Re: [Dmrpg-devel] First post

[Ely L. <el...@cs...>]

On Tue, 28 Jan 2003, Jouni Miikki wrote:

> Lemme introduce myself.
>
> I agree on next things that we should make it most abstract, well-designed
> data-stuctures so it can be stored in different formats on different
> places (files, db3, sql etc). If we do this, we can possibly get different
> kind of frontends for this whole thingie(tm) we are starting to make. Lets

The question is if to sperate the modules/code from the database,
if you know eggdrop and lambdamoo that would be the diffrance I mean.

> say someone wants to make java-applet for pages to view certain kind of
> stuff from space that we have allocated in our data-structures, only
> showing where some things are moving and what is happening. For that it
> only needslocation information and what kind of environment it is in. Also
> maybe news from that area and what has been happening lately. This way
> someone could follow up what users have been doing on that area and also
> keep up its information from work and other places without having highend
> 3d. That was just a simple example :).

yea the question is how hard is it to achive that?:)

>
> there isn't much on framework yet and that is good, so now we can talk
> about making certain kind of direction and targets what we are going to
> achieve. This might take few weeks or even months, but that doesn't mean

I think we should make it playble as fast as we can,
even if we are not sure on design that would make more people intrested.

> that we shouldn't code anything or do some basic work for it. Maybe
> defining what kind of libraries (if not own made) we are going to use and
> what platforms we are mostly trying to target our project. Lately I've
SDL/crystal space?
using those and you solved most problems in moving your code between
platfroms, as much as I dislike it I guess we should support windows as
well..

> been doing alot ogl stuff so it works on linux and windows and one ugly
> thing I found was that random isn't same thing on some c/c++ environments
> and that made me puke. well I made my own one. So thats already something
Yea I know that one..

> we should know about when starting to do this. (try to keep up with me,
> usually I'm writing with mind-flow, so it might feel little bit confusing
> to read my texts.)
>
> Any kind of documentation about things is good. Well I think everyone
> knows that :).

should we take GNU's coding guide?

> So the network idea what I had to introduce here is simple. There is
> interface for common server / client combination. Everything includes its
> own server and client .. lets see this as an component which is included
> in every frontend and programs we are going to do. Through that
> information is transfered to different clients who needs it and there will
> be interface for modules to register into this components offered network.

Well one of the biggest challenges is making a dynamic enviourment
so diffrent server could have diffrent sort of graphic or music and the
client won't need to spend hours on downloading it everytime it connects a
new server, that might help.

> So you can ask for data if that function is on that network. Its like
> corba/java rmi etc what there are just losing the basic server from it,
> which keeps up all the information. With this we can offer
> java/c/c++/perl+others if someone just codes that interface into that
> particular programming language, with this we can support the same
> features all over network. So routing is made in there components
> as many other things aswell. (I try to document this soon so we can talk
> more about this). Ofcourse if someone has better ideas to make this kind
> of system, I'm certainly open for ideas.

Can you give an example in which it solve a problem better than classic
approch would?

> If someone knows more coders to attend into this kind of project we should
> get those. Even if they are not professionals and they know what they are
> doing with coding, they could do routines, modules etc.

Btw when I talked about module API I didn't mean loadble module API,
I ment more if you want to have a database module or certain feature added
you need to have API inside the code, so for example the calling for
database would stay the same but the module would implement the functions
needed.

> About module interface making this whole thing as in scalable and generic,
> so experts on c/c++/java/+other have to think about multiplatform and
> issues what there can be on loading modules. About this I have only

I don't think we should use more than one langauge in the meanwhile:)
and I think c is good enough:)

> dynamic libraries as those can be loaded and also configured to load
> certain kind of modules as needed. So adding those in runtime to
> mainprogram isn't problem. Just as windows / unix systems has their own
> kind 9*of systems for c/c++, so that has to be notified also.
>
> Ok, I'm back to work. Cya all, mov.
>

Ely