The trouble with Bifrost graph

[Black.Bart]

The following is a comment I posted on one of the YouTube tutorials and was suggested to also post it here... I also made a few additions... Please do keep in mind that this comes from someone who doesn't have access to the dev team like major studios do.

Okay, I'm just going to throw my 2 cents here about the Bifrost graph:

1. You broke the first rule in the book of the UI and UX design: Never, ever, ever let a programmer design a user interface. This feels like it is exactly what happened.
2. You made something by programmers for programmers only, not the target audience that primarily uses Maya, which are artistic in nature and aren't very good with rigid logic.
3. You assume the user has the same knowledge or background that the people who coded this have.
4. This thing has so many gotchas in it it's not even funny any more.
5. When someone in the tutorial actually says: "We have to guess a little bit.." it's a complete fail of the devs to implement a descriptive objective structure for the end user. Not a good sign and again you assume knowledge in the end user that they actually know what to search for. A shelf on the top with node groups or ready made tools would be a good start to remedy this issue.
6. Node naming is in many instances adapted to service only one type of user - the programmer. This is continued in the arrays, floats, integers etc... Usually people don't know what these mean, what are they used for or when they should even use them. It's something from the programming world and it should stay there. You need to find more descriptive names like "Value list" or "Decimal", "X Y Z", etc...
7. Build actual tools. Don't give me nodes to build the most basic of tools. I just want a Translate node that does scaling, moving and rotating and gives me these options in the parameters tab. That's it. Don't give me everything in pieces and expect me to assemble it the way you envisioned it should work. This is not LEGO and we have a job to finish.
8. Error messages are also a problem. You give user an error message just by adding a non connected node to the graph. This sends a false signal to the user. The first time this happened to me I thought I put a wrong node in and was just confused. An error message should only be displayed when a user actually does something wrong, like inputting a decimal into an integer value box.

Look, I like Bifrost, I think it's a good and valuable addition to Maya, we need it, but it's confusing as hell.

I've seen a lot of software make the same mistake. You can have the most advanced piece of equipment on the planet, but if nobody knows how to use it, what good is it then?

I will now leave you with a question for the devs...

How many times did you say to yourselves: "The users just don't understand"?

[ihooper]

Hi @Black.Bart 

I wanted to respond to your post because I think it represents how many users currently feel and I hope to provide some context for why Bifrost is the way it is today. 

Your comment that we made something for programmers, and not the audience that primarily uses Maya is partially correct. When designing a UX, it is critical to understand how the user works and what their objectives are. For a tool like Maya which has many different users and many different workflows, we must necessarily choose which persona to target first. Our focus has initially been on technical directors and people who are already extending Maya through the use of Mel or Python. However, even from the start we knew that we wanted to eventually make Bifrost more accessible and useful for a broader group of Maya users.

Making a tool that makes sense to a programmer, which builds on their knowledge and expectations and which also is understandable to a non-programmer is the challenge we now face. As you pointed out, even the choice of what to call things can make the experience confusing and alien. In time we will have more high-level compounds where more "common sense" naming will prevail, but as a general design principle we also don't want to "black box" the complexity. Ideally, we will have tools built on Bifrost that are easy to use and accessible (such as https://www.envitscript.com/), but which also introduce some of the concepts of visual programming so that people who are interested can start digging in and modifying and extending the tools to better meet their particular needs.

You've made several good points (such as all of the "gotchas" and less-than-helpful error messages) and I just wanted you to know that I agree and that we are working on it. We have a long way to go still, but we know that for Bifrost to be successful it needs to be accessible and useful to a much wider group of people. We will get there and are actively working on it.

In the meantime, I know it can be especially hard for people who don't work in a large studio with their own developers. Fortunately, some of those technical experts are incredibly generous with their time and skills and have been building lots of great compounds, graphs and tutorials which they are sharing with the community. I'm incredibly thankful to everyone who is helping people cross the complexity chasm while we are still building the bridge.

[Black.Bart]

First of all, I am thankful for the reply. You explained your views very well. I understand you have a specific group of users you want to target first. That's fine by me. I have a better understanding why Bifrost is made the way it is. It is also good to know that you are actively thinking in the direction of simplifying the interface. But let me be the annoying customer for a bit... how much longer will ordinary users be kept waiting? A year? Maybe three? If we do get more high level tools later down the road, what sense does that make for me to start learning now and put all my time and effort into this if it's going to become easier to learn later on? Not much incentive there.

Making a simple translation node doesn't "Black box" anything, users can still dive inside to take a peek and modify anything they need and even start understanding the basic logic behind the scenes. This is what baffles me. You could have already made a couple of basic compounds that even TDs can use as shortcuts. You can satisfy both groups of users, just start making a few high level compounds, see what works, what doesn't. It doesn't brake anything and everyone can benefit from them. That's the point I'm trying to make.

When it comes to learning, yeah, you have a couple of people that help, make tutorials and whatnot. But it's not a good sign when even they stumble through Bifrost in the recording and are just as amazed, as to what is going on, as we are who are watching it (you can't tell, but I am laughing right now while writing this, also I've watched a lot of tutorials from different people). I also came across a person that made some truly amazing stuff with Bifrost, and he/she (not sure atm) was an MIT graduate. Ok, if that's the level of knowledge I need just to make something, I'm out. I also really don't want to buy scripts for something that I don't even know if I will keep using later down the line, mostly for the fact that you don't change anything for ease of use.

Look, I'm not angry or trying to be annoying or anything like that. I just want to explain the frustration that I and probably many more users are going through so at least you can get a different perspective and possibly a few notes on our experience. But for now, I congratulate you for basically making Matlab inside of Maya 😄