Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: Kooba, an interactive graph for finding new books (jake.run)
145 points by cloudbeary on June 12, 2017 | hide | past | favorite | 46 comments



Do many people have a hard time finding books to read? I have the exact opposite problem: I'm constantly overwhelmed by the tidal wave of books that is my to-read list. I'd give my right arm to be able to read 2x as fast as I do today.


There's always another book to read. But the books that I remember years later, the real gems, are few and far between. If I had some even marginally better process for finding those than dumb luck, I'd be thrilled.


There must be some fundamental difference between the way you and I read, because I feel like I'm wading through a sea of gems when it comes to choosing books. Every once in a while I come upon a dud, but not often at all. And the duds are basically always very specific to my personal taste. For example, you mentioned Anathem as a gem, and I love Neal Stephenson and have read most of his work, but reading Anathem a couple weeks ago I was bored out of my mind.

Google search and Amazon reviews are actually excellent book-finding tools and I don't believe there's much value to be squeezed out of what they leave on the table.


I hesitated to include Anathem on the list because it isn't like the others. I didn't love it because it a masterpiece of the novel form, but rather because I found so many ideas in it fascinating.

Perhaps if I had a more comprehensive background in a couple of areas (math, philosophy) I would have already been exposed to many of them and wouldn't have enjoyed the book so much. But as it turns out I didn't and I did.


Did you finish Anathem?

It was a brutal start for me, I didn't get into it until after the first 300 pages. Then it was fantastic. (I basically hate the overly descriptive setup because I don't really like wasting my time trying to visualize what the author is describing).


I didn't and for essentially the reason you described. The minutiae of the setting kept getting in the way of enjoying the story and the characters. Maybe I'll give it another go. Still, 300 pages is quite the hurdle.


I loved Anathem. Perhaps it has to do with the fact that I heard about the Long Now foundation and their work at roughly the same time. Stephenson's wanderings into tangential topics is annoying at times but I stopped judging and just read it like a stream of consciousness kind of thing and it worked out quite well.

Everyone I've recommended the book to has either been overwhelmed because of the size or bored because of the length.


Out of curiosity, which books?


I don't have a full list, but some that I thought of when I was trying out this website were Foucault's Pendulum; East of Eden; Sometimes a Great Notion; A Man in Full; I, Claudius; and Anathem.


> Out of curiosity, which books?

There were two great discussions about this on HN last week:

Ask HN: What language-agnostic programming books should I read? https://news.ycombinator.com/item?id=14486657

Ask HN: Books you wish you had read earlier? https://news.ycombinator.com/item?id=14477851


There's always a tradeoff between reading the next best book in your queue versus spending time sorting what the next best book would be. I certainly don't have the problem of not having a long queue of books to read, but always want ways to improve my sorting algo. I expect this might help me find lookalike books similar to my favorites, that might be more worthwhile for me to read versus my current queue.


I see what you mean. Something I've been doing recently to help with that problem is reading clusters of books that all relate to each other. For example, I'm currently reading a bunch of Hemingway's stories and next I'm gonna read a bunch of books centered on the American Revolution and the founding fathers. By clustering books the list I have to sort is much smaller.


Its a classic case of paradox of choice; having too much options makes it more difficult to pick any single one of them. The problem is not that I have hard time finding good books. I have hard time selecting which of the thousands of good books I want to read.


Reading the wrong books is worse than reading no books at all. Any tool that allows you to avoid 10% of the wrong books is worth it.


Despite having a formidable Mount TBR, I occasionally branch off into a genre I have barely touched and would like to read more of that genre (example I'm thinking of is horror after reading 14/The Fold). My to be read list is not a guidance, but a fallback :)


No but optimizing my queue is satisfying and much faster than actually reading.


I do i'm a picky reader and i stop reading more books than i finish cause I might not like the story.


Cool! I actually made a book recommendation service as well. [1] Its been indispensable for me, gives me much better recommendations than friends/Amazon/Goodreads give me. I've found Library at Mount Char and Hard Magic because I wanted something similar to Weaveworld [2] and both those have been perfect. I'm only two months into using it, but I feel like I have a never-ending reading list now!

I look forward to trying yours when it is back online (service was disrupted when I tried).

[1] https://www.booksuggestions.ninja/

[2] https://rpiai.com/why-i-made-a-book-recommendation-service/


I really like your blog post because it covers all the important and interesting parts concisely. It seems like the same can be said about other content on your website. I'll read more of it!


Library at Mount Char was reeeeally disturbing. Good, but ye gads... not what I was expecting... :)


Its funny, I was sort of expecting something disturbing because I got the recommendation through Weaveworld.


Hey! Wanted to let you guys know that we use Amazon data and currently the Amazon API limits the request rate so that if there's multiple concurrent visitors it might be temporarily delayed. Working on fixing it now!


Rate-limiting task queue and aggressive caching might help out... "Shown HN" loosely translates to "DDoS Me" :P


Yeah, at the very least slap a leaky bucket on there. https://en.wikipedia.org/wiki/Leaky_bucket

Also, a user navigating to /search/:title is currently shown a raw error .json that exposes both your HMAC signature and your AWS access ID


I'm still playing with it because of the, completely understandable, technical problems. But initial comments:

1) Love the idea, I'm always struggling to find the next book to read.

2) It seems to be pulling some scam books, I put East of Eden in and got some sort of help you with your homework summary book.

3) I'd echo the suggestion to deemphasize other books by the same author. If I loved Foucault's Pendulum, I almost certainly already know about In the Name of the Rose.



This is beautifully done. I'm happy about your successful implementation and half annoyed that now, I'm going to get even more books in my "to read" list. :-|

Back when I worked on openlibrary, we had something called "subjects" (e.g. https://openlibrary.org/subjects/architecture) and at the bottom, it would give out "Related..." books (by subject, place etc.). It was an extremely simple algorithm but gave me quite some nice things to follow.

Good work!


Interesting concept, and a really fun interface. The main issue I'm seeing is that the suggestions being surfaced are too close to what I've already selected - i.e., once I've indicated that I've read one or two of the books in the Hitchhiker's Guide to the Galaxy series, it should probably downrank the others, because while they are similar, it is likely that someone who's indicated they liked two of them has already read the others. There also seems to be an issue where the list of recommendations on the right still contains some of the books I've already liked.


Yes, totally agree! We're going to make adjustments to the algorithm so that books by the same author show up less (obviously if I'm searching for Harry Potter I'm probably not looking to be recommended six other J.K. Rowling books).


Is this just an item-item collaborative filter based on star ratings and cosine distance or do you have something more sophisticated going on?


As many have said here, finding books is easy, finding good books in a particular niche is hard. For me using this tool gave few new books to my reading list.

As a side note, I an quite social reader and like to read books that I can have conversations about. Amazon ratings etc. are quite good in that already, but the visualization helps big time.


Would really like a way to import books from my LibraryThing profile. I have too many to add manually.


503 (Service Unavailable)

Guess HN killed it :\


Suffers from my biggest complaint about modern JS applications as well: The assumption that the server is infallible.

The progress bar simply never goes away and my console is full of 503's. I sat waiting for like 30-40 seconds before I decided to check. Give me an error message so I know something went wrong at the very least!


If you're ready to invest some time www.librarything.com allows you to catalog your books and then gives you recommendations.

Still I hope to see Kooba working to find other books to add to my reading list :)


I would love to see this work better on mobile. I'm not sure if it's just not working due to the rate limiting but it's not really working for me (Safari, iPhone 6, iOS 10)


Suggestion: cancel pending requests when another one is requested.


On top of that, maybe only send requests when a user hits enter or submits the form. Could ease up the load a bit.

As of right now I am unable to get anything to happen.


I am wondering what api you used for book search, as I am currently thinking of undertaking a project involving books. Thanks :)


This looks awesome. Look forwarding to using it but right now getting some errors in the console. Maybe load problems?


When I click a details link, it sends me to the Canadian Amazon link. Any way to fix that? Looks neat otherwise.


Going to switch it to amazon.com!


Can anyone comment on the benefits of using the Goodreads API vs Amazon?


Never loaded...


It loads but it doesn't respond when I type in a title. Just three reddish disks that blink forever.


Same... typed in a book title in a couple different browsers, just the red dots appear :/




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: