Maya has been bundled with Mental Ray for years now, but why would they remove the scanline renderer? There’s still many things where it is still useful. Ask yourself this, in a toolbox, unless you need to put something in that doesn’t fit, why would you want to discard tools from it? ;D
That is pretty much my reasoning, but i keep hearing people say that BI will die out. Having both would be perfect in my opinion, especially if there was a way to work with both on the same scene.
I think it will die out in the sense of no real development will go into the blender internal (besides some bug fixing, probably. and maybe someone in the community will keep adding patches, if needed).
but i also believe, that it will be kept around. doesn’t make sense to remove it. but at some point, cycles will become the default engine (once all features are implemented and it’s as stable and fast as it can be).
I don’t know if development will die out, but there is little need to develop BI further, but bugs and such will still get mended I think… And as a simple example of something that is mote fitting to be rendered in BI than Cycles, just look at motion graphics or you were to design a HUD in 2D, you wouldn’t wanna mess around with Cycles doing that, faster simpler doing stuff like that in BI… 
Also, the better you are at creating good looking things using a crappy renderer, the better you will be using a good one.
the best way would be cleanup the BI, I mean delete useless and buggy stuff.
Ask yourself this, in a toolbox, unless you need to put something in that doesn’t fit, why would you want to discard tools from it? ;D
The analogy almost works. Except that in this case, if the tool (BI) is still there it will get bug reports, feature requests, need to be kept up to date with other changes in other tools in the kit… thus becoming an aging tool that requires it’s own maintenance.
Perhaps a car garage is a better analogy. The garage has many useful vehicles but a very limited number of mechanics (developers). You have old reliable Bertha in the back corner and a new hot rod that is fun to tinker with but only half built. Once the hot rod is fully built and faster, more reliable/predictable, nicer to drive, etc… do you keep your overworked mechanics tweaking old Bertha?
Durian and the Render Branch was an attempt to give Ber… Blender Internal an overhaul, tweak it further, make it more modern but it proved to be too much of a task to be worth it with the Blender Foundation’s current resources and the general state/structure of the renderer. So the plan is to eventually replace the BI with Cycles as I understand it. Not tomorrow by any means, but once it is feasible that the majority of users can do the majority of things they used to with BI at an acceptable level of speed and flexibility. Which won’t be tomorrow by any means.
PS. If a developer reads that and I’m completely wrong, please correct me. However it seems this recurring issue comes down to a matter of what is worth the allocation of time and effort to maintain.
Doesn’t the blender game engine use BI as its graphics engine?
Don’t quote me on this, but I’m pretty sure that the BGE renderer is waaay simplified so it can work faster. Also, the BGE is not useful for rendering, at all. It’s useful for physics animations. falling stuff, exploding stuff, destroying stuff, smashing stuff to pieces, and other fun stuff 
When Cycles first came out I thought I would probably carry on using BI simply because my graphics card does not support GPU rendering so I do not get the full benefit of the new renderer. But gradually I find myself using Cycles more and more because a lot of the new tuts use it. For the moment though, out of the two, I still prefer BI. I`m sure that would change if I bought a new laptop with CUDA though.
Hello peeps,
this is very interesting…
the question i’ve got is…
how much effort is there in maintaining blender internal…
Are new features being added to it
is it in the way…
…
i belive that if blender internal is feature rich and no longer requires mainence then wats the problem with keeping it in blender…
some poeple may wish to combines renders from blender internal and cycles as they both have there perks…
correct me if im wrong but inst freestyle also a render engine…
i think that users should have access to all render engines-----consumer choice
well, for starters, unlike external renderers cycles nodes clash with BI… this is sad because it makes switching back and forth a pain. most external renderers can switch back and forth without problem…
there are weird artefacts with shadow maps, strange “quad stairstepping” with shadows at terminators, durian’s raytraced indirect lighting never made it to trunk (only approximate), animation glitching with “halo” spotlight volumes…dreadfully slow image based lighting (that requires raytracing…)
so in short, blender internal is feature rich but sadly does require maintenence and bug fixing.
That said, BI is currently a way better choice than cycles for animation IMO. I hope that it hangs around for many years. Cycles cannot ever be all things to all people.
Okay then, consider this:
As of Blender X.Y.Z, it will no longer be possible to render any of your (perhaps thousands of…) existing Blender files. We’ve taken that capability out of the product, denying you ever having the use of it, because we think it sucks. :no: Sorry about your projects, and maybe, your entire business.
Ain’t gonna happen. Because, if it did, you would immediately have a fairly large group of people who would no longer (be able to…) use any version of Blender for their purposes. Face it, it simply isn’t your place to tell someone who’s been using a tool for decades that they just can’t do that anymore because you’re tired of the technology.
And then there is this point of view:
Sure, Cycles is great, say, for producing realistic full-room lighting in a hurry. But, sometimes “actual reality” is not what you want: you want “realistic,” and sometimes you want “pure fantasy, but properly exposed.” Cycles compels you to start looking for the real-world lighting tricks that real-world photographers have to deal with every day. And this is CG, not the real world.
A workable strategy is obvious: combine the two techniques. Make the GPU-rendering wizardry an adjunct to what you are doing now. Use a rapidly-generated overall solution in Cycles and then use BI to generate specific aspects of the shot. Then, bring the two together in one compositing-based work flow. I want to, say, produce a Cycles solution and at the same time inject the effect of a shadow-only spotlight. I want faster, please, but not real-time. Given that I have this nifty-fifty piece of silicon in my computer, I want to use it in many different ways. I want to leverage the GPU to get me results faster, without being strictly constrained by what my GPU is or is not capable of doing.
I want to see these two methodologies become very tightly integrated; not to have one usurp the other, and also not to have it work in an entirely “different” way. Perhaps in addition to Cycles, I still want GPU-accelerated mathematical render nodes, using the GPU strictly as a massively-parallel matrix processing engine.
Erm, NO! 
The BGE uses “Open GL” (Open Graphics Library) for rendering real time graphics. Loosely said, there is nothing more possible, than the viewport rendering. There are several functions / features, that let you make look your game pretty slick, though. That’s e.g. depth of field, motion blur, 2D filters and custom shaders.
so do the hundreds of millions of people who enjoy Disney/Pixar’s films.
no it doesn’t
Thanks. With that answer frankly if anyone was taking bets I would bet against BI. The vast majority of users are either working on their own or in small companies. I think the will gravitate towards an engine that is easier to use. Path tracers are slow but the are also easy to use and learn.
Take a simple light there is hardly any settings to adjust I wouldn’t say the same thing about a light using shadow maps for example and this thing kind of thing sort of carries on in every nook and cranny of biased engines settings to tweak and adjust everywhere.
Biased engines are fast but the also demand a lot in terms of adjusting settings and tweaking things here and there things like photon maps can take time to tweak and set in cycles you adjust bounces and that’s about it.
About the most complicated thing in cycles is the node based materials. Just take a look at BI closely every time you use anything that does not involve ray tracing watch how the number of things you have to set and adjust increases.
I’ll add to tyrant_monkey’s comment this http://wiki.blender.org/index.php/Dev:2.6/Source/Render/Cycles/DesignGoals
Please stop this nonsense with the die-out crap ! A little information never hurt anyone before posting on forums !
Just read through here http://wiki.blender.org/index.php/Dev:2.6/Source/Render/Cycles and try to draw your own conclusions.
Also don’t forget the PLANS AND IDEAS at the bottom.
There is nobody working on Blender Internal right now and there likely won’t be anyone working on it in the future. It’s as simple as that. BI won’t be removed as long as Cycles is in such an early stage, but it might get removed if Cycles is mature and maintaining BI ends up taking up too much time.
Face it, it simply isn’t your place to tell someone who’s been using a tool for decades that they just can’t do that anymore because you’re tired of the technology
Every blender release within the last decade is archived, so neither will anyone not be able to render their files anymore, nor need these people stop using blender. But you can’t really expect the developers to deal with this decade-old crap forever. Simply being a user of a free application doesn’t entitle you to anything.
It’s called getting with the times. With every advancement in technology, there are always people who are against the change. But in the end the change happens anyway, and we all benefit. Like I said, if you had ever used something like Arnold before you wouldn’t even care about BI, you’d be begging Ton to hire 4 more coders for Cycles. Physically-based does NOT mean physically accurate. There’s plenty of room to mess around with a more toon-y look, as well as lots of other kinds of artistic expressions thanks to the versatility of things like OSL (which I’m praying gets implemented in Cycles, even if just for CPU) or the existing nodes BSDF system. All physically-based means is that light and materials, under normal parameters, will act in an intuitive manner for the artist, instead of spending hours tweaking and test rendering just to get one material correct under one lighting situation.
BI still has many advantages compared to Cycles currently. I think as long as Cycles cannot do the same (speed) BI is for certain projects simply a better option. Not always is a path-tracer needed.
I love having both.