Resurrecting Softimage From Scratch

General discussion about 3D DCC and other topics
Posts: 4
Joined: 08 Jan 2018, 18:24

Resurrecting Softimage From Scratch

Post by taoros » 09 Jan 2018, 18:28

Hey guys,

I came here to ask a particular question that's been bugging me for a while. I am not a techie, nerdie, coding guy and there is no one in my circle who is like that, so this forum seemed to be the obvious choice.

Here is my question and i'm mostly directing it to Softimage artist-developers:

Hypothetically speaking... if someone had access to millions of dollars for a startup and wanted to re-create Softimage from scratch how much money, how many people, and how much time would be needed to get it to the maturity it was when Autodesk killed it?

How would YOU do it? How many people would you get and what would their qualifications, or specialties, be?

Could it be done by a small team of say 10 people within 3 years? Could you do it with fewer people within that time?

Also, ho many people do you generally need for creating a Photoshop type of software, or a video editing or compositing one, if you want to finish it to maturity within say 3 years?

Your help and responses would be greatly appreciated.

Best regards,

User avatar
Posts: 1551
Joined: 26 Sep 2009, 15:51
Location: Bonn, Germany

Re: Resurrecting Softimage From Scratch

Post by rray » 09 Jan 2018, 18:42

Hi Tim,

It would be more expensive. Autodesk bought Softimage for 35 Mil $, which is probably only a part of the development cost up to that point. I think I read somewhere that they had over 100 people working at the same time on it.

I guess with today's more modern devlopment environments and friendlier APIs all around it could be done with fewer people / much reduced cost. But I don't know much about project development. I think the biggest challenge would be hiring the right people, most of all a reaaaly good core architecture team.

"as a maya discussion grows longer, the probability of a suggestion involving deleting the preferences approaches one"

softimage resources section (formerly known as updated May 30 2018

User avatar
Posts: 396
Joined: 12 Mar 2013, 18:44

Re: Resurrecting Softimage From Scratch

Post by mc_axe » 10 Jan 2018, 00:36

If I had infinite money i would buy it from AD instantly hah! And if they said NO I would buy AD alltogether :ymdevil:

If you had to build it from scratch good and bad thing is that you have a very specific goal, that is to reverse engineering it, but this can be proven even harder compared to going with a new design, and because companies and you the boss dont want to loose time and money, it will inevitably lead to an almost new software that only reminds of Softy. Of course you need an epic team, maybe even some veterans to guide you with priorities, is not a small task at-all. The way i think about it it you dont need to go through all 30 years of development, Summatra XSIs core was developed from 95? to 97? not sure ...And till ~2000s they replaced SI3D with something new, if any1 knows how big was that team i would like to know as well.

Now lookalikes that implement very basic UI with alot of not working buttons and 1-2 basic features can be even a one man job
Remeber a friend did a PS clone that you could only paint with a square then he added brushes and a color picker and i was like whoah!

Posts: 4
Joined: 08 Jan 2018, 18:24

Re: Resurrecting Softimage From Scratch

Post by taoros » 10 Jan 2018, 15:45

Thanks guys, but I'm really hoping that developers who worked on Softimage (or any other similar software) will respond.

But, to some of your points, there are software companies that have 50-100 employees and work on 2 or more software apps that fall under 3D or Photoshop-like software.

And yes, I would hope that "old-timer" Softimage devs would jump on a project like this.

User avatar
Posts: 1279
Joined: 09 Jul 2009, 08:59
Location: Close to The Hague, Netherlands

Re: Resurrecting Softimage From Scratch

Post by Rork » 10 Jan 2018, 17:10

If you want to spend money on a 3d package, my guess would be invest in Blender and accelerate the development there.
It's open source and free, but could use some TLC regarding UI and workflow. Money would really help there get it done more quickly.

We have to accept SI is gone and done, and nothing out there is going to replace it the same way. It's nice to have some philosophical discussion about it, but that's all it is. ;)
SI UI tutorials: Toolbar | Custom Layout | RenderManager View

Posts: 2389
Joined: 24 May 2012, 09:44

Re: Resurrecting Softimage From Scratch

Post by Bullit » 12 Jan 2018, 07:26

I agree with Rork, to invest, go for Blender, even do a fork if it is necessary, maybe you can even Softimagine it...

User avatar
Daniel Brassard
Posts: 827
Joined: 18 Mar 2010, 23:38
Location: St. Thomas, Ontario

Re: Resurrecting Softimage From Scratch

Post by Daniel Brassard » 12 Jan 2018, 15:53

Resurrecting Softimage would be a difficult venture as you would be at risk of infringing in a lot of patents own by AD. As already mentioned, the effort would be better spent in improving Blender (open source) or developing your own solution (expensive and time consuming).

What I would be looking at would be improvement and modernization in the animation side which is still weak or old in many commercial packages like Maya/Max or Houdini. You need a solution similar to Softimage (Non Linear Animation Workflow) that is robust almost unbreakable and expendable with a very good SDK. That alone would put you in the radar to all kind of studios that are struggling with Maya intricacies or Houdini bare bone solution. Something akin of Presto on steroid come to mind.

Then you could expand from there to the other area of the industry (modeling, sculpting, texturing, rendering, etc.) if needed. There are a lot of other solutions out there that do well in the other areas so that would not be my priority. You could make it a fork of Blender or a stand alone solution that can take a scene from another package, animate it and sent it back for rendering.

My priority in development would be as follows:

Phase I
NLA Rigging and Animation Core
Story boarding/Animatic/Visualization
Scene Layout

Phase II (Houdini is strong here so a well developed SDK/Import/Export could be an advantage)
ICE like Core system
Particle and Special Effects
Hair and Fur system

(Maya/Max/Houdini/Modo/Lightwave/C4D/Blender compete in those areas)

Phase III
Rendering c/w pass partition system that can be sent to Arnold/Redshift/Renderman/Mantra/Game Engines etc.

Phase IV
Modeling and Sculpting
Set Dressing

Phase V (The least important, any compositing package would do if you can import/export for post-processing)

Food for thought.
$ifndef "Softimage"
set "Softimage" "true"

User avatar
Posts: 356
Joined: 28 Sep 2011, 10:33

Re: Resurrecting Softimage From Scratch

Post by myara » 13 Jan 2018, 10:43

Modo was developed by a considerable small, but very experimented team that was the core team of Lightwave. I don't know how many they were, but since they had 7 key members of the Lightwave core development team and were like 30 when they merged with The Foundry, I guess the average could be like 15 people ? Anyway, it took them like 3 years to finally release a first version that only had modeling tools. No animation, no render, no nothing else, but it was a very good modeling tool. It took them another few years to introduce some sort of animation without skinning, and a skinning system took a few more.

To get the maturity level that Softimage 2015 had after 18 years of development, it takes time. I don't know how may people were involved in average, but I'm pretty sure they were more than 10. A lot more than 10.

So in short, 10 people in 3 years ? Softimage 2015 level ?
Absolutely Impossible.

You could invest time modifying as much as you can other packages with their respective APIs, and accepting some limits that you will have.
Blender is open source so it should be easier to dig and modify very deep, but Max and Maya are more used in the industry and have more plugins so it could be easier to get projects, and you could even sell your plugins and modifications.
Last edited by myara on 16 Jan 2018, 04:51, edited 2 times in total.
Character Modeler | Softimage Generalist (sort of)

Posts: 4
Joined: 08 Jan 2018, 18:24

Re: Resurrecting Softimage From Scratch

Post by taoros » 13 Jan 2018, 15:59

Thanks for the replies guys, especially Daniel and myara. That's the kind of response I'm looking for, but I'd like more.

I know about Blender, but there is a reason I'm asking specifically about SI. There must be some Softimage devs or devs who are also artists on this board, maybe even "old-timers". I REALLY would like to hear from them, with specifics to my exact questions.

Posts: 108
Joined: 16 Apr 2013, 13:28

Re: Resurrecting Softimage From Scratch

Post by Nox » 14 Jan 2018, 13:22

You should take into account that software is not something that can be made in vacuum by handful of devs working in a basement. It's a product of constant feedback from users, improvement and research. You cannot expect devs to come out after a couple of years and deliver you a finished product.

So, for such complex piece of software as Softimage, besides actual core programmers you would need software testers, devs to write automated tests, UI/UX designer's, team leaders, project managers, IT support, HR, web devs, web publisher, marketing guys, customer support and whatnot. Just assembling full team can easily take a year if not more.

As for image editing software, sure you can do some sort of PS clone in 3 years. But will it be a real competitor offering same features with high level of polish? I don't think so.

Posts: 1201
Joined: 22 Jun 2009, 00:08

Re: Resurrecting Softimage From Scratch

Post by luceric » 14 Jan 2018, 15:51

A team of less than 10 people and three years sounds reasonable, yes. (Bigger teams are just less efficient and don't get any more work done)
But I'm not good at estimation, or more precisely I hate to do it for real.

When you do actual software estimation, you have to do the very frustrating thing of breaking down the work to do in bits, estimate those bits, and add it all up. It's usually very discouraging. There are lots of bits in Softimage, for example the modeling kernel, animation mixing, I.K, property propagation, etc. Stuff I'm not actually familiar about and would vastly under-estimate. It's also likely that those things would be done differently in 2018 than they were with 1998 tech.

1) I think those semi-mythical patents that protect 3D app would be 20 years old and expired.
2) Didn't these modo people also write a renderer? I might be mistaken, but I'm pretty sure that they did and that Modo 101 was a modelling/rendering package. That's like writing two products.

User avatar
Posts: 1710
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia

Re: Resurrecting Softimage From Scratch

Post by Mathaeus » 14 Jan 2018, 16:35

luceric wrote:
14 Jan 2018, 15:51

2) Didn't these modo people also write a renderer? I might be mistaken, but I'm pretty sure that they did and that Modo 101 was a modelling/rendering package. That's like writing two products.
Actually it was Modo 201, around 2006, while they wanted to have renderer from beginning, according to linked article. Modo started in public as modeler and nothing else. As far as I remember, Lightwave people weren't so happy to pay 1K $ or something, to only modeler.
The renderer "turtle" used by this scene, is not currently available. The "turtle" renderer will be used instead.

Posts: 1201
Joined: 22 Jun 2009, 00:08

Re: Resurrecting Softimage From Scratch

Post by luceric » 15 Jan 2018, 02:41

Well ok, it came in Modo 201, one year later. I figure It must have been in the works long before that.

User avatar
Posts: 805
Joined: 09 Oct 2012, 20:48
Skype: ondraise
Location: Colombia

Re: Resurrecting Softimage From Scratch

Post by Draise » 15 Jan 2018, 20:20

I have no experience.. but This.. would be great if someone invested in something new with similar mentality - visual node programming directly in 3D content creation and topline animation and rigging - with UI tweaks that make life easy.

You'd need to build a framework, or adopt one
Also work with the team on milestones
Manage all the other stuff like logistics, payroles, accounting, budgeting, publicity and social media, testers and whatnot.
Havea business strategy to cover costs and the whole investment
I guess pretty much like... a full feature animated movie or game developer business - the model - would be the same. Scrum tasks, nightly/weekly builds, modules and testing, user experimentation and feedback, etc.

Fabric Engine I think had the potential to become a framework for a user friendly modeling/animation/rendering experience, a full on software. I certainly hope that is what they are venturing to but probably not.
Blender sourcecode is decent, ok framework, still evolving, not well commented and a little convoluted concerning code, but easy to work with using Python and C++.
Bforartist also has a sourcecode, which is a bit better commented, and still has compatibility with the code resources used in the Blender community - and they are trying to tackle what I personally don't like of Blender (on top of exisiting with future Blender updates that I'm looking forward to - including a la Softimage style partitions with "Collections" and nice viewport, also visual node programming for particles and object data). So far so good, though it's not a final release yet.
Either branch of Blender you could fork could ultimately - over the next 3 to 5 years of developing your own personal goals - be a good path and shortcut.
There are not many other opensource platforms other than maybe Unreal Engine or Godot as game engines that you could maybe manipulate to be content creators. Odd that game engines are building faster than most content creation....

So.. you could also think about crowdsourcing for financing and development over 5 years development (thus invest more into social media, community and education on top of a playground someone can play with - like Fabric Engine as if it were opensource and targetted not into splicing, but being a 3D authoring package) instead of releasing tiers of products with increased features like Modo has - with particular modules as milestones - or hacking what is already there with a team on payroll through a huge account and investment with clear vision into another platform (UE viewport for modeling, rigging, animation and then rendering anyone!? you can already hack it for texture painting and basic modeling and animation) bending it to your will.

I highly recommended trying out a business model using opensource code like Bforartists, which allows you to hire and raise funds for development, or even try fork Blenders source code with your own team to target your goals - either giving back to the Blender community or splicing in your own vision of what a 3D authoring software should be. There is an addon for sticky keys. I don't know of anyone working on tearable menus. I know that maybe it is possible for someone to develop the "apply to all over active selection" that could be applied. I know there are node visual programming addons already built that could be assimulated into the workflow (Sverchok for parametric procedural modeling or Animation Nodes for procedural animation - and rigging to a degree - using visual programming like ICE) or other systems in development, that you could also just invest in so those developers make it happen. You could be another foundation based of certain goals developing the same sourcecode...

Or go the long route and just make a whole new sourcecode - then maybe boosting it with potential developers by making it also opensource and unlike a foundation, actually hiring those who want in on the opensource for particular milestones.

Making a software development project that requires the top end of the spectrum latest technology developed with enough flexibility to last 20 years and still be relevant and flexible in the present of the future... is a hard model to push for investors and long term contracts to programmers. Specially with a small artistic market, huge risk, large competition, difficult software milestones to develop and the like. Unless you try alternative business models, or go a route that is as modern as say a blockchain technology for a 3D authoring app hiring and paying and generating funds from some kind of programming blockchain developer network to get your milestones and goals realized.... then I might think twice before... trying out a business and developer model that trueSpace, Bryce, Softimage, Lightwave (to a degree) and even the old Blender (before going opensource) have all failed.

Posts: 2389
Joined: 24 May 2012, 09:44

Re: Resurrecting Softimage From Scratch

Post by Bullit » 15 Jan 2018, 21:23

I think to develop software correctly you need first to develop a narrative. Without language to streamline it you can't develop a software of this complexity.

I would start using the road/traffic terminology added with aircraft navigation terminology and logic. We are in space we deal with quantities, front, behind, up , down, speed, stop, acceleration, turbulence, random, all, only, between etc.

So having a dictionary as streamlined as possible would be my start to help conceptually design a software. These words would be all over the software and always mean same.

Posts: 701
Joined: 09 Jun 2009, 20:33

Re: Resurrecting Softimage From Scratch

Post by NNois » 15 Jan 2018, 22:57

Draise wrote:
15 Jan 2018, 20:20
... visual node programming directly in 3D content creation and topline animation and rigging...
Read about Gaffer ;-) That's Open Source too.