[After a brisk nap]

With the release date for Dreamfall pushed back, does that give Ragnar Tørnquist an opportunity to observe what works and doesn't work in Fahrenheit (Indigo Prophecy) and maybe lift some of the better ones? Or are the games likely to be too fundamentally different for that to work?

I would argue that when a game is delayed, the developers not only have to update the graphics to the new state-of-the-art, they should also implement gameplay and storytelling innovations that have been developed in the mean time.

[Fop]

Changing the design late into development is a stupid, stupid idea.

[After a brisk nap]

Not necessarily. Good designers stay flexible, test their product on users (players) at all stages of development, and adjust the design accordingly. Dreamfall already dropped their scripted camera late in development, and that's a pretty major change. I don't see why it wouldn't be possible to overhaul the interface, change some of the interactions, or even tweak some of the puzzles.

Of course, it all depends on how tight their deadlines are, the stability of their production process, and the state the game is currently in.

[Aj_]

That change in the camera system was followed by a delay.

1. A fundamental change in the graphics engine would result in a big delay, you'd most likely have to reprogram, reproduce, and redesign models and textures through out the game.
2. A overhaul of the interface would affect the gameplay, if you had to change how the interactions worked you'd ahve to test the whole game again.

Testing should take place on alphas when the principles are set in stone, and beta when the content is there. Testing takes a very long time. Adding new features means you have to test the features, and test all the interactions those features have with what you already have in place. An overhaul of something like the interface or graphics would require complete retesting.

[Steve Ince]

Fop and Aj - I completely agree with you both.

[RLacey]

Surely Dreamfall and Fahrenheit are very different types of game, even if they broadly fit the term 'adventure'. I love Fahrenheit, but, from what I've seen of the full game, it's not really a style that I'd advocate for a sequel to The Longest Journey...

[Mattsius]

Originally Posted by Fop

Changing the design late into development is a stupid, stupid idea.

Indeed. If you go down that road you will soon have a DNF on your hands.

[RLacey]

I think it will be interesting to see how many people think that Fahrenheit-style gameplay in Dreamfall is a good idea once they've completed Fahrenheit ...

[After a brisk nap]

Wow, universally negative responses so far. That's OK, it was meant as an "out there" idea, not a suggestion I'm strongly committed to. However, the view that this is a bad idea in principle is one I disagree with vehemently.

Originally Posted by Aj_

That change in the camera system was followed by a delay.

Well, they changed the camera system, and a while later they announced that the release date had been pushed back. I don't think it was ever stated that one caused the other. The originally announced release date had looked implausible for quite a while.

That said, so what if it caused a delay? A delay isn't the worst thing in the world, if the improvement is worth it (and if your budget can take it). Judging by the reaction on this forum, I'd hazard that most people would consider it worth the wait.

1. A fundamental change in the graphics engine would result in a big delay, you'd most likely have to reprogram, reproduce, and redesign models and textures through out the game.

Depends on the degree of fundamentality. If you've been following Ragnar's blog, you'll know that they have recently been upgrading the graphics, going from DirectX 8 to DirectX 9 and rewriting the shaders in the process. And that they are currently replacing a number of the character models and textures.

This kind of thing is pretty routine in the development process, especially when your game gets delayed, and I only mentioned it as a comparison: Developers improve the graphics during development to stay up to date, why shouldn't they change other aspects of the game as well?

2. A overhaul of the interface would affect the gameplay, if you had to change how the interactions worked you'd ahve to test the whole game again.

You have to keep testing the whole game anyway. Presumably the game hasn't reached alpha stage yet, much less feature-lockdown, so it's not like they'd be taking a game that is known to work and potentially breaking it. For comparison, FunCom changed the fighting system earlier this summer. (Which sure as hell changes gameplay. In fact, that's rather the point.) Ragnar mentioned:

Originally Posted by Ragnar Tørnquist

It’s still not done, of course, and we’ll keep adding to it and tweaking it until the day we ship the gold master

See? That's the right attitude!

Originally Posted by Aj_

Testing should take place on alphas when the principles are set in stone, and beta when the content is there. Testing takes a very long time. Adding new features means you have to test the features, and test all the interactions those features have with what you already have in place. An overhaul of something like the interface or graphics would require complete retesting.

I think we're talking about different types of testing. You're talking about bug testing, while I was talking about play testing, or what is known in other fields as user testing. Testing your product on users during the design and development processes, and then changing your design based on their feedback, is essential in all fields of creativity. You have to start early (because it's cheaper to make changes early on, and to avoid going down a fundamentally wrong path) and you have to keep doing it until the very end (because some problems don't become apparent until your product is essentially complete).

The FunCom team recently had a major round of testing with users:

Originally Posted by Ragnar Tørnquist

Focus testing is an important part of game development. This is when we bring in people who aren’t familiar with the game, sit them down with a controller and some basic instructions, and let them loose in our world.
...
Focus testers aren’t bug testers. We have people to do that
...
Feedback from focus testing isn’t taken as gospel. It’s dangerous to put too much weight on what a small group of people has to say, because they may not represent the majority of players, but if everyone - or even most - are unhappy with one element in the game, you’d better listen, because it probably means there is a problem.

This kind of testing should start as early on as at all possible, which is usually before you've written a single line of code. Ron Gilbert tested Maniac Mansion by creating a boardgame version of it before implementing it as a computer game. He tested the story of Monkey Island by writing it up as a short story and having people read it.

Yes, the further along you are in the development process, the more difficult it is to make changes. However, it's not impossible, and it's often worth it. Sticking with the Monkey Island example, Herman Toothrot was added during development because Gilbert, Schafer, and Grossman realized that the eponymous island got a little dull and lonely. And one part of the game (Meathook's three tasks) was axed late in the development process because it slowed down the pace of the story. And of course, the interface was modified for the CD release (reducing the number of verbs and adding graphics for the objects), proving that you can make an interface improvement without having a major effect on the gameplay.

[After a brisk nap]

Originally Posted by RLacey

Surely Dreamfall and Fahrenheit are very different types of game, even if they broadly fit the term 'adventure'. I love Fahrenheit, but, from what I've seen of the full game, it's not really a style that I'd advocate for a sequel to The Longest Journey...

That's an interesting objection. While I'm sure the differences between the two games are massive, I think there are also some important similarities that would make it possible to take lessons from one and apply it to the other. Of course, I only know Fahrenheit from the demo and from what I've heard (the game is currently in the mail), and Dreamfall only from what I've heard, so there's going to be quite a bit of speculation here.

For one thing, they are both third person, fully 3D games with a direct-control interface. That already creates a huge area of issues that are relevant to both. Dreamfall's focus field concept appears quite different from the interactions in Fahrenheit, but I can imagine other useful ideas that could cross over, like how to deal with conversations, the use of split screen and camera angles, and possibly even the "action-mimicking" controls.

Secondly, they both feature multiple playable characters, allowing you to experience the story from different points of view. There's potential for Dreamfall drawing on Fahrenheit's earlier example both in terms of storytelling and gameplay. Just for one example, the way the camera cuts away as you're hiding the knife at the murder scene, which allows you to actually search for it later in the game, is a cool design detail that has been praised in Fahrenheit, and could inspire something equally cool in Dreamfall.

Thirdly, they are both (though this is disputed) adventure games with action elements. I don't know how graceful Fahrenheit's fighting system is, but if it works it might be worth having a look at for FunCom. After all, they did say they'd keep tweaking their own fight system all the way up until the game goes gold.

Based on the feedback for Fahrenheit, Ragnar could tweak things like the rate at which he reveals the story (as of last report, the dialog was not yet finalized, and voice recordings have not begun), how the story is introduced (both games are told in flashback by the main character, within a framing setup), or how characters are portrayed. (There have been some criticisms of how Carla is "the worst cop ever". If, hypothetically, Ragnar looks at Dreamfall and realizes that Kian is the worst assassin ever, that might be something that should be fixed. Which brings up the point that it's not just about imitation, but also learning from the mistakes of the earlier game.)

Some changes would be more major than others. Some might have more impact than others. It would be up to FunCom to decide whether they would be worth the effort. I was just hoping to start some speculation and brainstorming around the idea.

[Aj_]

Originally Posted by Snarky

Depends on the degree of fundamentality. If you've been following Ragnar's blog, you'll know that they have recently been upgrading the graphics, going from DirectX 8 to DirectX 9 and rewriting the shaders in the process. And that they are currently replacing a number of the character models and textures.

Dude, I'm in the comments for that post on Ragnar's blog.

As far as I can remember Ragnar stated in the past that DirectX 9 was planned to be used, shaders added, etc... All those things, already in the Shark3D engine. Those things didn't need any reworking, although they are going to tweak somethings they aren't happy with. Updating the graphics in an unplanned way compensating for advancing technology is a whole different ball game.

You have to keep testing the whole game anyway. Presumably the game hasn't reached alpha stage yet, much less feature-lockdown, so it's not like they'd be taking a game that is known to work and potentially breaking it. For comparison, FunCom changed the fighting system earlier this summer. (Which sure as hell changes gameplay. In fact, that's rather the point.)

That's not what I was trying to explain, you have to think that this change took place in June, and I'd be surprised if the game isn't in alpha by now, it's set for spring release.

Focus testing took place in July, after major changes took place, of course they're going to test it then. Major reworkings would require the testing to start over.

That said, so what if it caused a delay? A delay isn't the worst thing in the world, if the improvement is worth it (and if your budget can take it).

Delays can kill a game, funcom doesn't have the biggest budget in the world, increasing development time costs money.

[Martin Gantefoehr]

See? That's the right attitude!

Sure. But at some point - and that point is always earlier than desireable - the tweaking will have to stop, you'll have to freeze the code and the script, and go with what you have. It may be possible to tweak and fix a game in some areas last minute, but it's impossible to turn a game into another game.

If a superior game is released earlier than yours, then that's bad luck, and it can be hurtful. But trying to "learn" late in development ( - by trying to force things into your game that the orginal design totally didn't plan for in the first place - ) is a strategy that will likely leave you with a bizarre ruin.

[Fop]

Yeah, this was something that was reiterated again and again in the book we had to read for our admission exam. While the design has to have a small degree of life even late in the project, you can't go adding new features all the time. What's *especially* dangerous is seeing what new technological developments have taken place and revising your software to match. Sometimes you have to do that to still be competitive, but you should avoid it if at all possible. The costs increase sharply the later into development you are.