12 replies to this topic

[eddie]

Man, there's no better feeling than hitting a basic milestone, is there?

I just had this rush as I finished something *so* basic, but it's awesome.

I basically got my 'game' infrastructure up and running.

In a nutshell:

• A server that accepts clients
  
• A client that sends out data in the form of a proprietary "protocol" (over TCP)
  
• A mechanism for both the server and client to handle the data flexibly (lua extensions)

Which alone is pretty neat. Especially to have it where it's largely unit tested, doesn't memory leak, and is easily extended.

Now I've got it sending ChatProtocols (spamming all clients). It's probably the most over-engineered chat game there is, at this point. :)

Anyways. I'm rather happy. Not because it's "fantastic", but the framework is there, it's all code I understand, it's all OO and well documented, I can now poke at it with a stick and make it do back flips.

Anyone else got anything cool going on? Otherwise this post just sounds like I'm bragging over nothing. :)

[moe]

Sounds pretty cool to me :)

I wouldn’t say that server-client applications are easy so it might be a tad more than a basic milestone.

Quote

it's all code I understand

I’d say that’s the most important thing anyway. Sometimes when I build something it works already yet I am not satisfied until I also have a proper source. Hence I recode until I am happy with it before moving on :)
That’s one of the big advantages as a hobby coder. You can move on or recreate as much as you like :)

However milestones are good and it’s always a nice feeling if you reach one and everything seams to work as planed.

As for my own projects I am unfortunately a bit stuck right now. All the ideas I have require a pixel shader 3.0 capable gfx-card and I just can’t afford to buy one right now… developing in ref mode has proven to suck quite a bit... there are shaders I made working in hal and ref as ps_2_0 but if I change the shader version to ps_3_0 they don’t compile anymore :( And it’s far from convenient if you have to wait a few seconds until you see the output. You got to love realtime...

[eddie]

moe said:

Sounds pretty cool to me :)

Why thank you. :) I'm pretty excited... There's nothing like the feeling of potential.

moe said:

I wouldn’t say that server-client applications are easy so it might be a tad more than a basic milestone.

Yeah, I totally agree. It's just funny because to an outside user a basic chat program (even if it can hold a lot of clients) probably doesn't seem monumental. But all the stuff under the hood is what's so cool about it.

moe said:

I’d say that’s the most important thing anyway. Sometimes when I build something it works already yet I am not satisfied until I also have a proper source. Hence I recode until I am happy with it before moving on :)
That’s one of the big advantages as a hobby coder. You can move on or recreate as much as you like :)

Agreed! You get everything you want, and your own milestones are your own. Bend them, break them, add to them, remove from them, you're your own boss.

The best feeling I get from that, is that every part of my system can have *potential*. So many times when I'm coding on-the-job, I have to cut corners, or dive to the solution, and sure, it works, but it's not extensible, reusable, or even nice to look at. Having a system where I look at it and say, "Wow. It's not perfect, but I can use this for SO many things!" is a great thing.

moe said:

As for my own projects I am unfortunately a bit stuck right now. All the ideas I have require a pixel shader 3.0 capable gfx-card and I just can’t afford to buy one right now… developing in ref mode has proven to suck quite a bit... there are shaders I made working in hal and ref as ps_2_0 but if I change the shader version to ps_3_0 they don’t compile anymore :( And it’s far from convenient if you have to wait a few seconds until you see the output. You got to love realtime...

Woah man. :) You're way ahead of me, so look at being stuck as a blessing then. :D

I was originally going to code this with my friend who's a rendering guru, but he has no time to commit so I'm on my own coding graphics.. I expect to be using the graphics forum heavily! :D

That said, good luck mang... You could always hawk a card off eBay perhaps? If you're in Canada, check out www.redflagdeals.com, or www.ncix.com - good place to pick up cheap hardware. :)

[moe]

Quote

You're way ahead of me, so look at being stuck as a blessing then.

Hehe, this might not be true. It always depends on what you do. I have no clue in networking :(

I am all into gfx programming and physics is already something I take a lib rather than doing it myself. I tried at first but I have a rather poor math skill... Unfortunately what I am trying to do now requires a solid base in math. That’s not very advantageous :)

The main reason why I would need pixel shader 3.0 is because I build a little ray tracer witch is supposed to run on the gfx-card. So it would be quite shader involved and also need a bunch of render targets and some other things... :)

[monjardin]

eddie said:

So many times when I'm coding on-the-job, I have to cut corners, or dive to the solution, and sure, it works, but it's not extensible, reusable, or even nice to look at. Having a system where I look at it and say, "Wow. It's not perfect, but I can use this for SO many things!" is a great thing.

I agree whole heartedly. That's why I like to tinker around on my own. Then you can quickly incorporate the pretty stuff at work with those tight schedules.
On a related note: does all that task and defect tracking (I think your SCM calls them tickets) really help on a one man project? I'm refering to this: http://www.kickingdr...s/trac/timeline

[eddie]

monjardin said:

On a related note: does all that task and defect tracking (I think your SCM calls them tickets) really help on a one man project? I'm refering to this: http://www.kickingdr...s/trac/timeline

Well, there's a bit of history.

First off, I'm an SCM nut. I started out life as an SCM consultant, so I'm always looking at new toys.

Also, I was supposed to start this project with a friend who's a rendering God (SIGGRAPH published, knows his stuff, and he's fantastic in bed (J/k)), however he wasn't able to commit as much time as he'd like so he dropped off.

So I started up the SCM in the time we were "working" together, mostly because I'd heard so many good things about Trac, and I wanted a way to record my progress, and jot down what I thought were important milestones.

All history aside, I do find it rather nice to have around. It's all really low maintenance, and when I'm spending time pondering I often flip through Trac and post up new bugs, or add notes to old ones, etc. It's also rather energizing to see real progress being made in terms of 'deliverables' (even though they're self-fabricated), and knowing you're accomplishing stuff.

And one day, if I get the game up and running, the idea was that it could be the public interface for people to see, file bugs against, etc.

I'll probably have to flush all my implementation specific notes, but that shouldn't be an issue.

Where'd you find the link to that anyways? I thought I didn't advertise that.. :)

[TheNut]

Good stuff! What plans for software do you have with the new framework (aside from chats =)? One thing I find most fun when it comes to networks is being able to monitor what other people are doing, especially with wireless networks >=). I’d like to eventually support passive SSL sniffing and packet inflation for my Packet Spy utility. I spy with my little eye… heh.

[eddie]

<beams> I like people asking me questions about my stuff. :)

To be honest, for now the main goal is to build a game out of it. It's extendable and easy to build any kind of client/server on top of it, but I might just leave the other ideas for a later date.

I've always fancied ideas like a nice lightweight HTTP Server, or a source control system of some sort, but who knows... I'm liking this gaming thing so much, maybe after I make my millions on this video game, eh? :) J/k

Sounds interesting on your SSL stuff. :) I know so little about the deep innerworkings of networking, it's scary. But programming's a humbling experience!

[monjardin]

You linked to it from your blog post about DevMaster. That's where I found it. Sorry, if you didn't want it posted here!

[eddie]

Oh, don't worry. I don't really have "trade secrets" or anything. :) If it generates some interest, why not?

Although it's pretty far off. :) Having screen-shots of a DOS window won't exactly count as good advertising. :)

[geon]

My little "3D editor" is drawing things and I can rotate the camera around them. All objects in the scene can be instanciated any number of times, with different transformations, colors, etc.

Next: Load and save to disk...

[eddie]

Cool geon. :)

So are you trying to write something like a world editor? Or more like a Blender/Maya/XSI software package?

[geon]

It's supposed to become a 3D modelling and texturing tool. Nothing more, nothing less. Well, maby object iarkys should be supported as well, but I consider that part of the modelling.

The focus will be bezier patch editing, since I feel there is a severe lack of good tools for that. No tool I used was efficient to work with.

Then, I hope to support meshsmooth also, since it is basically the same thing. Just let booth derive from a lowpoly class, and you are set. (Sounds simple, doesn't it.)

Also some primitives like spheres and teapots would be trivial. (I'm playing with torii for the moment...)