Rack-a-thon? Is there any interest in an online hackathon-like thing in a similar form to the Zurihack ? ( https://zfoh.ch/zurihac2020/ screenshot in thread) ( I’d like to find a better term as hackathon has a bad reputation as being exclusionary to women and minority groups) :+1: for yes :–1: for no,
I like the idea of https://zfoh.ch/zurihac2020/#beginner-track\|beginner and *https://zfoh.ch/zurihac2020/#advanced-track\|advanced tracks.*
Whatever name you choose, maybe just say early on that everyone is welcome and include a code of conduct?
Yes good point. Will do.
One of the impressions I have of hackathons is there are sometimes a few people who know exactly what they want to do, and a larger group of people who have much less of an idea. So when I think of organizing events like this, I think in terms of: people with a goal already, people with lots of questions, people who don’t know what questions they should ask, people who want to show their skills, people who are afraid to show they know less, people who want to learn, people who might see something and then get excited, people who have a vague notion of what they want to build but would like some advice from people who know more to learn how best to start and build the thing and test it, etc. There are lots of different kinds of people. Then there are organized larger Hackathons (e.g., PSUhack at Penn State where I’ve volunteered as an AccuWeather person). Those are a bit different. Companies sponsor them and give prizes. So for AccuWeather, we hand out developer API keys for free valid for 24–48 hours, tell students all about the API, and offer prizes for the best weather related [whatever they build and it can be just about anything]. Usually the winner just has a really creative idea that’s just our API, or it’s something that uses our data along with the data from another company that’s at the hackathon.
I’ve only recently become aware of Twitch. ScalaLove couldn’t get on YouTube for its conference, so did everything on http://Twitch.tv\|Twitch.tv, which is normally for gamers, so we had the whole conference on Twitch, then it was changed over to YouTube (for archiving) about a month later. Anyway, I discovered this whole world of people who are just live coding stuff with people from around the world watching. It’s pretty random but sometimes very interesting.
Hm. I see the solution you offer here, but maybe not so clearly. (Maybe I’m requesting the wrong thing.) While it’s true that any comparison sort algorithm is in the class Omega(n log n), I’d like a way to somehow distinguish the n log n ones from all the others. If they’re all in Omega(n log n), I can’t use “Omega(n log n)” to single out the fastest ones.
My intuition tells me I’d like to say: some sort algorithm X belongs to Theta(n log n) while algorithms A, B, C do not belong to Theta(n log n), so X has this distinguishing property. However, I don’t have the Theta()-fact, so I can’t say it. All I know so far is that X is O( (1 + 1/lg n)^n ), where n is the number of bits in the input. All other algorithms (I’m comparing X to) belong to O(y^n) where y is greater than (1 + 1/lg n). They’re all superpolynomial, but X has the slowest growth.
It’s fun discovering such communities on Twitch/YouTube.
@me1884 has joined the channel
@me1884 Welcome :joy_cat:
Writing. Do you write O((1 + 1/lg n)^n) or would you write the prettier O(1 + 1/lg n)^n?
The first.
Really? That’s so ugly. I think I’d have to increase the size of the outer parentheses to improve it a little.
Find a book with an author with a good taste (like Knuth) and see how he does. Concrete Mathematics perhaps.
Excellent idea! Thank you!!
on big-O notation, I highly recommend https://cs.uwaterloo.ca/~plragde/flaneries/FDS/Tools_and_Techniques.html#%28part._.Order_notation_and_its_discontents%29
Ragde has a section on @Any ’s question: https://cs.uwaterloo.ca/~plragde/flaneries/FDS/Tools_and_Techniques.html#%28part._.Abuse_of_notation%29
He does like you. He says ln z = z − 1 + O((z − 1)^2). But he does increase the parentheses a little.
Page 445, second edition. Concrete Mathematics.
I’m lost here. My question is only on notation, how to write it. Am I abusing notation by writing O((1 + 1/lg n)^n)?
No that’s okay.
It’s O(f(x))^n that is abuse.
The idea is analogous to log n or sin x. Sometimes we don’t write the parentheses in “function calls”, say.
But the suggested “prettier” expression is not that kind of abuse. I’m just removing the parentheses from the “function” O just like we do with sin x or log x.
The idea is that people wouldn’t read it as a set raised to a power, but anyhow — the points raised here make it all clear. :slightly_smiling_face:
soegaard2 gave the idea of checking Knuth in Concrete Mathematics. He increases the sizes of the outer parentheses. I’m here to follow the rules.
We have a winner. Thanks everyone!
Btw - I stil think sin^2 x vs sin x ^2 vs sin x^2 is a mess.
I totally agree. But, you see, programmer-mathematicians [or mathematician-programmers] (as you [likely] are) are not totally like mathematician-mathematicians. :slightly_smiling_face: Mathematicians don’t really care for all the pedantry on syntax. They don’t see their art that way. That’s why they’re kinda messy. Though they all seem to agree [on] adjusting notation in each different context — kinda like Lisp programmers into language oriented programming. I can’t quite disagree with either group.
I also wish I could write 2xp(1 + 1/p) instead of 2^(1 + 1/p) just like I can write exp(1 + 1/p) instead of e^(1 + 1/p). :slightly_smiling_face:
As Sancho says, if you say so in the notation section and if it’s justified (you have lots of them) then at least to me it’s totally fine. What I dislike is people using untold local conventions, or preferring simplicity over clarity.
sin is different. sin produces a number, so you can exponentiate it. But big-oh produces a set. It doesn’t make sense to exponentiate a set (well, it actually does, but that results in a set of tuples, which is definitely not what you want)
And even for sin I really hate it as well.
sin^-1(x) is arcsin, not 1/sin(x). So much for this idea.