Half-Life 2 / UT2003 engines comparison.

[legacy-nickelplate]

What are going to be the main differences between the two engines? What are their respective strengths and weaknesses? What kinds of mods are best suited for each engine?

I know it might be too early to talk about the HL2 engine, but since some of the most successful HL mod teams already have access to the SDK, I thought some of you may have valuable information on it.

[legacy-iboo]

Re: Half-Life 2 / UT2003 engines comparison.

Originally posted by nickelplate
What are going to be the main differences between the two engines? What are their respective strengths and weaknesses? What kinds of mods are best suited for each engine?

I know it might be too early to talk about the HL2 engine, but since some of the most successful HL mod teams already have access to the SDK, I thought some of you may have valuable information on it.

Maybe you should be asking the HL people this question.

[legacy-TseTse]

actually, i tihnk it's a perfectly honest question and id rather see it on this board than a flamer thread on General Discussion. lol

ive studied the "e3" video rather carefully and kept an eye on HL2 developments. generally, i dont see ANYTHING really, that cant be technically done in unreal (perhaps the deformable terrain, but i thought i heard you can do that in unreal now, somehow).

however...

they do seem to have a class of "dynamic" mesh which are neither static mesh nor merely animated mesh. seems like a whole class of objects with rather robust interactive properties available. aspects seem akin to karmalized features, but this seems to be deeper than just additional physics properties. it's like a whole entire class of objects integrated from the core.

likewise, seems that the material system is really geared towards similar interactivity. that's something i think which can be done within unreal, but... again, i get inkling that it's designed to be really easy out-of-box.

im not really knowledgeable about unreal scripting, so im not sure what's even possible from the unreal side. im just going from what ive seen from games thus far and from my general understanding.

overall, i suspect both will be "equally" powerful for modding.

depending on the mod's vision, i think the initial questions will emerge around: vehicles; netcode % num of players; size/scale of terrain/maps; physics systems.

those are the core issues as i see it.

[legacy-AkumuX]

I'm not really the best suited to answer this question either, but I did want to make the comment that I'm pretty sure it will be much easier to make single player mods in HL2 than it is in UT2k3. UT being a primary multiplayer DM experience, you have to get pretty familiar with unreal script to make a SP experience. Or at least from all that I've seen you have to.

This is rather frustrating to me, because I'd like to make single player map campaigns, so that I could tell stories and also give examples of what I'm capable of, without having a lot of programming knowledge. I understand that UScript is a scripting language and not a programming language, which saves a bunch of time, but its still code, and that just makes my head ache when I write bugs. Since Valve has been designing their engine around the idea of making it way more accessible to moders, I imagine it will be very user friendly and easy to work with. When I look at the tools for Unreal it looks like a lot of pretty road concepts that are yet unfinished (IE: tools in the editor that don't work the way they should and are left in, which only creates confusion).

[elmuerte]

the only thing missing in the UT2003 build of the engine to create decent SP games is the SaveGame feature.
At this moment you can only save at the beginning of a map, not in the middle of it. But iirc the SaveGame feature should return in UT2004.

As for UnrealEd not being finished, it's because it's in ongoing development. Ofcourse they could create "finished" builds of the engine (e.g. all non working feature removed), but what's the use of that.
I don't think you can say anything about Valve making their engine "way more accessible to moders". You can do about anything you want in the Unreal engine because most of the game code is open, you just need the skills to create the content you want. You can't expect Epic to include building blocks for most ideas in the engine when they don't use it in the rest of their game.

[legacy-Doc_EDo]

Besides SP savegame features I want AI to be better and easier to set up for SP.
Right now it's very hard to implement FEAR and some other functions.

[legacy-AkumuX]

Originally posted by El_Muerte_[TDS]

I don't think you can say anything about Valve making their engine "way more accessible to moders". You can do about anything you want in the Unreal engine because most of the game code is open, you just need the skills to create the content you want. You can't expect Epic to include building blocks for most ideas in the engine when they don't use it in the rest of their game.

The only say that because the articles I've read about HL2 indicate they have built their new engine from the ground up to make the moding process easier to new people. HL1 owes most of its success to all the great mods that were created based off their game. Its kept HL1 sales strong many years after most products go to the $5 bin. Valve is VERY smart to give that back to the moding community by developing their product to be easier to work with. Granted, none of us have gotten to get our fingers into their new stuff yet, so for all we know it could blow compared to the Unreal engine. But so far they have a lot of my respect by A) Making a great SP game. B) Not buying into this marketing **** of building separate SP and MP boxed products (the way ID and Epic have done).

Anyway, I'm not the most experienced in this field, so I'm don't want anyone to think what I've said is all 100% accurate. I'm just going off my own experiences. I agree with you that it would be rather pointless to remove features from the Unreal Editor that don't work, but you also have to understand how having those features in there, create a lot of frustration and confusion to new moders like myself. Of course the new HL engine may be just as bad.

*EDIT*

I would like to give mad props to Epic for investing in those online training courses they had 3D Buzz do with the masteringunreal.com site. That is TOTALY awsome to give that to the moding community. That alone gives me a lot of respect for the Unreal engine.

[elmuerte]

Originally posted by AkumuX
The only say that because the articles I've read about HL2 indicate they have built their new engine from the ground up to make the moding process easier to new people. HL1 owes most of its success to all the great mods that were created based off their game.

Well HalfLife was everything except mod friendly to begin with, the unreal engine however was pretty much designed to be mod friendly through UnrealScript.
The only thing Valve could to better compared to what Epic did is to provide documentation, Valve must be large enough to have some dedicated documenters around.

[legacy-Doc_EDo]

Could do better?
Epic doesnt even have any documentation becides a few comments that often contain "//FIXME".

[elmuerte]

Epic does provide information, it's just limited to the major part of the UnrealScript language.
The only thing missing is the game flow documentation, also a lot of game code documentation is missing.
But since you have most of the source it doesn't matter much.

[legacy-JaFO(JBE)]

There's one big mistake in trying to compare HL2 and UT'03 ...
HL2 is a singleplayer-engine that can do multiplayer as well.
UT'03 is pure multiplayer only ...

Because of the focus on single-player the need for interactive objects is greater as well.
In this aspect it is a bit like comparing a stunt in a movie to one that's done in a amusement park. The movie-version doesn't need to be repeated so it can be a one-shot setup.
Otoh the amusementpark needs something that can be done a million times without causing any problems.
In a SP-environment it makes sense to have crates that can break, but in a MP-environment you'd have to take care that the broken crate doesn't cause any problems in the level for the other players.

And I'm sure you'll find better SP-oriented AI in Unreal 2 and other SP-versions of the Unreal-engine too ... (IIRC it uses plain text-files for the dialogs ... )

As for building an engine to be easily modified ... I say Do'h!
Of course you'll want an engine that has that. Not just for publicity, but for their own licensing needs as well (the less training any potential customer would need the more customers you can serve ... ).

Valve/Sierra may not have bought into the separate SP & MP-games 'marketing-****', but they did manage to buy into the 'rerelease half-life to squeeze every penny out of the popularity of one of the add-ons' -marketing-**** ...

[legacy-Doc_EDo]

Sure but UT2003 lacks the SaveGame feature and for modders that's a HUGE drawback. For Epic that would be very easy and fast to implement if they wanted. I mean you could cope with almost all other stupid things that it has but lack of SaveGame is just a drop too much.

[legacy-JaFO(JBE)]

Why would a mod for a multiplayer-game need a save-feature ?
And why isn't the ini-file-system not good enough for saving ?

[legacy-Doc_EDo]

[elmuerte]

Because of all things that changed the savegame function got broken, and since there was absolutely no need for it in UT2003 they didn't set a priority to fix it.
And like JaFo said, UT2003 has not been designed for story based SP games.