One thing I find generally worth improving with these sorts of introductions is that they always seem to introduce data structures as a pick-and-mix bag of prebuilt parts, rather than a set of techniques which allow you to enforce properties which allow you to build objects that do what you want efficiently. Teaching the former is like teaching a physicist all the different kinds of gear systems, rather than the math that lets you model them.
Knowing how binary heaps and hash maps work is great, because binary heaps and hash maps are very useful data structures, but knowing the fundamental reasons behind them in a way that composes to building new things is much more useful. What happens when you want properties that someone hasn't already built a data structure for, at least as far as you can find? (Pet example: ordered container with efficient append and the ability to query the maximum of any subslice length k in O(log k). Extension: what are the minimal assumptions you need to do better, eg. amortized O(1)?)
I think this contributes to the attitude where programs are built out of monolithic, ill-fitting parts rather than actually looking at the problem and building the solution which does the right thing, the right way. (Pet example: if you designed websites as data structures and the browser to support doing so efficiently, 90% of requests on a site like Reddit would be a single `memcpy`.)
What I would attempt is rather than saying
"Here's one way we could do it: <explaination of pre-baked hash table>."
talk about the underlying tool first
"One tool that can help us is hashing: <explaination of general technique and why its properties are useful>. In this example we want <these properties> but can't immediately get it because our data has <these properties>. But we can apply hashing to get <technique that looks decent but has caveats>. Handling the caveats by doing <hash table things> gives us a data structure called a hash table."
Yes, this is probably a harder way to handle things, but I'd like to think it'd be worth the trouble.
That was a clearly-written, knowledgable, info-dense and effective bit of writing you just did, @Veedrac. Ever considered doing more of the same, and delivering the kind of educational material you just described?
You can subscribe to Veedrac's HN comments on RSS like I do. I've found a bunch of people like Veedrac regularly write really interesting things in HN comments, so I wrote an app that lets you subscribe to them: http://hnblogs.thume.ca/
I was expecting a far more mixed reaction, so I'm glad to hear this engaged people. Writing educational material to takes more work (and expertise) than I'd be willing to guarantee, but I'll keep the sentiment in mind :).
> (Pet example: if you designed websites as data structures and the browser to support doing so efficiently, 90% of requests on a site like Reddit would be a single `memcpy`.)
You'd still want a compressed form, even if you had a good way to bypass parsing.
Ah, yes, I botched that description, since there has to be more to support things like encryption. A more accurate claim is that the bulk of the work currently done would be replaced with a memcpy.
> (Pet example: if you designed websites as data structures and the browser to support doing so efficiently, 90% of requests on a site like Reddit would be a single `memcpy`.)
Formatting a webpage cannot be done by a simple memcpy, especially since you don't know the size of the browser window in advance.
But perhaps I didn't understand the point you were trying to make.
I mean on the server-side. Right now the process between "I want thread 35va5k" and "I want to render this html" makes extremely inefficient use of both servers and network.
I'm all for stuff like this in theory, but in my experience I never use said knowledge for any productive purpose and eventually it degrades... until I read a post like this, feel a surge of professional inadequacy and go re-learn it. :P
I’ve wondered why we don’t typically use algorithms that will self-adjust when we have so many VMs that can do all sorts of statistical analysis. We know that quicksort is inefficient on smaller and mostly-sorted lists so we oftentimes switch to selection sort below 20 items or so, but we typically have libraries that have statically picked this behavior as directed by their developers instead of over time selecting the more appropriate algorithm. Sure, if you’be never seen the data before there’s not a lot you can assume, but if over a lot of executions your VM can perform native code substitution and replace certain sections of code out based upon runtime profiling in HotSpot why can’t we come up with heuristics that say “just use a list instead of a hash table until runtimes to access start getting slower than n, in which case reallocate the data to a hash table, btree, or a heap depending upon which operations are most sensitive to system design parameters.” A lot of problems seem like we’re repeating ourselves writing the base algorithms and gluing things together when we really should be investing problems around resource constraints and system trade-offs and priorities. Then we could run test suites against these systems to determine how well they’d meet certain guarantees programmed in.
>> A lot of problems seem like we’re repeating ourselves writing the base algorithms and gluing things together when we really should be investing problems around resource constraints and system trade-offs and priorities
What's the minimal platform that would enable library developers to monetize such things ? and does it or something close to it exist ? because that seems like the fastest way to make this happen.
A small thing: the slide in subscription popover contains language which is dangerously close to the dark pattern of framing refusals as guilt/shame inducing.
A simple no thanks would do.
Because “they won’t remember what you did but hey will remember how you made them feel” and this made me feel “eww”.
Interview cake is amazing. The way the problems are presented and the hints are given is often better than what you'll get in an actual interview setting.
After using it myself, I could not recommend it any more highly.
Nice! Actually in all seriousness the site was pretty awesome and helpful. Six different companies are flying me out to SF in a few weeks for final interviews with them all. Can't say what percentage of that was due to your site and what was general study, but I found your questions intuitive and super helpful. For real, thanks!
Grats on the interviews. It's a little weird to read this comment as someone in the same industry but in the polar opposite position. As another commenter asked, I'd be curious to hear about your path up to this point.
Sure, yeah, that's basically where I was coming from. I'm not 100% set on moving yet, just testing the waters. I have some personal connections at large companies that I intend to use to get referrals into the hiring pipeline when I do actually move, because that's apparently the only way to reliably get interviews at the huge places. For now, it was a lot of applying to jobs online with targeted resumes/cover letters at smaller to mid-sized startups in SF/bay area. There really just are a huge amount in the 20-50 person size range.
I found that once I actually got an interview or two, it was like a bit of a waterfall. I applied to companies X, Y, Z. When I got an interview at X, I emailed Y and Z and said I'm interviewing elsewhere and if all goes well plan to be in SF, and would love to interview with you. I think this is confirmation bias at play or whatever, because that really seemed to help get additional interviews.
Thanks for the info. You're where I'm planning on being in about a month or so. A handful of personal connections that I'm banking on to get my foot in the door. Fill that in with targeted resumes to my top picks and a maybe untargeted to a dozen or so others.
That's an interesting strategy. More specifically, when you emailed y,z saying that you had an interview elsewhere, did you have their direct contact information or did you just mention this in the cover letter box on their application page?
I've had the same concerns but I think--this is all my thoughts, no official sanction--Interview Cake started by catering to people who had interviews coming up, and so it didn't look like they will need more than a year's access. Neither would they need the material after they got the job. But I think @gameguy43 will see your comment and hopefully take it up with his team. I'd love to have eternal access, but then the 1-time charge might be bumped up too. Who knows?
Shoutout for including the bit about constants in O(n) notation, 99% of things I see cover it gloss over it.
I've seen radix sort walk all over quicksort for common datasets many times. Knowing your dataset is at least as important as knowing the right algorithm.
I really love this. Frustratingly few articles/posts/courses actually follow through on their promise to make difficult material intuitive. I'm only halfway through your Big-O post but it's already obvious that you've put a lot of care into really making it accessible.
That is soooo great to hear because every time we "finish" writing a new piece we end up spending literal /months/ obsessively polishing it before taking it live.
I always worry that we're just spinning our wheels. But it's great to hear that the extra effort is showing through :)
This is something I've entered my email into after at least a year or so. Great presentation and design and I'm sure it's good content too -- I'll get to know over the coming week :)
We have a team of like a dozen and a half contractors from around the world who do the translations.
It was important to us to not just translate our Python into C++, but to have C++ that was actually good, C++-style C++. So we hired C++ experts to do all the C++ translations. And same for the other languages.
Now, none of us are C++ experts. So how did we know if the contractors we hired were any good? That's a whoooole blog post for another time.
In short: Noah (an engineer on the team) came up with this clever system where we hired a bunch of people to translate the same code sample (as a paid trial task), then we hired a bunch of people to rate those people's translations (again, as a paid trial task). Then we hired the translator who had the best ratings, and we also hired the /rater/ who was most /thorough/--they do code reviews of the translator's code.
I use Swift and Objective-C, but I've consulted InterviewCake content for every other language except PHP at some point. The sample data structures and question answers are nice code samples when I need a quick reminder for how to do basic things in languages that I don't use on a regular basis.
Not having Go or Rust isn't a big deal, if I could choose one it'd be Go. I want to learn Go for myself, and asking for a friend for Rust.
I also never use the editor on the site, I write/run the code on my machine. I never liked using the editor the few times I tried it. I see in another comment you made in this thread that a new one is coming soon, so I'm excited to try it. Any chance I could beta test? :)
I think these styles of puzzles borrow a lot from competitive programming (e.g: ACM, IEEE).
And while competitive programming requires skill, what you do most of the time is collaborative programming: create software that interoperates with other software, with a user, that is maintainable, etc. Software without these skills quickly becomes really expensive and hard to maintain.
This is really beautiful. I love that the images are SVGs as well, makes perfect sense for these line drawings. Though it is a little bit of a bummer that all of them have embedded fonts, even those that don't use text.
Oh iiinteresting I guess we are doing that! Probably not too hard to fix. Thanks for pointing that out!
They used to be just inline svgs (instead of img tags), but (at least a couple years ago) our research suggested using img tags was better for SEO and screen readers and even page load time.
edit:
(Of course, when they were just inline svgs, they got the fonts for free from the surrounding webpage).
So far I'm thinking we'll just turn off the embedded fonts for images that don't have text. Definitely not expert on this stuff though--open to other advice!
Knowing how binary heaps and hash maps work is great, because binary heaps and hash maps are very useful data structures, but knowing the fundamental reasons behind them in a way that composes to building new things is much more useful. What happens when you want properties that someone hasn't already built a data structure for, at least as far as you can find? (Pet example: ordered container with efficient append and the ability to query the maximum of any subslice length k in O(log k). Extension: what are the minimal assumptions you need to do better, eg. amortized O(1)?)
I think this contributes to the attitude where programs are built out of monolithic, ill-fitting parts rather than actually looking at the problem and building the solution which does the right thing, the right way. (Pet example: if you designed websites as data structures and the browser to support doing so efficiently, 90% of requests on a site like Reddit would be a single `memcpy`.)
What I would attempt is rather than saying
"Here's one way we could do it: <explaination of pre-baked hash table>."
talk about the underlying tool first
"One tool that can help us is hashing: <explaination of general technique and why its properties are useful>. In this example we want <these properties> but can't immediately get it because our data has <these properties>. But we can apply hashing to get <technique that looks decent but has caveats>. Handling the caveats by doing <hash table things> gives us a data structure called a hash table."
Yes, this is probably a harder way to handle things, but I'd like to think it'd be worth the trouble.