« first day (4172 days earlier)      last day (1163 days later) » 

00:02
oh no
I didn't increase the width and height constants
I'll get my coat
Mines still running
Damn it Dijkstra, why aren't you faster?
yeah, first time I tried it went over 35 minutes before I stopped it to optimise and introduce bugs
Im gonnasleep and see what it outputs in the monring
@Wipqozn A* m8
@fredley I know, but I'm already 20 minutes into this
In economics and business decision-making, a sunk cost (also known as retrospective cost) is a cost that has already been incurred and cannot be recovered. Sunk costs are contrasted with prospective costs, which are future costs that may be avoided if action is taken. In other words, a sunk cost is a sum paid in the past that is no longer relevant to decisions about the future. Even though economists argue that sunk costs are no longer relevant to future rational decision-making, in everyday life, people often take previous expenditures in situations, such as repairing a car or house, into their...
00:10
@fredley Exactly!
Hopefully it'll be finish by the time I finish this spire run
learning more than one algorithm per day is against the law
@fredley lmao @ concorde being the image for this
00:27
I think the last time I did A* was for some kind of ant program in AI class
Although I'm starting to think I might need eat this sunk cost and just do it
THis is terrible
Was Dijkstra a fucking turtle?
@Wipqozn The "a" between "fucking" and "turtle" is making me raise my eyebrows
@PrivatePansy What are you talking?
@Wipqozn Mod abuse!
Proud of it too
I don't think I've ever had cause to use chat history purging for legit reasons.
I've done redactions in the main site for PII, but never in chat.
It's very easy to do a redaction wrong lol
On chat anyway
00:32
Can you do chat redactions? I thought you could only purge message history.
test oh my
I think you're mixing up main site and chat. I see no redact option for chat, just purge.
00:55
fucking zeus it's still running
I might just rewrite this with A* tomorrow
@Wipqozn You can purge history but you can't redact the current version
So a delete and history purge isn't a redact you need to edit too
I was wondering if A* has some weird pathological edge case but it looks like worst case complexity is just O(|E|) which is not terrible
I feel like Dijkstra shouldn't be taking this long. I suspect I fucked something up somehow.
01:22
Dijkstra's is O(E+VlogV) which is not that much worse than A* in the worst yeah
01:33
OKay so
I decided to do a quick timgin estimate, and basically my solution will take 11 hours to complete.
I think I know what's slowing it down as well
Problem for tomorrow.
 
2 hours later…
03:44
annnd done
@PrivatePansy Problem was that I was pre-populating the dist[v] with "infinity" like the algorithm normally calls for, but since I wasn't sorting it and also because the number of nodes was huge it meant the call to min(dist) took forever. Just populated the dist[v] as I went along instead, and just did a check if a node was present in dist. Super quick then.
 
4 hours later…
07:47
yeah, mine took 16 minutes, which while slow was a lot faster than I expected
shame I left it to run or I would have been within 24hrs
 
1 hour later…
09:08
My mod swag arrived
you were a true mod now? you are a true mod then?
I will have had been a true mod
09:37
@Wipqozn awww dang it. I thought that wouldn't work so I didn't try
 
3 hours later…
13:00
@Wipqozn I did A* and clearly I did it right because I got the right answer. But clearly I did it wrong because it took 87 minutes.
Me reading today’s instructions though…
I don't know how you're managing to make it take so long, are you recalculating the same nodes over and over again or something?
@Sterno Once you get it it's not that complex, but it's certainly a lot to read
@Ronan theoretically no, but obviously in practice probably yes
13:40
@Sterno just give me a proper ICD
14:09
@Sterno very same
 
1 hour later…
15:13
@Ronan nah it was the call to min(distance). Was way too slow when I preloaded it.
And also didn't sort it
Really? There aren't that many nodes
@Ronan 250k is slow enough to add up apparently
Since it's running on every iteration
Weird
When I stopped preloading it ran in a few seconds
@Ronan I haven't bothered yet since I'm too lazy to read them giant wall of text
15:48
Ugh, I still need to do yesterday's
16:13
Today's was fun. After writing all that code to parse packets, though, I'm sad there wasn't something more interesting to do with it
16:53
Epic started a daily free game for the holidays. Today is Shenmue 3, but tomorrow will be something different, so get it while you can.
17:17
@Batophobia 15 days of free games in total
Though keep in mind that the store appears to be a bit buggy right now. I wasn't able to redeem the game at all
Kept getting an error when I tried
17:54
Found a problem in yesterday's code that cut my time in half (at least, based on the test case... I haven't wanted to run it for the full thing), but that would still be 40 min :(
 
2 hours later…
19:59
@Sterno So you don't think this will be a reusable bit? I miss those.
 
2 hours later…
21:38
finally ferreted out the bugs and got the answer. feels pretty satisfying
22:37
lol
There is a broken crossbow on the tile where swampletics died and wiped

« first day (4172 days earlier)      last day (1163 days later) »