Computer Science

Realz Slaw

00:12

@vzn bad data :/

vzn

00:30

so you think the whole thing was a mistake? did someone find it or you found it?

Realz Slaw

yes

someone else found the mathematical error of the result, I found the program error

vzn

hmmm ok

a worthwhile direction of investigation nevertheless it seems

nothing ventured nothing gained

am thinking ought to do a blog on primes or riemann or something like that

awhile back sketched out another idea wrt factoring in SAT to explore. think it could go far toward explaining the hardness of the problem & maybe others. supported by many prior papers.

need to write up some of that too sometime...

hey are you still writing up an arxiv paper? hows that going? marzio recently published a new one that got major pro attn, did you see that?

Realz Slaw

no

link?

I haven't started an arxiv paper

vzn

didnt you say you were working on a paper?

Realz Slaw

I was contributing to someone elses

vzn

00:43

no kidding who?

in theory salon, yesterday, by vzn

Mario di Biasi showed Min TSP Tour co-NP-complete answering question http://bit.ly/1dsiygZ based on my blog post http://bit.ly/19V1FPQ

Tweeted by fortnow on March 20, 2014 at 7:54 AM

Realz Slaw

and contemplating writing up the answers I have on cs.SE

I am unsure if he wants me to say publicly

but not someone famous :P

vzn

marzio told me in chat the paper duplicates an earlier result (oops)

lol ok well if the paper is published it will be public soon enough...

what is the topic?

Realz Slaw

it is an NP-reduction

have you seen Vor's website?

he is way into NP-hardness reductions for games and puzzles :D

vzn

yes its a cool web site. might have commented on it dont recall

yes computational hardness of video games is a neat advanced topic. theres several papers out on it now. just saw a new one.

Realz Slaw

haha

vzn

00:47

another good blog topic, gotta cover that sometime.

Realz Slaw

I actually dreamed/designed a game that has NP-hard puzzles

a 3d game

vzn

a nice mix that takes away some of the formidable aura of TCS...

a funky factoid for you:

the chat software tracks top writers (by lines) in a chat room.

guess who's #1?

ps babibu your old friend showed up recently :\

Realz Slaw

lol

vzn

he seems to have lost interest in TSP & is now working on sorting

with similarly stellar results :p

Realz Slaw

lol

vzn

00:57

r. spent a bit of time going over it... lol

Realz Slaw

well sorting is fun

so is TSP

either way he learning things

so thats some good news

vzn

:) gotta run l8r

Realz Slaw

cya

vzn

01:14

fondue break :)

spking of that, you defn have real teaching abilities... teaching in chat really challenging

Realz Slaw

01:25

haha thanks

vzn

btw disappointed you deleted the question. think it served as useful record. now theres nothing. also went to some trouble to publicize it elsewhere eg MO, & now its just a 404 dead link.

Realz Slaw

:/

but it is totally wrong :(

vzn

esp since you dont blog...

Realz Slaw

I don't want to keep the upvotes etc.

vzn

it serves as a useful record of a lot of decent effort.

imho you deserve the upvotes even with the mistake.

Realz Slaw

01:26

nah, it was a very trivial mistake

vzn

dont like how many halfway decent questions vaporize without trace around se.

disagree "trivial"...

poof

Realz Slaw

heh

vzn

science advances thru this process, & its not always helpful to "airbrush away all blemishes"....

Realz Slaw

but IMO it was a non-question

vzn

aka photoshopping

Realz Slaw

01:28

the question contained a mistake

vzn

2020 hindsight

just sayin.

ofc its your choice but want to let you know.

←leaves up all kinds of low-voted Q/A on se sites based on this "let it all hang out" philosophy

heck you even got upvotes. hah. whatever man.

Realz Slaw

lol

vzn

←gets a lot of flak for it too. whatever

Realz Slaw

I am still doing R&D on this though

looking for correlations between different tests and prime-ness

vzn

ok.

let me know if you ever want to consider graph-analysis-based approaches

seems quite promising to me

Realz Slaw

01:32

I can't really spend a lot of time on this

this is actually relevant to one of my projects

vzn

based on solid theory

another top secret project...?

Realz Slaw

yes :D

vzn

ah you would fit right in the "defense" industry

right there with Nakamoto :p

Realz Slaw

lol

an awesome figure, right?

the lore

vzn

spking of that a bit, a new beta, have you heard of it

Realz Slaw

01:35

nice :D

vzn

there is also something like area51 open science proposal... guess you saw that?

Realz Slaw

no

vzn

did you hear theres a $30K+ fund of BTC donations for dorian nakamoto? crazy world

Realz Slaw

I don't have time to spend on browsing area51 :P

yes

vzn

me either but some topics/proposals are way cool...

Realz Slaw

01:36

here is an idea

ability to upvote with bitcoin on SE :P

I wish

vzn

41

Open Science

Proposed Q&A site for scientists interested in openly sharing their research; for anyone interested in open access to research results; for publishers, librarians and other Open Science stakeholders.

Currently in definition.

Realz Slaw

but it is against SE policy

i.e they are against monetary tips

vzn

wow was just thinking something along those lines myself.

se needs to figure out a way to build incentives/prizes.

Realz Slaw

ah yeah I saw this @vzn

I am already following

vzn

there would be mass support for that.

Realz Slaw

01:37

they have a blog post on it somewhere

er no

it was a long meta answer

vzn

they do have a program to donate $100 to charity for each mod. did you hear of that?

over $30k

Realz Slaw

no I haven't

vzn

do you have that link to the meta answer? yeah am very interested in that kind of stuff

Realz Slaw

no, but you can easily google it

meta.stackoverflow.com/questions/97657/…

here is one

but there are bigger ones

it gets asked often

vzn

yeah saw someone talking about it on tcs.se once

heres that 2013 giveback announcement

Blog - Stack Exchange: Stack Exchange Gives Back 2013

Robert Cartaino on December 23, 2013

This is a time of year of traditions and celebrations — and we have a tradition at Stack Exchange where we set this time aside each year to give back to the groups and organizations that need our help. Each year, we reach out to our moderators and offer to make a $100 donation to charity on behalf of each moderator for their Stack Exchange community. It’s just a small gesture of thanks for the tremendous amount of work every community has contributed to make this entire thing possible.

This “giving back” program actually goes waaay back to the beginning when we started with on …

Realz Slaw

01:44

nice

vzn

they used to have a stackexchange store with some swag, it got closed. too much administrative difficulty apparently.

they would occasionally award prizes from the store.

thought they would continue to scale it up but it went the exact opposite direction.

its like how as corps get bigger they take away benefits instead of adding them yaknow...

my big question for stackexchange is their so-called "exit strategy"...

do they want to go public some day? acquisition? usually staying private is not so much an option for dotcom startups with VC funds/investors chomping at the bit....

iiuc se started out as a ycombinator prj....

Realz Slaw

mmm dunno

vzn

wondering, do you have a day job?

Realz Slaw

no

vzn

more time for research :)

Realz Slaw

01:55

:)

Hellovart

anyone know sql or database

Realz Slaw

yes, but cs.SE prolly not right place for it

vzn

Hellovar sounds more like a stackoverflow thing

Realz Slaw

Hellovart

why does everyone say that!

Realz Slaw

01:56

or that

@Hellovart cs.SE is more theoretical

not programming directly, unless it is about an idea, not an implementation

so, basically it is language agnostic

@Hellovart if your question is about databases in general, then it is appropriate

and ofc you can use sql to demonstrate the question

but if it is about SQL, then it is prolly not appropriate

anyway, doesn't hurt to ask it in chat

Hellovart

well it's about normalization

Realz Slaw

no one's gonna downvote you here :P

could be appropriate

Hellovart

i dont understand unnormalized form

lol

Realz Slaw

users (userid, username)

Hellovart

the grouping of columns

Realz Slaw

01:59

transactions (recipient_username, amount)

you see I used username twice

Hellovart

yes

Realz Slaw

if the username every changes, I have to change it in both places

which can be error-prone, and annoying

if I had used user_id instead, it would be normalized

then, if the username changes, it requires changing it in one place

Hellovart

see that I get but it was the way my professor is teaching me, she is incredibly incoherent

Realz Slaw

heh

grill the prof

you paying

if you listening and don't understand, don't let it pass

Hellovart

i'm paying to teach myself evidentally

Realz Slaw

02:02

lol

Hellovart

or paying to ask you in this case

Realz Slaw

well no problem here

I like teaching

Hellovart

then you should teach at my college because they hire graduate students

which is retarded

Realz Slaw

heh

not a graduate student :/

Hellovart

why not

Realz Slaw

02:04

too expensive

Hellovart

tell me about it, i had to pay an extra 250 dollars for a storytelling class,

has nothing to do with computer science

Realz Slaw

yeah, the system is insane, don't get me started

they'll lose out in the end though

Hellovart

like, i want to be able to get a job without having anything on my resume

is that so hard to ask

Realz Slaw

intern ?

Hellovart

no

actually my program has co op and if i keep a >= 3 gpa i might get in

Realz Slaw

02:06

nice

Hellovart

you make money on the job and it counts for credits

Realz Slaw

ah thats nice

child labor :P

jk

Hellovart

postimg.org/image/isjowvyeb/

postimg.org/image/isjowvyeb

ok so that's a table

i dont understand how to make that in UNF

clearly the projnum is the primary key

Realz Slaw

well

you can just make one table for UNF

I assume that means unnormalized form?

or somesuch?

Hellovart

yes, theres unf, 1nf 2nf and 3nf

Realz Slaw

02:08

ah right

Hellovart

only those steps we need to know

Realz Slaw

gosh its been a long time

Hellovart

i failed test #2 im sure it was on normalization

got 99% on the normal sql commands

on first test

Realz Slaw

mmm yeah, I don't remember the differences between 1nf 2nf etc.

Database normalization

Database normalization is the process of organizing the fields and tables of a relational database to minimize redundancy and dependency. Normalization usually involves dividing large tables into smaller (and less redundant) tables and defining relationships between them. The objective is to isolate data so that additions, deletions, and modifications of a field can be made in just one table and then propagated through the rest of the database using the defined relationships. Edgar F. Codd, the inventor of the relational model, introduced the concept of normalization and what we now know a...

Hellovart

1nf gets rid of the multi valued dependency

from the the unf

whatever the hell that means

Realz Slaw

02:10

so if you want it completely non-normalized,

you would make it one huge table

and duplicate

so the table-as-is

but, for the blank spots, you would duplicate it

Hellovart

yes, that's what my other examples look like

like: projectemp[projnum(pk), projname.....

apparently projects can have multiple employees

Realz Slaw

there is no need for a pk

because it has duplicates

well

projnum would not be pk

it would be employee-name

or num

oh wait

that is duplicated

it would have to combine several columns to make a pk

Hellovart

employees are used more than once in projects

Realz Slaw

right

so (project-num,employee-num) combination can be a pk

Hellovart

the professor said there needs to be at least one pk in the unf form

Realz Slaw

02:14

mmm

Hellovart

UNF*

Realz Slaw

well, you can add a unique-id column

or you can combine two columns for the key

vzn

23 mins ago, by Realz Slaw

no one's gonna downvote you here :P

lol except for all the mods & acolytes who will downvote/close it :(

Realz Slaw

here= in chat

vzn

exactly. all 3 mods hang out here & also juho, who has voted to kill quite a few of my questions :(

Hellovart: look into MOOCs. wave of the future.

Realz Slaw

02:28

my point stands, you can ask in chat, without worry

vzn

oh ok

Realz Slaw

ok I g2g

vzn

Hellovart: you may be "paying" but rs is not recving...

all unpaid volunteers around here

Hellovart re your diagram. any time you have repeated rows, its like a redundancy. imagine attempting to eliminate repeated rows by creating new tables with new relationships. in other words there are relationships hidden in the given unnormalized tables. what are those relationships? find new relationships & the new tables associated with them.

the diagram is slightly misleading because it doesnt repeat the columns, it only has spaces where there are repeated columns, but imagine filling those in to see the issue.

think like a db: imagine having to store the same strings/numbers over and over. inefficient right? so normalization has some similarity to a data compression operation.

←got a software engr degree/BS yrs ago & was a bit shocked "long ago" that they didnt cover this in my edu either

btw the general topic of db normalization seems like it would be on topic for the group.

1nf, 2nf, unf, not so simple concepts even for experts

note the whole idea of normalization is that you can reconstruct (so to speak) "compressed" (normalized) fmt via joins.

so hv are you at a university?

what are you studying?

Raphael

11:05

@Gilles Ah, sorry; I only say the "such as" which is too minor. My bad. Should have read the edit message.

Looking at the right column, I suspect someone does not understand the starring feature. O.o

Patrick87

18:19

@Raphael Regarding the question I closed on summing Big-Ohs - both questions ask "why does arithmetic with Landau notation not work?" In both cases, the answer is that it doesn't and don't do it. At least, that's the gist I got from those questions.

Raphael

19:56

@Patrick87 I read the one you closed as "How do I analyse this algorithm? Here's what I've tried: ..." Parts of the answer are "not like this" and coincide with that other question, but that one does not answer how they analyse the code.

@Hellovart Easy. The question is: do you want to do these jobs?

Transcript for