From a 25 years experience programmer: there is no such thing as a “quite stable” API !!! <insert your favourite facepalm image here>.
API stability is a binary thing: either it is stable (all previous calls work in the same way, with the same arguments, with the same names) or it is not. Even just renaming a single call out of 10’s of thousands is UNSTABLE because somewhere there is a program/plugin which does no more compile/link/run.
This is a real block for Blender adoption into existing 3D shops: those people rely critically upon commercially/internally developed plugins for their workflow. Blender development makes currently impossible to develop plugins because every two months people would have to check and fix their code to run after underlying API changes and people do not want to do so and existing non-trunk plugins simply rot away.
Anyway the discussion is pointless: Blender PTB have decided this way and therefore the situation won’t change (until the day in which, like e.g. ngons, all of a sudden the impossible/not needed becomes the latest fashion…).
“All of a sudden”? There were at least 2 ngon-capable mesh structures attempted before bmesh and bmesh itself was under development for years. I’m not going to get into the hows and whys on that, but to characterize Blender development (particularly regarding bmesh) as only following what’s fashionable is grossly inaccurate.
@Zalamander, IIRC there was some change to PyNode API, but this is a new area are very few scripts use.
@jpb06, point taken, if you treat it as binary - strictly speaking blenders API is unstable, if you Auto-generate an API from an application - then extend and improve the application, you pretty much can’t help breakages when you make minor changes.
but saying this is why blender has not gained popularity is a stretch (probably its a contributing factor in some situations).
This API discussion is getting a bit old, it keeps getting raised but I don’t see things changing.
You can point at an unused fluid-sim-property being removed as breaking the API, but in practice we have a bunch of addons included with blender that didn’t need changing because of API updates in quite a while now.
The alternative is to have a much more limited API which we keep stable but this means you loose access to lots of useful stuff - its a tradeoff.
Perhaps we could define a subset of the API as stable, but not sure this really helps so much… if one area of blender is re-written (Particle-Nodes for eg), then the API for it changes (breaks) too.
Okay then, as a professional CG animator and TD for 20 yeas now, I don’t think this has anything to do with Blender adoption. If anything it’s the other way around: Plugin developers aren’t going to port their code to a program until it’s perceived as professional.
If you ask me, I think there are three main things going against Blender for professional use:
1.) There has been lack of professional looking work out there for it. It’s all about the work that is being consistently done with it. This is changing day by day but it doesn’t happen over night and you can’t keep showing the same images over and over again. It has to be a consistent improvement and it has to be outstanding. You have even more to prove due to the next item…
2.) The software is free. Really, folks, a lot of people just have a stigma about free software. Like, “It’s free, how good could it be?!” Recently Brad from Luxology reviled that one of the most common issues that they ran into while trying to sell Modo to large studios was that they actually thought the program was less professional because it was so cheap. People would literally ask them, “Why is it so cheap then?” It’s one of the main reasons they sighted for raising the price in the last few releases. $1,000 software might be good but $10,000 software must be really good for it to cost that much. Obviously not everyone falls for that. But I think you would be surprised if you knew how these buying discussions are made. Think about how popular Apple products are right now. There are much cheaper devices out there but there are still a lot of people who believe that you get what you play for.
Obviously this will/should never change. It’s just something Blender and the people who use it are going to have to live with. We need to work even harder to become better artists and prove to them otherwise.
3.) the UI navigation and user interaction model is very foreign to users of major 3D programs. And no, I’m not talking about how the GUI looks. I’ve spoken to many people who learned on Maya and have a story or two about the first time they tried to use Blender and how off putting it was. It took me quite some time to fully understand how to change specific keys and clicks to work more like I was expecting them too.
As far as developing plugins for Blender, most small studios want a program that already works. Not one that they buy and then have to buy a bunch of plugins for. I think the main reason plugins are so popular in other programs is because the development of those programs is so slow that others can come in and fill a need when it arises. Basically, it’s because the software is lacking in the first place. With Blender, we never really have to wait to long to get newer and better tools. I consider this to be one of the advantages of Blender if you ask me.
So, if you’re talking about studios that do a lot of custom development for their projects then you must be talking about bigger studios like Sony and ILM. Are you really expecting these companies to switch to Blender? Actually, are you expecting them to switch software at all any time soon? No, surely not.
So you have to start with the little guys first. You get in with the cutting edge crowd and help them show that Blender can work for commercial production too. Little Medical animations or industrial videos and half baked ninjas or pretty car renders aren’t going to help here. Sorry guys if that sounds mean and insensitive but that’s the way the rest of the industry sees it. Like I said, the quality of work has been improving recently but we need more and more in order to grow out of these stigmas. So, I don’t see how larger studios who do a lot of in-house development have anything to do with this.
And, over-and-above what you’ve well-said here, Indy, there is also the very pragmatic notion that “a large studio (or a data-processing shop of any kind at all) is not going to be inclined to change a-n-y-t-h-i-n-g about a production-process that they know will consistently bring home the bacon.”
When I was at University, never-mind how many years ago, a department maintained an IBM/360 computer (with magnetic core memory … the first and the last time I actually saw that), because it could run IBM 1401 Autocoder emulation. (The 1401 was an even older machine, and Autocoder was an even more archaic way of writing programs.) They did this strictly to maintain software continuity across the entire lifespan of a substantial Government project. Years later, another department paid to maintain a specific version of the FORTRAN-G compiler, even though IBM considered it to be obsolete. Once again, the requirement was that “the latest” computer-generated results for the project had to be identical to “the earliest” ones. Any re-runs had to produce the same results, and do so in the same way. I’m quite certain that the same kind of reasoning applies to a substantial CG project. (There are probably still some projects-in-progress that are using “Blender 2.49 release exactly-this.” Same song.)
Today, real, professional-quality work is coming out of Blender: it’s not a toy anymore. But, even so, software products almost never “displace” one another in a production environment. The risk is simply too high, and the return-on-investment is not there. I don’t think it makes any sense to expect it.
Indy_logic i deeply and honestly agree in everything you mentioned, in everything you said, in our studio, my partner hasnt gotten into blender because as he describes it: “its a ferrari with the steering wheel on the ceiling”, it’s the future, and its great he acknowledges it but, he feels “the main things like clicking and selecting are cumbersome and he will wait until they make blender more human” and honestly even though i can make excuses and i totally love the way blender works for some things, it took me quite a while to deal with stuff like right click cancels and left click accepts operation, right click selection and such…
“its a ferrari with the steering wheel on the ceiling”… Hahaha! That’s an awesome quote!
I was just thinking, it’s more like a free V12 super car that you drive with a joystick in the center counsel where the gear shift usually is. You know, I could totally see a lot of people thinking that would be cool and would love it. They would say, “I tried driving other cars but they never felt right.” But for the rest of us, we would have to figure out how to Jerry-rig a steering wheel in there just so we could drive it.
As someone who has never used any other 3d software, I can confirm that it all is alien until you use it. I have never “used it”, but I have clicked around 3dstudio max at one point and it was just as confusing to me as Blender was when I first started using it a year ago. Using blender makes complete sense to me now (for the most part, at least as far as my experience will let me refer to), but as I have said, I have never used any other 3d software, so my brain has not be predisposed to alternative ways of navigating the 3d world and working on meshes and the variety of other features. For once I feel like I am at an advantage for coming late to the game and not playing other sports… if you know what I mean.
People seem to forget that what feels natural is just what they are used to. Particularly with intangible tools, how you expect them to work is defined by your previous experiences.
Yes, this is completely true. I’m all to aware of that point. That’s why I said, “It took me quite some time to fully understand how to change specific keys and clicks to work more like I was expecting them too.” Notice I didn’t say, “…to get it to work the right way.”
The issue however is that there is a whole world of 3D programs out there that all behave basically the same way and use a lot of the same basic tool paradigms. Over the corse of my career I’ve use a lot of other 3D programs. I’m very ambidextrous with most of these programs but this is because they all work basically the same way. You just need to figure out what the different programs call the same features. If I click on a tool and then click with my mouse in the 3D view, I know pretty much what’s going to happen. If I click and drag on the background, I’m going to get a box selection; If I click once on the background, I’m going to deselect everything; etc. This allows me to work in many different environments with many different programs. It’s very important to my business as a freelance artist. With Blender, if I click on a tool and try to click in the view it doesn’t behave like I would expect it to (Sometimes it does though but that creates inconsistency). Sure, if you stick with it, Eventually you will get used to it. But then, what happens when you go back to Maya or Max? If you want professional artists who use these other 3D tools to be able to use your 3D tool too, you can’t be that different. It’s okay to be different but no so much that you disable users from other software. I’m not saying Blender needs to change everything and be like everyone else. But I think at some point development resources should be spent to make a kind of compatibility mode that helps ease the transition into Blender. Right now, you have to click a bunch of buttons and set up your own hot keys in the preferences (but of course, you would need to know Blender well enough to know how to change theses things) to get it to behave how you like, but I see no reason why this couldn’t be a simple button in the UI somewhere. “Software Compatibility Mode” would be fine. In fact, I would bet that if one day you announced that Blender had a new “3D software compatibility mode”, you would get a flood of new users.
P.S. I’m sorry for derailing this thread but this is a subject that is very important to me. I want Blender to get past some of these stigmas and start getting used in more professional animation studios. So I take any chance I can get to talk about it.
Very true. I often work with autocad on one monitor and blender on the other. They have completely different control schemes and often throw me through a loop when I switch from one to the other. (of course, I primarily use blender, so I blame it all on autocad ).
On basic tool paradigms, I’d like to raise another point. Photoshop. The most widely used image editing software, definitely the industry standard has completely foreign controls. I don’t know any other program that uses spacebar as the pan control, while being blissfully unaware of middle mouse. GIMP does too, but I imagine that is only a control scheme to make it easier for photoshop transplants. Spacebar is not the best way to pan, by any stretch of the imagination. It isn’t 1997 anymore. In GIMP you can at least use the middle mouse as well. Almost any other program with 2d panning will use the middle mouse. This is the sort of tool paradigm incompatability that blender is constantly being held to the fire for, but since adobe is the industry standard, any control weirdness is just a given.
Maybe I’m just bitter because every time I switch to inkscape from photoshop, I end up making thousands of copies of an object while trying to pan
I knew I started having fun in Blender when I was going in to work, building assets in Photoshop, and then trying to scale selections using ‘s’ -> ‘x’, and zooming with my ‘+’ key instead of using control. I wasn’t disagreeing with Indy either, I definitely think that similar controls to other applications could make other professionals use blender more easily. I was just stating that as someone whom of which has had his first 3d experiences in blender that I have no qualms with blenders controls and actually like them a lot.
Actually, a lot of other 2D programs use the space bar for panning. It’s quite ancient actually. Even before Photoshop, programs like illustrator and PageMaker used the space bar for panning. It’s also pan in a lot of other 2D drawing and painting programs. I’m surprised you’ve not come across it before.
Yes, I totally get what you are saying. The difference obviously is that Photoshop defined a standard back in the day. Whereas in many ways, Blender is just another up and coming 3D program among many others who are all just trying to be as good as the current standard.
This is similar to the argument that if control schemes mattered so much, why is ZBrush so popular? The thing is, there was no 3D sculpting before ZBrush and it has stayed consistently ahead of the curve since it was established. People endure it’s weirdness because it provides something they couldn’t easily get with other 3D programs. The same it true for Photoshop too. Although there were other 2D painting and photo retouching programs, no one went as in depth into it as Adobe did.
The interesting thing about this too is that all of the times that I was tempted in the past to really try and use Blender were because it offered something that the other 3D programs were lacking. Smoke sim was the first thing then it was spline based bones and editing and sculpting in pose space or dynamic topology sculpting. All of these things are ahead of the curve and are great reasons to use Blender. Not it’s funky control scheme.
Well, I haven’t dabbled much outside of photoshop and inkscape on the 2d side, so that makes sense why I haven’t run across that control scheme more. And it is a logical control system in the time before middle mouse button. Autocad has a lot of old legacy things like that too. Every button I click in the interface is just typing into the console. Anyway, good conversation, but i’ll stop derailing this thread now.
Blender is highly accessible due to its price point and availability. Its much more likely that any kid or noobie will turn to Blender first to learn 3d.
Blender’s controls and layout are counter to what we would deem not just industry standard but computer standard as well. By learning the controls first and only knowing those controls, not only are habits formed but a way of thinking is formed as well, and this shuts or rather builds a very difficult mental wall/barrier for those going into the professional field of which are expected to use and adapt to the industry standard apps.
Mixing those two together is bad, because it hurts everyone in the long run. It means many new to 3d are building a wall that makes it difficult for them to adapt to other key 3d applications and the industry built around them.
I highly discourage the use of Blender’s default control, but rather advise those who do want to use Blender to pick a Maya or 3DS Max keyset. In fact one of those should be the default.
Additionally, hotkey based workflows are not necessarily the best either, as more visual control schemes work better across multiple platforms and applications. In Blender’s case, its excessive. You literally have to be a keyboard contortionist to use Blender effectively. We are even at the point where developers are running out of room for hotkeys, and to make matters worse each window has its own set of hotkeys making the user experience far more mentally tedious than it needs to be. Blender is a great app, but its still a mess.
So while it may seem normal to you, it becomes a barrier for working with the rest of the 3d applications that make up the larger industry. It cant be better or worse until you have experienced many of the apps and their pipelines, because it helps give context as to what actually is good and what is not. Some people in the past may have thought carrying buckets of water was a good workflow for them, where as until you see the romans with their aquaducts, then you realize the bucket carrying wasnt so great after all.
Thankfully I believe Ton knows this and is slowly leaning in that direction, I could be wrong though. The pie menu integration will be a nice addition, and the reworking of the UI will also hopefully do some good. The best thing they could do however is get rid of the current default keymap and control scheme, leave it as an option but promote something newer and far more compatible to industry standard apps. No more keyboard acrobatics.
Ok, back to derailing…Maybe it’s because I’m still young and spry, but I don’t have that much difficulty adjusting from one program to the next. I’ll go from meshlab to blender to autocad to sketchup to inkscape to excel, all in the course of a day. I’d used blender for 8 years and started using autocad professionally with 2 days of training. There is always a degree of adjustment from program to program. (Hell, half the time I forget how to walk after a long drive). I don’t think that we should be focused on being more like other programs, just for the sake of being more like other programs. Blender is a whole program. It isn’t just a step you take to get to maya or max. We should be focused on making the controls work well. That can mean using other control paradigms that other programs have set up, or it could mean using something new or different. Should we remove mouse pan from the image viewer and replace it with spacebar pan? No, because that’s not an improvement. It’s (apparently) a standard in 2d painting, so why would we want to cripple people who might end up using photoshop instead of blender?I know we’ve talked about hotkeys before, but I will continue to advocate for them. I do agree that the amount of hotkeys that blender relys on really steepens the learning curve. But I think that more advanced users who take advantage of hotkeys will be significantly faster than an equally advanced user mousing around the interface. Why would I want more buttons on the screen? I’ve got 103 buttons right in front of me. I have 1 cursor, but 10 fingers.
In terms of the visual part, William Reynish has posted some interesting mockups in the UI redesign thread in taking the existing UI concept to the next level (and if you don’t know yet, he was the designer of the current interface).
I don’t know if there’s a lot more options exposed as a result, but the idea is to try to remove all unneeded visual noise so as to allow the user to just concentrate on the important things.
The problems with Blenders default controls aren’t necessarily that it doesn’t match other 3D software hotkeys, but that it doesn’t match any conventions for basic user input. Adjust icons on your desktop; select multiple files in your file browser; open Inkscape and select multiple shapes; Launch up a game of Warcraft 3; nearly all software uses the same conventions, not just 3D software:
LMB selects
LMB-Drag does a box select.
Shift+LMB adds to selection.
Ctrl+LMB removes from a selection.
Clicking outside deselects all.
So with most software you don’t need to relearn this basic stuff. With Blender there is always this 2 week to a month extra adaptation period JUST to learn how to select things; not even how to actually model. That isn’t even getting into how those of us that are used to defaults find it completely inefficient to have to hit the b key a toggle a all the time to make even the most basic selections.
So a lot of the time, even in situations that Blender could fit into people’s pipelines, it usually doesn’t. Say you wanted something for retopology and have never used Blender. Is it worth spending an entire month learning Blender in order to retopo, or just buying Topogun for $100 and retopologizing that night? In the professional world where deadlines are a factor, the answer is always going to be the latter.
I was going to say this but didn’t want to get flamed.
People always say blender should change to left click default because other 3d software has left click,but they fail to mention that EVERYTHING has left click default,the internet,microsoft word,you name it.
i’m sorry but this right clicking select business is ridiculous.I just think its some kinda “Pride” thats keeping that from changing.Since the new 2.7x aeries is coming and lots of things are changing (including the interface I think) ,now would be the best time to address that.
