29 replies to this topic

[NomadRock]

knackered3 said:

So, you should have a section called "Show me how to...." followed by a series of links to these articles...if no article has been written, you simply leave an email link for so someone can submit one. I may drop one in there if it were that simple.

That is actually a very good idea. We should also probably have someone serve as editor for that section, as it should be kept to the bare minimum to get the desired effect. In normal articles no one makes any claim to how difficult they are, or how bloated with extra information they may be. In a section like that it would be useful to have someone checking to make sure theory stayed out of them.

It would probably also be useful to catagorize them by api under the topic. For example if you had a topic of "stencil shadows" you would want one for opengl and another for DirectX. "how to make a hinge joint" would also need seperate sections for seperate physics systems.

I would be willing to write up a few, especially since they would be small enough that they could be written relatively quickly. This might solve the cyclic problem since they would most definately include code that could be dropped in without fully understanding the theory. Then as the user grew to know more they could work on tweaking the code to their special needs, or for the likes of quaternions, never touch it again because understanding the theory doesnt really help you much.

In short, I like it.

[JSoftware]

"show me how to...":

Sounds like a brilliant idea! that's surely been my great problem in this short time i've baffled with 3d graphics not to be able to read and understand all this math, and afterwards transfer it to code!

i would love to contribute with, maybe some pascal/delphi ports of eventual code..

[SpreeTree]

Wouldnt this be along the lines of the code snippits from other game sites?
Still a good idea though :)

Spree

[NomadRock]

It could be, but I had always thought of code snippits as being interesting pieces of code that is small enough to grok easily.

This stuff would entail a few code files usually, perhaps a small library. At least that's what I had in mind.

[baldurk]

hmm. I'm not convinced that removing all theory and providing complete sample code is a good idea. Then you end up with nehe where people simple copy-paste the code, and when it doesn't work come to complain here. What we want is simple to understand explanation of how the technique works, and some pseudocode or similar to demonstrate in practicality. I don't mean looking at the deep mathematics behind proving the technique works, but at least trying to make sure that the reader understands the technique.

[NomadRock]

You have a good point, but NeHe is wildly popular with good reason. It is easy to start on. Sometimes there are several dependent effects I want to play with but I kind of need the others while I am playing with one. It is helpful to have some useful placeholders in these situations. I am not saying theory should be thrown into the wind, but there is more than one way to learn.

[Polar Sleuth]

Is the DevMater site dead? No, people are still here and are posting in the forums. But by some of the indicators, it is no longer really active (e.g., Image of the Day, number of daily posts, no new tutorials, etc).

As to flipcode, they returned with a good series - but additional articles are ... lacking. Even flipcode is struggling to maintain a stream of the daily gristle outside the forums.

If you measure the vitality of a site by the contributions of its members to the articles / tutorials and images, then no game development site is alive. The rate of contribution per member is too low. Most people have enough time to write a few lines in a forum post, but not the time to write an actual article or save a screen shot.

[baldurk]

NomadRock said:

You have a good point, but NeHe is wildly popular with good reason. It is easy to start on. Sometimes there are several dependent effects I want to play with but I kind of need the others while I am playing with one. It is helpful to have some useful placeholders in these situations. I am not saying theory should be thrown into the wind, but there is more than one way to learn.

Heh. I help out on an opengl IRC channel (#opengl on irc.enterthegame.com </plug>). Unfortunately I have to deal with the downside of NeHe which is that it gives people too easy access to working code. I think a good mix between code and theory is best - so that people have to complete the code themselves, by understanding the theory.

[NomadRock]

Perhaps, at that point it comes down to the age old question of letting people pick what is best for themselves or picking for them. I would like to pick for myself, but it is true that there are a great deal of people who are poor at picking.

[SpreeTree]

Polar Sleuth said:

But by some of the indicators, it is no longer really active (e.g., Image of the Day, number of daily posts, no new tutorials, etc).

I agree with that point, but as devmasters is a pretty small community, theres nothing you can really do especially in regards to the image of the day. The articles have seen little increase, but the majority of the articles are top notch, some of the info you cant find on other sites, and I thnk thats how it should be... For now anyway ;)

Spree