As the resident c# evangelist, I hereby take custody of this thread. icq me 520608. or email me if you must, mgambrell@gmail.com

It's 1997 2000 2003 all over again!

Waitaminute...

Edit: regardless of snarkiness, we need the goddamn editor to be completed.

I figured that there would be resistance. ;-)

Adding the extra functionality will actually make it EASIER to use verge not less. This is for two reasons:
1. C# isn't hard to program in at all and is a standard. VergeC was a scripty javascript like version of C that was somewhat non-standard and quite slow. VergeC has no Intellisense editor, no debugger, and limited options to turn to for support. In addition, porting to .NET would let a user use any .NET capable language they want. So if a user doesn't like C# then maybe they'd like VB.NET, Python, or JavaScript.
2. Porting to C# allows for easy inheritance and addition to verge classes. This allows much easier customization of verge than the current C++ code which must be recompiled. With .NET it would be possible to inherit from a class that is within a DLL and make your own customizations. It would be possible for someone like McGrue to make a Sully DLL that extends the core VERGE and allows for more things like menus and battle sequences. The sky is really the limit.

About the .NET runtime. Two more points:
1. The Mono core download is under 4MB. So long as one does not use much of the .NET this is all you need. Verge would be mostly doing calls to external DLL's (directx, SDL, fmod, etc) and so won't need most of the .NET libraries.
2. Most people who have XP installed will probably already have .NET 1.1 installed. If we targetted that we'd most likely be OK. .Net2.0 might not be standard for a while but Vista will introduce .NET3.0 so things are already progressing past 2.0

I'm not necessarily advocating abandoning the verge code base. For one, have you seen C# and C++ code side by side? There is quite a bit of similarity. I'd bet that a large amount of the code could be directly used without changes or with minor changes. Sure, the code would have to be more object oriented but that isn't necessarily a bad thing. As an aside I say it would have to be more object oriented because in C# everything must be contained in a class. There are no global scope functions. There are no header files. The philosophy is different but the language is very similar.

So, anyway, I think that modernizing the engine could be a good thing. Maybe the best idea is to create a verge fork to .NET and maintain the code bases seperately. Then developers and users can pick for themselves and if one method turns out to be much better than the other then maybe the loser can be discontinued at that point. Or maybe they can coexist forever.

The way I see it, expanding functionality in all directions could potentially open up the way for some serious issues. For one, people will start throwing DLLs around, plug in additional functionality that requires a particular runtime (say, a specific movie codec or code library) and before you know it, you're hunting around the web for seven different downloads just to have all the files to play a tech demo.

However, most fundamentally of all is my second point: mad as it may seem, I believe in placing limits on the resources people have at their disposal in order to realise their ideas. That way they are forced to make the best of what they have and, consequently, make a better game because of it.

Consider this scenario: I start with 2d VERGE but soon work out how to incorporate some 3D graphics routines into it. I have something akin to the Quake 2 engine. So I ditch all the 2d art I spent the last month making and start again. Half-way through making some fancy player model and an early level I work out how to incorporate bump mapping and volumetric lighting into the engine, so now I have to re-do the model to take advantage of the bump mapping and re-light my level with the new lights. A year later and a few more technical advances down the line and I *still* have just a player model and a bit of level. I mean they look nice, but it's bugger all in the grand scheme of things. Christ, I think, I've turned into 3D Realms and this is my Duke Nukem Forever. Disillusioned, I quit.

Now, if all I ever had was a 2D game engine and some strict limitations I'd have been forced to work within I might have made a game. Instead, with the near infinite number of technical possibilities to realise my idea, I'm overwhelmed and ultimately screw things up royally.

See where I'm going with this? As I see it, in games development "everything" is just an infinite magnitude of "nothing".

Hey,

I know that I've been basically absent, but I'd like to throw in a few things myself. Let me just start by saying I don't know much about .NET/C#, so for all I know there's some magic techology that solves all these problems.

Having a codebase in .NET would be great. It would be a zillion times easier to add things to verge for us developers. Once it's done. Getting it there, however, will be a lot of work. I don't really know much about how close C# and C++ are, but seeing as how most of the benefit of such a transition for us would be to have everything C#-like, I'm assuming a significant porting time. Especially because you're proposing using the .NET toolchain for verge games, so we'd have to rip out the current verge compiler and runtime while keeping all the library functions, etc. That sounds hard to me, but maybe you've got a good idea how to attack it.

Ok, so, if we get there, it's easier for vergedev. But for people using verge? Let's look at the current process for making a simple ("hello world") verge game: Download verge, open system.vc in a text editor, type, save, open verge. After a transition to .NET in which we're using visual studio, that turns into a dozen-step process involving installing the .NET framework, the MS dev tools, opening/making solution files, and I don't even know what else to get everything to link together into a game and run. I don't know what the mono toolchain is like, but I'm guessing it's harder to install/use than Visual Studio, and probably dissimilar enough that platform issues will be a big problem. Is there something here I'm missing? Can you make the process as simple as it is now?

In that same vein, C# being "not hard" to program in is fine, but you have to realize that VergeC doesn't require you to think about references/classes/members/etc. It's a very simple language, which is intentional. Yes, it gets in the way when you want to do something complex (to put it midly) but for people just starting out that just want to make their first game, you want something that easy. Using something like IronPython is going to introduce more complexity into the installation and support. But hey, maybe in a few years, everyone will learn to program in grade 1.

That all being said, far be it from me to discourage working on something verge-related.

Quote:Originally posted by mcgrue

I tend to agree more with int's reasoning. Can't we just find a way to speed up the function calls and the other whotzis that vecna said are the main vc performance bottlenecks?

I'll buy tacos for whomever does it!

As evidenced before, I've advocated the same thing. Now I'm bringing up another choice. Whether it's good or not remains to be seen. I'm really open to *either* choice. From the sound of it, it would appear that my first option was the more accepted one.

It should be noted that, technically, the revolution has already happened. Zero has already got a lot of work done on an engine called CAROT that is kinda-sorta like a .NET version of VERGE. If we did do a verge .NET port then chances are there'd be a bit of his work in there or at least that would be the starting point. However, since he's already gone off and done it on his own there is still the option to leave it called CAROT and not use such ideas in Verge proper. He, and potentially others, could still work on carot as a completely seperate entity and everyone else can stick with normal verge if that's the way everyone wants to go.

I do not advocate more complexity for verge.

I do advocate completed dev tools so the linux and mac people can finally make with it, and the windows people can make competently.

If you, or others, wish to make a vc interpreter in .NET, that's way awesome and cool and stuff, and I'll set it up a SVN for y'all, but that is such a significant deviation from the main project as to be it's own project.

Verge is all open source and such now, so there's no need to really split hairs about it. Remember, though, any such vc interop layer would need to find a version of vc to remain compatible with, and update from there, rather than branching to it's own features. Which seems frustrating and pointless ultimately, since if you're already favoring .NET, you've got far better options to program in than vc.

I don't think moderately tweaking Verge will really get us anywhere new. Verge is perfect for that niche of people interested in making games, but not savy enough to code it all themselves. I really wonder if that niche even exists anymore. If you ask me, Verge3 is a good enough tool as it to make pretty much any kind of game you want (except for MapEd maybe), so what's stopping them.

I'm curious though, did anyone ever try (with a large effort) to spread the word about verge to attract people other than random web hits? Anyone recently? Maybe that's all we need.

Er... don't assume that everyone that uses Verge "wants to make a game but isn't savvy enough to code it all themselves". Some people just don't have the time or inclination to bother coding it all for themselves, or have a nostalgic attachment to the community that makes them want to build stuff with it.

... Not that I speak about anyone in particular, of course.

*cough*

... Your post count is at 1049 for me?

Also, calling what I do "godlike shenanigans" makes me really want to work on a game now. But I have some work-work to do first.

So I'm going to do that, and then shenanigans. Brilliant.