ... I thought ... my professor would take it as an insult to his integrity and probably use it as an excuse not to pay me at anything at all. In a sense, I was being intimidated by my own thoughts.
Finally the big day arrived-my first real estate closing. I talked to my professor just a couple hours before the closing ... he said he felt terrible about it, but, after adding up all the figures, there was no way he could spare even $100 out of the proceeds of the closing, let alone $6,500.
...
I then made a bold decision. I scurried over to the office of an attorney friend of mine, showed him the document my professor had signed, described the conversation I had just had with him, and explained that the closing was about to take place. The attorney and I then went over to my professor's office and found that all the parties involved were in the process of preparing for the closing.
Whereupon my attorney sat down with the other two lawyers, and the THREE of them went through the mechanics of finalizing the deal. Although I didn't understand the principle at the time, what I had going for me at that closing was the unwritten, universally accepted understanding among all attorneys that I subsequently dubbed the Universal Attorney-to-Attorney Respect Rule. It's kind of analogous to "honor among thieves." ...
...
P.S. I got my $6,500 fee-at the closing.
...
First, not only did I not have a written agreement with my professor, I didn't even have a clear verbal agreement. Our understanding was vague, at best. all I had done was volunteer to try to "solve the financial problems in Cincinnati" and if "successful" (which also was not defined), I was supposed to be "paid handsomely in return." In other words, there was no agreement, written or verbal, that morally obligated me to present my professor with offers to buy out his interest in the property. I was strictly on my own, and my fiduciary responsibility was to myself.
...
In addition to learning how a Type Number Three [2] operates, I received a bonus in this deal in that I got an answer to my naive question, "Why shouldn't a real estate agent also be represented by an attorney at a closing?" Firsthand experience had emphatically given me the answer: "He should!"
Regardless of the business you're in, never allow yourself to be intimidated into believing that you aren't entitled to the same rights as the so-called principals in a deal. I say so-called because, from your standpoint, you ARE a principal. If you have a vested interest in a deal, you have a right to protect that interest, regardless of the size of your stake relative to the other players' shares. Just don't expect the other principals to agree with your viewpoint. To be forewarned is to be forearmed.
[1] he refers to the people he learned a lot and got an expensive education from as his professors... these would be either the buyer but more often the seller
[2] Type Number Three, who, like Type Number Two, assures you that he's not interested in your chips. Unlike Type Number Two, however, he sincerely means what he says. But that's where the difference ends. Due to any one of a number of reasons-ranging from his own bungling to his amoral standards for rationalizing what's right and wrong-he, like Types Number one and Two, still ends up trying to grab your chips. Which means that his supposed good intentions are irrelevant to the final outcome.
[3] he refers to himself as a reptile like a tortoise.. he's "slow", but slow and steady wins the race
EDIT: What is missing above is the setup. How he went about trying to obtain the deal and "setting" the buyers offer price. He opened himself up to be easily squeezed and excluded from the deal. Luckily he had the seller sign a "letter of understanding" as things started unfolding that mentioned his fee so that his lawyer could at least retain it and keep the deal on the rails.
I think it's an honest mistake. The fact that you dismissed the popup seems to be stored in a cookie, so it bothers you only once ever (twice if you use the back button).
Sidenote: since I'm in the EU I think you need to need to ask me for permission, but let's not open that can of worms.
Indeed, get rid of it, your site design is so sparse it's hard (I think) to miss the "Newsletter" item at the top. Or put it inline somewhere else, e.g. at the bottom of the list, if someone's bothered to scroll down that far the odds they're interested in the newsletter are quite a bit higher, I'd think.
I can't help but see the world as Fragile Robust Antifragile.
Getting back to the individual though, it's very hard to know how big of an inference gap there is for the model(s) you're trying to convey and how much the person "knows" already. Some books are masterful if you've read precursor books a, b, f, and p. Some books are masterful if you're coming from the right cultural context. And then there's tightly held biases that can get in the way of some core assumption of the book, ruining the whole effort.
Top 30 books ranked by total number of links to Amazon in Hacker News comments http://ramiro.org/vis/hn-most-linked-books/
I am surprised that someone doesn't solve this problem once and for all by allowing decent search parameters (e.g., all time, last month, by karma, by genre, etc.).
This latest submission weights by karma and bins the results weekly. Those restrictions make it personally unappealing but it is a great first step. I wouldn't have thought to consider karma.
Edit: clarity
He/she did a lot of work to make this project and to maintain it. I think they are done tastefully and sensibly in a service that adds value.
Moreover, I'd like to add that affiliate links are less intrusive and dangerous wrt to malicious code and user privacy. I wonder why Adsense et al rarely get called out, but affiliate links do. Maybe it's my confirmation bias or it's because people don't see those other ads, because of using an ad blocker.
I think your hunch is correct — we hover over the links and see it in the query string. There's probably a Chrome extension out there somewhere to strip affiliate links from URLs, but honestly they've never bothered me because they don't detract from my experience.
When somebody posts a link to Amazon in a comment thread that contains an affiliate code, I think that's fair to criticize. When somebody points to something they built that uses affiliate links, it comes down to how it's presented (and more importantly, how often it's reposted).
Who wouldn't be interested in a followup post 6 months from now with a breakdown of traffic and revenue?
As soon as I heard books I automatically assumed it would have links to amazon. And if you are linking to amazon from your own website you are stupid if you aren't using affiliate links.
Will soon code an aggregation for the two years, good point.
https://www.amazon.com/Feynman-Lectures-Computation-Richard-...
With the frequency of comments being 1 for each book, it will ultimately come down to karma ranking of each comment author.
- People online right now according to GA is about 500 and has been stable like that for the last two hours or so
- ~80% of the traffic comes from HN, rest is FB, Twitter, etc.
- Last two hours - 9,414 users, 16,338 pageviews
Will write a blog post with more info.
The nuanced answer is that there's a bit of a sliding scale. Known users, who have been participating in the community for a while, can do some things that might be spammy in other contexts, such as when a new account is created only to promote something. Of course that only holds as long as they do it tastefully (e.g. don't try to hijack someone else's launch) and don't do it much.
People seem to mind if you post referral links to threads as replies.
Internet citizens are sort of programmed to attribute high value to items near the top of a list when no explicit negative context is mentioned. What I am trying to say is that it would be truly beneficial if there were a bit of context mentioned along with the book, but I guess it would have to use AI/ML for NLP which is by no means a trivial thing. But that would make this spectacular. If that is not feasible (for whatever reasons), I suggest there be a prominent message saying to the effect of "this is not a recommendation list but just lists top mentions made by HKP".
https://news.ycombinator.com/item?id=12309500
> To this day, I recall with fondness poring over Palm's documentation and the book by Neil Rhodes and Julie McKeehan [1] and writing code using the MetroWorks CodeWarrior for Palm OS [2].
Like others have said, for the love of god please get rid of the popup subscription, or at least make it not alter history.
I ask because I was recently rejected for my application (http://addonbuddy.com/). Their reason was "lack of original content." Here's part of the message I received:
"A part of our criteria is that your site has to be established with enough unique content. We rejected your application due to one or more of the following reasons. - Lack of content which is original to your site and beneficial to your visitors - Pages that are mainly empty when advertisement content is removed"
I like your site and think it offers a lot of value, but Amazon's affiliate program seems to favor blogs that have a lot of content. They can also be real sticklers about including certain text in your website indicating that you're part of the affiliate program.
https://news.ycombinator.com/item?id=11704954 --> "I loved this book for being exceptionally clear and terse. I was hooked from the first sentence: "Probability is a mathematical language for quantifying uncertainty." That one sentence makes the concept clear in a way that the entire chapter on probability from "Statistics in a Nutshell" (http://www.amazon.com/Statistics-Nutshell-Sarah-Boslaugh/dp/...) did not."
--> "Statistics in a Nutshell" got added to list 20 of year 2016
That said, I have zero experience working with neural networks, so it might not be very easy at all.
Fundamentally, I don't have any problem with it. If you're going to link to something anyway, as would be obvious with a book list like this, it's a no-brainer to use affiliate links, get a cut, and earn back some hosting fees. The price is what it is for the consumer, whether they get there through an affiliate link or not.
This posting, as is, is basically the same as posts to blogs that are not much more than thinly veiled sales pages.
Great site, looking forward to the new reads.
Example #17: http://eli.thegreenplace.net/2009/10/07/book-review-c-interf...
For example, for one book in the list the review (with lots of ALL CAPS SENTENCES, but still interesting) in 5th place of the negative reviews by "truestara" seemed to be a good summary: https://www.amazon.com/Iodine-Crisis-What-About-Wreck/produc...
nkurz has over 37,000 karma, and is currently about 30-35 in the list of people by karma. https://news.ycombinator.com/leaders
edit: meant @DanBC in above line but wrote nkurz. Never mind - thanks to both. :)
Is there an 'all time' list that would be better? This contains a lot of garbage that got referenced once in an argument (and often a bad one at that).
It would be concerning if unpredictable, unrelated things didn't appear, still more concerning if all the things made sense and were correlated with some (any) convention, and most concerning of all if the convention were that of a majority.
I think it gives too much weight to me.
Look at 2016, and there's a bunch of books I've linked to. I believe these are great books, but it's a bit weird to see them so high on the list.
how are those accounted for?
I remember loving them then going to hate.
Guess it's where you are in life.
Which is really disappointing.
I should apologize to "someguy@gmail.com" and "ihateyou@gmail.com" for all the things I've signed you up for.
I get tonnes of emails at my, what i would assume to be uncommon [navarr] @ [gmail] address getting signed up for all sorts of things. Dating websites, Car dealerships, ugh its awful. Do they not accept @example.org which is a reserved domain? That would be a much better address to use
EDIT: You guys suck [1]
The more annoying thing for me is that they often won't accept my actual email address. It seems to be due to either the period in my name or the hyphen in the domain name.
Or hit back and don't sign up at all.
http://www.theverge.com/2016/8/23/12610890/google-search-pun...
It looks like the author has removed the popup.
Pleeasse don't do it... especially on a website whose target audience is the HackerNews folk, because they will not put up with it.
Those pop-ups are jarring. I think it's great you've got an affiliate link, and I get you want to get as many recurring eyeballs on this as possible, but those pop-ups are everyone's pet peeve... it's the MIDI autoplay, MARQUEE and BLINK tag of websites today.
Edit: Removed capitals for emphasis and adjusted tone to better communicate my intention and comply with HN Guidelines.
The comment also breaks the HN guidelines by using uppercase for emphasis. Please don't do that.
I'm sure you'd get better conversion with it as the first thing at the top of the list, maybe with a pastel yellow background for the form.
It's sort of the same thing with pop-under ads or those video ads, sure you're likely going to get more clicks on those, but you've made your website look (1) sketchy (2) difficult to use and (3) undesirable to return to.