Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

Shmup

Name: Anonymous 2009-08-17 19:46

Hey /prog/,

University's gonna start in a week for me, and I know that for the semester I'm in, our final project is to code whatever the hell we want in Sepples following certain guidelines that will be given later. I decided I'd go with a shmup game and that I will most likely start it very early in the semester.

The reason I'm posting this here is that I want you all to tell me why you think this will inevitably fail, so that I can avoid it by preparing for it. If this helps you, I have 3 years of experience with Sepples and I've already programmed 2D games before (although they were all ARPG-like) using SDL.

Name: Anonymous 2009-08-17 19:53

NO EXCEPTIONS

Name: Anonymous 2009-08-17 19:55

You will fail because the guidelines won't allow copyright infringement, and therefore you will not be able to include Touhous in you're work, ergo it will stay unpopular.

Name: Anonymous 2009-08-17 20:45

>>3
Since it's nothing that will be sold, "copyright infringement" is usually accepted for projects here. Then again, there's no real infringement with Toehoes since ZUN doesn't give a shit. I'd still like to make my own sprites though and o be honest, so far I think that's going to be my biggest problem. I'm absolutely ass at drawing/art/making sprites. I don't know how I'll manage to do this part (yet).

Name: Anonymous 2009-08-17 21:10

Shmups are incredibly easy to make. All you need to do is load some sprites, make them move, and let them shoot other sprites from them.

If you're just going for basic move and shoot, you'll succeed. If you want something fun, you'll fail because you need good AI and enemy patterns and whatnot.

Name: Anonymous 2009-08-17 21:11

programming games
( ´_ゝ`) i remember when /prog/ was all about programming fibs, not this supreme faggotry.
i miss those days

Name: Anonymous 2009-08-17 21:16

>>6
Don't worry, we're all programming a fibs a day to keep sane.

Name: Anonymous 2009-08-17 21:38

>>4
Steal all your sprites from Soldat.

Name: Anonymous 2009-08-17 21:40

What about a fib danmaku pattern?

Name: Anonymous 2009-08-17 22:06

>>1
How much are you paying in tuition?  You can do better than a SHMUP.

Name: Anonymous 2009-08-17 22:07

>>6
I can include a FIBONACCI BUTT SORT in my shmup if it'll make you happy.

Name: Anonymous 2009-08-17 22:10

>>10
Could I do better? Of course. Would it be as interesting and as rewarding as a Shmup? I don't think so, tim. It's not like it's the only assignment in the semester, I'll do tons of... whatever the hell they want us to program, but the last project is for ourselves.

Name: Anonymous 2009-08-17 23:17

>>1
University's gonna start in a week
Where are you going?

Name: Anonymous 2009-08-18 13:00

>>11
Sort my anuses!

Name: Anonymous 2009-08-18 13:14

>>14
Recursively!

Name: Anonymous 2009-08-18 13:31

Create a multi million dollar programme. I'll contact you via email with further details once you've done this step.

Name: Anonymous 2009-08-18 13:47

You will not fail. Write it in a decent language first; when it works, translate it into C. Then throw some shitty template in for something and--surprise!--it is C++.

Basically: easiest way to avoid failure is to avoid C++.

Name: Anonymous 2009-08-18 13:48

embed lua / scheme

Name: Anonymous 2009-08-18 14:12

>>17
The teacher might get suspicious when he sees all the base_GHCziBase_*_closures.

Name: Anonymous 2009-08-18 19:00

ive been wondering how u write an efficient collision defection algorithm. obviously it has to be fast with more than 1000 bullets on screen.

how does touhou do it? also are the hitboxes of the bullets squares? what about strange bullets. like lasers that are thinner at the ends and fat in the middle.

Name: Anonymous 2009-08-18 19:07

>>20
CPUs are good are doing this sort of stuff

Name: Anonymous 2009-08-18 19:19

>>20
Don't know about the latest ones, but the ones I've looked at has had three kind of hitboxes: Squares, circles and beams.
Beams have uniform thickness.
I doubt it's written with efficiency in mind, there won't really be that many things that can collide at any one time. Maybe some hundred bullets against twenty enemies, at the most. And on a slow computer, even that much causes lag, which wouldn't be the case had it been OMG OPTIMIZED!11!.

Name: Anonymous 2009-08-18 20:57

>>22
but still, lets say there are 1000 bullets on screen. doesnt the players hitbox need to be checked against every single one of those bullets every frame to see if it "overlaps" with one?

Name: Anonymous 2009-08-18 21:15

>>20
[CODE=python]
for bullet in bullets:
   if bullet.rect.colliderect(player.rect):
      player.kill()
[/CODE]

Name: Anonymous 2009-08-18 21:19

Name: Anonymous 2009-08-18 21:20

>>23
in the worst case it does
(moving every bullet to a random() location every frame or something)

you can schmuck around with random datastructures
and hope to trade many cheap operations for less full collision checks for a speed up in average as in
http://www.metanetsoftware.com/technique/tutorialB.html

Name: Anonymous 2009-08-19 3:11

To be successful, you must include enterprise-quality industry-standard XML-based BulletML1 support. Even Touhou(R) Danmaku 12g supports BulletML now.


---
1http://www.asahi-net.or.jp/~cs8k-cyu/bulletml/index_e.html

Name: Anonymous 2009-08-19 17:04

>>23
With a naive implementation, yes. And this would indeed pose a problem if you were running on a 66kHz processor.

Name: Anonymous 2009-08-19 21:59

>>28
Are you suggesting there is an optimization by which you wouldn't have to iterate across each bullet once per frame?

Name: Anonymous 2009-08-20 3:04

>>29
To check for collisions? Several come to mind off-hand.

Name: Anonymous 2009-08-20 4:30

>>30
Then your retarded.

Name: Anonymous 2011-01-31 20:18

<-- check em dubz

Name: Anonymous 2011-02-03 4:39

Name: Anonymous 2011-02-03 5:21

Name: Anonymous 2011-02-04 16:43

Name: 2012-01-25 7:18


Don't change these.
Name: Email:
Entire Thread Thread List