Why do we seem to fear improving blender?

[Ace_Dragon]

@Luftmensch

That’s an assumption based on not thinking of alternate possibilities. An alternate way to think of it is, the interface should add options that make no-scroll possible, which currently it is not. The ability to split windows isn’t contingent on never allowing the interface to overlap and restricting task-based workspaces.

When you look at the horizontal bar in Andrew’s video, you discover that it relies on the wideness of the HD screen to not require scrolling, but if you split the screen in half to make room for both the 3D view and the UV editor, the edges of the top bar get cropped off the screen and you then need to scroll to see everything (only it’s horizontal scrolling this time).

So you have to decide, is scrolling such a major abomination that we need find a way to nest options and tools inside options and tools to remove the scrollbars completely (never mind the fact that I just use the middle mouse button instead of the actual scrollbars)?

Which would be quicker for the user do you think…

top bar > face tools > more > advanced tools > quick move down menu to activate tool or…
middle mouse down > move mouse down the list a little to activate tool

Or maybe it’s still just as fast to do…

space bar to get search menu > type ‘My’ > press entry labeled ‘MyTool’ (can also be improved if users can attach tool entries above the search box to make a menu of oft-used tools).

[Luftmensch]

Right, so who does it hurt to have no-scroll options if you can just use the search tool?

The problem with scrolling isn’t alleviated whether you use the scroll bar, middle mouse, or scroll wheel. The problem is that useful information is obscured without being represented on screen (in drop-down menus at least you can see the higher-level category), and when you find the appropriate info, which usually only needs a quick glance or a single edit, if keeps taking up screen real-estate because the zones system makes it hard to quickly open and close menus so it’s easier to leave it. Besides, scrolling is tedious. There’s more eloquent arguments against heavy scrolling, but I basically covered them.

[Ace_Dragon]

@Luftmensch

Right, so who does it hurt to have no-scroll options if you can just use the search tool?

When looking at the mockup, it might actually force more dependence on the search tool and I would prefer that I don’t need to use it whenever I need to use something that’s more than a basic operation.

In fact, I think that the search tool is a necessity for a few too many cases right now, which is part of the reason why I see a good point in the right-click context menu idea along with pie menus, it’s also a major reason why I say users should have the ability to use a small button or a hotkey to dock the specified operator name to an entry above the search menu, essentially making a custom menu (which would then be preserved after saving the scene and quitting Blender or becoming part of the startup.blend options so you have your entries in all new scenes you make).

In my opinion, I find the pie menu project along with the GSoC toolbar improvement projects as good examples on how we can build some nice improvements onto the current system (which I hope they find their way into trunk in some form). I would note that the toolbar improvement project will also help to address the issue of access to various tools, so overlapping partly with the crux of Andrew’s proposal without changing the existing concepts and workflows completely (there might be some change, but it’s not a tear-down in the vein of installing a new one).

[arcr]

Blender needs to be changed to facilitate the work for the professionals. If that happens Blender would facilitate the work also for beginners. Complete beginners that doesn’t work with any 3D software and doesn’t know nothing about 3D don’t need to be the target. The target needs to be the newcomers from the other 3D software, but that have some acknowledge about 3D.

[moony]

@danilius

That was not a random example out of a hat.

Microsoft spent hundreds of millions of dollars researching that interface, which was spent on focus groups, user groups, psychologists and what not. Hundreds of millions of dollars, and they still got it wrong. That’s the point. Making changes means that one can make really bad ones quite easily.

This is quite dissimilar to the 2.49/2.5 changeover. That was a genuine leap forward; the whole interface was upgraded from wood and stone to carbon fibre. Lighter, faster, cooler-looking with bling. In all seriousness, the new interface brought Blender into the professional realm in and of itself.

Of course I now realise that the people who first whine that “Blender isn’t professional” then whine “Blender is too professional”. I simply despair.

Actually you are correct - it wasn’t a random example out of a hat…its worse than that. It was a single example chosen specifically because it suited the argument.

It doesn’t matter how much Microsoft spent…its irrelevant to this discussion. The Blender development team aren’t Microsoft and they way they go about developing Blender doesn’t follow the Microsoft model. Microsoft have demonstrated time and again that they are arrogant and assume they know what is best for their users. Blender development doesn’t have to go down this route.

As I said before - most of what is being discussed and proposed constitutes minor tweaks, clarification or harmonisation of functions within the existing Blender UI (which is actually quite good). I haven’t seen many (if any) suggestions of a fundamental and radical redesign (a-la windows metro) - so comparisons to Microsoft’s balls up doing exactly that is nothing more than a straw man argument.

You clearly acknowledge that the change to the existing blender UI was a major step forward. Given the successful track record so far - why assume any further evolution would be negative? The existing blender UI was a great move - its actually the reason I finally made the leap to Blender…but it still has flaws, inconsistencies and inefficiencies. Many of these flaws can be easily fixed, whereas some will require more thought - but we cant deny they exist.

[moony]

@m9105826

Polling the user base to make decisions is a TERRIBLE decision, and anyone in design knows this. How does the saying go? A camel is just a horse designed by a committee?.

Yet going back to the Microsoft example - the fact that they didn’t poll the user community (or if they did - they didn’t do it properly) lead to a huge backlash when windows 8 was released - and is responsible for their backtracking with windows 8.1 (and reinstatement of the start button).

You cant do everything by committee of course - nothing would ever get done - but you have to be mindful of the requirements of your users.

As for that saying - its a bit silly IMO as it appears to assume the camel is inferior. In fact - depending on what the user requirements are for the beast being designed (e.g hot weather, low water survivability) which you only get by asking the users - a horse may not in fact be fit for purpose. What good is a horse - when the users needed a camel.

I have been involved in the design of many systems - and have seen what happens when users aren’t consulted properly - it’s rarely a positive outcome. Of course the devs have to arbitrate and not everything users want is possible - but a balance has to be struck.

[DruBan]

@Luftmensch

That’s an assumption based on not thinking of alternate possibilities. An alternate way to think of it is, the interface should add options that make no-scroll possible, which currently it is not. The ability to split windows isn’t contingent on never allowing the interface to overlap and restricting task-based workspaces.

I guess I just don’t understand, but I would like to, how you can present all the information on screen, thus avoiding scrolling, and still take up less space simply by changing its location. If it takes up less space then there’s less information because even in the smallest form it takes up so many pixels.
Either that or you are trading scrolling for opening and closing tabs. Or by having the info hidden until you click and open a drop down panel. I think that’s all the options.
I’m not even talking about splitting work (display) windows, so let’s just leave that aside…

[Luftmensch]

Well, you rarely need all the information at once. Although it’s a little too close to modal for my liking, Price’s interface has the advantage of being task-focused. Certainly it’s designed with very basic needs in mind (relative to the breadth of Blender’s total functionality), but it’s designed to present everything you need and only what you need.

Trading scrolling for switching tabs isn’t great, but having collapsible tabs is actually pretty handy. I usually find that in Illustrator, I only want a small handful of tools and info most of the time, but if I need to see something else, fine, I click the appropriate icon, and a portion of my screen is sacrificed to show what I need to see. I can even set it to automatically collapse as soon as I click anywhere else.

The advantage that tabs or icons has over scrolling is that when you have to scroll, you’re forced to hold a picture of the menu in your mind or search it manually each time you want to find any information. Tabs avert this problem by giving you a smaller, easier to recall list. Icons help even more because icons are even faster to recall. You could improve it even more by displaying some information in the compacted version. At least one nice thing about the properties panel is, it does give you an icon for each tab. Andrew’s idea, however, is rather than to give you everything at once in a packed form, is to give you what you need at the moment in an open, readable format.

Perhaps the Properties Panel would have to remain even in Price’s design, since you can’t predict every marginal use case, but the theory is that you will generally only need a limited information for a given task and more than that is clutter. It’s still a very young design so in sure with some expert counsel that could be ironed out.

[moony]

@jskurias

They must have tried millions of different UI layouts and finally came up with the most efficient one, which they call “ribbon”. And big companies like Autodesk followed suit. First time I saw it was with AutoCAD’s 2010. It was still a WIP back then. I didn’t like it at first, but it finally made sense. There’s no other way to organize a huge program like a 3D app other than what Andrew Price demonstrated in his video.

I worked as a Microsoft Excel spreadsheet developer for 5 years. The ribbon is far less efficient IMO. Instead of having all functions 1 click away on a single highly customisable toolbar - you now have to navigate tabs on the ribbon.

Also - unlike the original toolbar - the ribbon tabs cannot be undocked and left hanging over the working area - so tasks like drawing shapes is much slower since the ribbon function collapses after each shape is added - meaning you have to navigate to the add shapes function each time.

The concept of the ribbon has some good elements - but it hasn’t been executed to the best of its potential.

[danilius]

@moony

Actually you are correct - it wasn’t a random example out of a hat…its worse than that. It was a single example chosen specifically because it suited the argument.

It doesn’t matter how much Microsoft spent…its irrelevant to this discussion. The Blender development team aren’t Microsoft and they way they go about developing Blender doesn’t follow the Microsoft model. Microsoft have demonstrated time and again that they are arrogant and assume they know what is best for their users. Blender development doesn’t have to go down this route.

As I said before - most of what is being discussed and proposed constitutes minor tweaks, clarification or harmonisation of functions within the existing Blender UI (which is actually quite good). I haven’t seen many (if any) suggestions of a fundamental and radical redesign (a-la windows metro) - so comparisons to Microsoft’s balls up doing exactly that is nothing more than a straw man argument.

You clearly acknowledge that the change to the existing blender UI was a major step forward. Given the successful track record so far - why assume any further evolution would be negative? The existing blender UI was a great move - its actually the reason I finally made the leap to Blender…but it still has flaws, inconsistencies and inefficiencies. Many of these flaws can be easily fixed, whereas some will require more thought - but we cant deny they exist.

You are failing to understand the thrust of my argument, which is this: UI changes are not inherently good. There is no way of predicting how good a UI is going to be until you suck it and see. This has proven to be historically axiomatic. I was simply using WIndows 8 as a great example, since here you have the ultimate development and research environment, with more PHDs than any other company on the planet doing this stuff, more researchers, academics and other professionals involved in the field that in sum total outrank BF in every way, and despite the enormous resources available to them, the entire WIndows 8 interfaced is almost universally loathed.

Theoretically all those resources available to Microsoft would have ensured a great success for the WIndows 8 GUI; in reality it simply did not deliver.

There are many more examples of well thought-out UIs bombing for the silliest of reasons. There are many examples of seriously ugly UIs working OK since the users learned them as they had no choice because their jobs depended on learning how to use them.

The point is, there is little guarantee that any kind of UI design is actually going to work

Now, given this situation, you have to ask yourself a simple question. If we institute a tabbed ribbon for example, is this going to be an enhancement or a disaster? Well, we don’t know until we try it out - this is historically axiomatic. So, should we divert development away from bug-fixing and improved or new functionality for something which we don’t know will work?

If Blender really did not work well, everyone would argue it’s time for an overhaul. That’s why 2.5 was so successful. It was a huge paradigm leap, and resolved several key issues with 2.4x that really sucked badly.

In this case, though, you yourself are arguing that they are minor tweaks being proposed. That implies that Blender itself is useable, but perhaps not quite as easy as you or others would wish it to be. More so, when you ask users who have a great deal of experience using Blender to make a living, they will tell you as they have told me that they manage quite well, thank you very much.

So, the upshot is this. We have two competing requirements: minor UI fixes or faster/better/new features. Which one takes priority, when we can only serve one of them? Which one carries more weight?

More importantly, which of these outcomes are more likely to succeed?

The answer is simple. Bug fixing, feature improvement and development has historically proven to be far more likely to succeed than UI changes. That is not to say that every project has succeed. Many have been started by one-man dev teams and failed.

So we examine successful feature development projects and discover that you need a critical mass of developers for the best chances of success. BF has that critical mass, which is why BF-led feature projects have had a really good track record, and GSoCs have a much higher failure ratio. This is true in many major FOSS projects - think of Inkscape, Scribus and Apache etc.

Experienced users manage to work every day with the existing UI which implies that while it might be able to improve, it’s nonetheless very usable.

In summary, it seems more logical to bet on a near certainty that is definitely required (feature development) than one with a 50/50 chance of success that does not have an absolute requirement (eye candy).

[danilius]

@moony

silly IMO as it appears to assume the camel is inferior. In fact - depending on what the user requirements are for the beast being designed (e.g hot weather, low water survivability) which you only get by asking the users - a horse may not in fact be fit for purpose. What good is a horse - when the users needed a camel.

You have chosen to deliberately break the analogy in order to make your point.

The committee was charged to design a horse. They ended up with a camel. That means how ever amazing camels are, the committee failed singularly in their task, which was to design a horse. Not to design a camel.

Here is another example: a tank is a car designed by committee.

Now what was the task? Design a car. So, five seats: check. Space to put your shopping: check. Engine: check.

Now, both a car and a tank fit that requirement. There are a few minor differences between them, in that the car weighs one ton and the tank 40, the car has wheels and the tank has treads, the car can travel 50 miles on gallon of fuel, and the tank around 2.

Which one would argue best fits the original task? Would you argue that the tank is better because all the armour makes you safer? Because it can cross rivers no car can? Because you can drive it across the countryside? Because it can survive all kinds of weather, gas and biological warfare?

[Luftmensch]

All these mixed metaphors are really putting the cart before the camel, don’tcha think?

[danilius]

@Luftmensch

All these mixed metaphors are really putting the cart before the camel, don’tcha think?

Is there a smiley for a facepalm?

[system]

One question to all you nay sayers: how many Software and GUI`s have you designed in your life? How often have you dealt with customers and feedback? What qualifies you to lead the GUI debate? How many clue do YOU really have about GUI design?

“If I had asked people what they wanted, they would have said faster horses.”

― Henry Ford

Leave the question of GUI design in the hands of GUI designers. Not in the hands of the Blender Users. Also not in the hands of the Blender developers. The GUI designers are the specialists, and they know what they do. At least in most cases.

For me Andrew has made the jump from a Blender user to a GUI designer. He has proven in his videos that he has already more clue about GUI design than anybody else in this thread here. There might still be a chance that he is completely wrong. Nobody`s perfect. And even big companies makes nonsene, see Win 8. And his mockup still needs to pass the Ton hurdle. But when looking at the Blender Userbase, which is to 99.9% made of hobbyists, then i would say he is completely right with nearly all of his points and his mockup. It is very consistent and well thought. Even when details like the size of the ribbon area needs a bit more adjustment.

I remember the discussion when 2.5 with its new GUI was announced. Lots of the mentioned fanboys popped up, with swearing that they will quit on blender with 2.5, that the end of the world is near, and so on. And with flaming everybody who wants the new GUI. And what happened? Blender got more users, a better reputation. And the ones who cried the loudest are the most happy 2.5 users now. I see the same thing happen here again.

[danilius]

@Tiles

One question to all you nay sayers: how many Software and GUI`s have you designed in your life? How often have you dealt with customers and feedback? What qualifies you to lead the GUI debate? How many clue do YOU really have about GUI design?

I can only speak for myself. I wrote commercial code, and invented features, some of which appeared much later in Windows 7 (no, they didn’t steal it from me). I developed interfaces for engineers to migrate from existing hardware to software; they were well-received when reviewed by a senior archivist in the BBC. I developed UIs when we stilled called them GUIs for advertising marketers to analyze ROI on their campaigns.

That does not make me a UI specialist, it simply answers your first question.

@Tiles

Leave the question of GUI design in the hands of GUI designers.

In no way is that a guarantee of success. UI design is still a balck art.

@Tiles

For me Andrew has made the jump from a Blender user to a GUI designer. He has proven in his videos that he has already more clue about GUI design than anybody else in this thread here. There might still be a chance that he is completely wrong. Nobody`s perfect. And even big companies makes nonsene, see Win 8. And his mockup still needs to pass the Ton hurdle.

He read two books, designed some widgets and now he’s a GUI designer? Just how low are your standards?

[system]

Thanks for pointing that out. And i develop games since several years

In no way is that a guarantee of success. UI design is still a balck art.

Indeed. But there are some Design standards that you can follow. Some Design standards that you have to follow. And there are some no go`s. Unfortunately there are more than enough category 2 and 3 flaws in the current Blender GUI.

He read two books, designed some widgets and now he’s a GUI designer? Just how low are your standards?

I didn`t say that he is the best GUI designer. Just that he did the jump, and that he has definitely more clue than most ( i stand corrected here, i thought all … ) people in this thread. And he did a bit more than just reading two books and designing a few widgets. You may have noticed it when you really watched the videos. He has put the finger into nearly wound and flaw of the current GUI. Which are nagging me in the same way like Andrew since several years.

I don´t even say that the mockup from Andrew is the best i`ve ever seen. But it is definitely much better than what we currently have. Make a better one

My standards are in fact very high. As told, i make games. Means GUI design is also nothing uncommon for me. You better do a good GUI job at games, or you don`t have any players.

[danilius]

@Tiles

Indeed. But there are some Design standards that you can follow. Some Design standards that you have to follow.

This might be what you were told by some author or lecturer, however it is simply not true. Every software developer has two ways to go: use an existing widget, or invent your own. The latter prevails when the former fails.

@Tiles

And there are some no go`s. Unfortunately there are more than enough category 2 and 3 flaws in the current Blender GUI.

If this were true, Blender would be unusable.

@Tiles

I didn`t say that he is the best GUI designer. Just that he did the jump, and that he has definitely more clue than most ( i stand corrected here, i thought all … ) people in this thread. And he did a bit more than just reading two books and designing a few widgets. You may have noticed it when you really watched the videos. He has put the finger into nearly wound and flaw of the current GUI. Which are nagging me in the same way like Andrew since several years.

Even the most experienced UI designers will have no experience of developing for anything like Blender. It’s a largely untouched field. UI designers do games, websites, general software (Office et al) and little else. Blender does not fit into any of those categories. Their experience will not translate directly into something useful.

@Tiles

I don´t even say that the mockup from Andrew is the best i`ve ever seen. But it is definitely much better than what we currently have. Make a better one

My standards are in fact very high. As told, i make games. Means GUI design is also nothing uncommon for me. You better do a good GUI job at games, or you don`t have any players.

So you will willingly divert precious development time for a UI that you don’t even know will work? How does that make any sense? Development time is precious, and should only fix what is broken. The UI is not broken. Ubly - yes. Broken, no.

Why oh why have so many people fallen under Andrew Price’s spell? He is selling eye candy, that’s all. I like eye candy too, but I would rather have volumetrics in Cycles than shiny buttons.

[EnV (Enrico Valenza)]

@danilius

I like eye candy too, but I would rather have volumetrics in Cycles than shiny buttons.

• 10.000 and 10.000.

[system]

This might be what you were told by some author or lecturer, however it is simply not true.

Yes it is true. This teached me 14 years of experience and own mistakes with GUI design for my games and apps. I don`t need a book nor somebody to think for me to know that RMB select is nonsense while all other apps uses LMB select.

If this were true, Blender would be unusable.

It partially is indeed unusable. There are hidden tools where you simply have no clue that they exists. And it is slow and complicated to use in more than one area. Unwrap twice to get the conformal result? Please …

Even the most experienced UI designers will have no experience of developing for anything like Blender. It’s a largely untouched field. UI designers do games, websites, general software (Office et al) and little else. Blender does not fit into any of those categories. Their experience will not translate directly into something useful.

Oh, so Max, Maya, Cine, etc. did get their UI out of the blue? I see …

And again, there are some general GUI design principles that applies to all GUIs. One of the most common principles is KISS. Keep it stupid simple. I wouldnt call it simple to have hidden tools and workflows where you need to write down a list of commands because it is too much to remember, while you need one mouse click for the same task in another software.

Another is consistency. Not radiobutton here, and regular buttons there, RMB select here, LMB select there, a standard value of 1 here, and a standard value of 100 there for a similar object, and so on. For the other points watch the videos from Andrew. You might learn the one or another thing.

So you will willingly divert precious development time for a UI that you don’t even know will work? How does that make any sense? Development time is precious, and should only fix what is broken. The UI is not broken. Ubly - yes. Broken, no.

I see so many flaws, its a shame not to fix them. Its not just ugly. It slows down the workflow, makes the work complicated, and scares away generations of new users. And in that regard it IS broken and not working. It can just get better.

The ribbon GUI attempt is a proven and working concept in many other softwares. It may not be the best attempt, but it works. And is better than what we currently have.

Why oh why have so many people fallen under Andrew Price’s spell? He is selling eye candy, that’s all. I like eye candy too, but I would rather have volumetrics in Cycles than shiny buttons.

To your surprise, you can have both! The development at Cycles can still continue. But while Cycles still renders without Volumetrics, you will scare away tons of users every day with the old GUI.

No. Its not just eye candy, not just shiny buttons. A GUI is an essential part of the software. Function is not everything. Sculpt with Sculptris, then sculpt with ZBrush, and then tell me again that GUI makes no difference, and that buttons are eye candy. Thats a really scary statement for somebody who claims to have a clue about GUI design.

Its the opposite. Why o why do so much fanboys defend the flaws and quirks? I definitely dont need a spell nor somebody for me to think, as already told. I said some of the same things that Andrew tells now before years already. I have lead the very same discussion here with the 2.5 UI already. But got bashed away then. The Blender Fanboy problem …

It`s a proven fact that can be seen at any 3D forum, even here, that new users, and also old users, have massive problems with the standard Blender GUI. This needs a change. And the Blender devs agreed at that point, finally. They are convinced already. Not about the Andrew GUI, but that there needs to be a change in general.

So the only remaining question is in which direction the change leads. Maybe into the ribbon direction from Andrew, maybe into a completely other direction. I don`t mind in which one, as long as the GUI finally becomes user friendly and consistent, and follows common standards.

The mockup from Andrew is so far the most convincing attempt i have seen. As told, when you really think that you can do it better than Andrew, when his mockup is really this bad, then, where`s your mockup?

[danilius]

Very briefly, most 3D apps use their own paradigm. So does Adobe.

You clearly have no idea how complex UI programming is Blender. It really is one or the other. Ask Campbell Barton.

So, given the choice, what would you have? Eye candy, or volumetrics in Cycles? Tabbed menu bar or better hair physics, a faster water sim, a faster smoke sim that works in Cycles?