What if Softimage died?

[angus_davidson]

I could live with that strategy, given one thing:
the SDK HAS to be opened further, as consistently as possible, including bugfixes.
Because it should be possible for 3rd parties to patch on the features that AD itself will not care for anymore. Good compromise, actually!
Even better: devs wouldn't take the risk of writing something that will come from AD itself, without knowing.

We would have to learn, though, what that focus is going to be. Makes me really curious where you guys see Softimage fit in officially...

But PLEASE open all the doors for customizability, and don't cripple anything intentionally or by lack of (a little) action!

While I agree with your suggestion 100% its probably the hardest thing to do. Part of the problem with deep SDK integration is the way that Softimage has been put together over the years. If you dont plan for that at the start (as was the case with Maya) its something that is very difficult to expose later on. Luceric is probably the better person to speak to this though. I last worked actively with COM based stuff a decade ago. It may have become less of a Bitch to work with of late.

[Eugen]

True, that was maybe the biggest failure architecture-wise. The SDK was added too late, ans wasn't part of the initial layout.
There's still no direct way from the outside to manipulate the scene graph (complex a task or not), at least not deeply enough.

This, and maybe the core dependencies on legacy Microsoft technology, is what makes Softimage look old.
Darn... it's so brilliant otherwise...


Think about it, before everything quickly boils down to just new ICE nodes...

ICE is a kind of visual operator SDK - which is fine, but besides that, some prefer or need using good old lines of code - in the SDK.
ICE is somewhat patched on, it's not a full SDK replacement, but a complement. It only lives INSIDE of an operator.
The SDK should be there to handle the operator graph - make or break port connections as needed.

I'd love to see AD bite the bullet NOW and improve the SDK... as a gift to it's faithful userbase or whatever... then they can do what they want, retire, or throw new ICE nodes together all day long.
It's doable, and Luc-Eric mentioned that there were plans/thoughts to do so in the past.


There are other places to dock new stuff, too - custom primitives, tools SDK, ICE SDK, PyQt for flexible UI development, shaders, rendering API... should be part of the strategy.

[angus_davidson]

I would love that but I think at this stage its pretty unlikely. I think getting them to focus on one thing and putting in proper deep development instead of just tacking something on would be the biggest win we could get. One of the things I would love would be the ability to skin and or customize the UI to a much greater degree. However I doubt its very high on the list compared say a better viewport. Soft has always had this massive potential. Suppose it just irks me when it isnt always realised. Kinda like when we get a student who has ability and talent in abundance and and just skims through doing the bare minimum.

However I think this type of discussion is very positive. allows the folks watching to get some idea of what we need.

[SreckoM]

ok, interesting poll, but let me throw a couple of possible hypothetical curve balls in there.

What if Softimage continued as normal, but it was channelled and focused into being a more specialised package for something. Kinda in the same light as something like a Mudbox or Motionbuilder, which are good at doing particular things.

And before people start saying - 'yes but Softimage can do loads of stuff, etc, etc'.......we know that, but lets just say that you're faced with two stark choices, either the product has to die, or it continues but with this particular focus. I'm not saying what the focus might be and it might not be to everyone's liking, but it would suit Softimage and it would mean the product lived on.

Btw, just to be clear here, I'm not trying to conduct a secret poll/survey for Autodesk, I'm just throwing something out there in the light of this thread.

Lower the price, include it in rental (lower price than Maya or 3dsMax), continue fixing bugs (better Linux support), and let me know where to pay And if it is ICE that will be developed further, even better.

[Eugen]

Regarding "unlikeliness"...
After all, the Rendering API was changed under AD jurisdiction.
Unless development is throttled down to an absolute crawl, why shouldn't the same be possible again, with the operator SDK?

@ Bellsey:
If you guys want to focus on anything, than do so mainly on the 'core' of Softimage, meaning simply the parts nobody can reach from the outside.

Doing ICE nodes, however complex, would strategically be the most irrelevant thing to do for you, because others can do so, too - and are actually doing it all the time.

Negative example: CrowdFX (no disrespect to the achievement!). People could use Creation-Horde instead. Some company created their own crowd system as well, although in-house, but the point is, it's possible, and it's happening.

Positive example: operator stack.
THAT's a neuralgic spot for you to start with, and it can only be done by you.
Lots of improvements could be made (and lots of requests have been placed) in the way the scene graph is presented/manipulated, via UI or via scripting/coding.

A thought:
Imagine the operator SDK would fully cater everything that is needed for scene graph manipulation - 3rd parties could theoretically even write a custom scene explorer, maybe even a graphical one - Maya style...
It would mean exposing and documenting much (clusters and plenty other nodes behind the curtain I guess), but what keeps you from doing so?

[Lord Futzi Voldemort]

ok, interesting poll, but let me throw a couple of possible hypothetical curve balls in there.

What if Softimage continued as normal, but it was channelled and focused into being a more specialised package for something. Kinda in the same light as something like a Mudbox or Motionbuilder, which are good at doing particular things.

And before people start saying - 'yes but Softimage can do loads of stuff, etc, etc'.......we know that, but lets just say that you're faced with two stark choices, either the product has to die, or it continues but with this particular focus. I'm not saying what the focus might be and it might not be to everyone's liking, but it would suit Softimage and it would mean the product lived on.

Btw, just to be clear here, I'm not trying to conduct a secret poll/survey for Autodesk, I'm just throwing something out there in the light of this thread.

I don't think there isn't even the need to define a specialization. (On the contrary: Being an all-round tool is the "specialization" of SI already.)
How would a focus help to continue Si as opposed to just continue it? Again: the focus, if any, should be fixes, fixes, fixes, and that at a lower price. It hasn't to die at all only because it doesn't get new features.
On the other hand, a clear commitment would encourage third party devs to continue develop addons, render integration and the like for it. And if the SDK doesn't change every year all the better.

[Eugen]

And if the SDK doesn't change every year all the better.

Could you please be more precise here, so we can avoid misunderstandings? Thanks!

[Lord Futzi Voldemort]

And if the SDK doesn't change every year all the better.

Could you please be more precise here, so we can avoid misunderstandings? Thanks!

I mean it would be less hassle for 3rd party developer to maintain addons, if they wouldn't need to adapt to changed SDKs very often. Not that this is the case right now with Si ;-)

[angus_davidson]

And if the SDK doesn't change every year all the better.

Could you please be more precise here, so we can avoid misunderstandings? Thanks!

If there is one worse thing then not being able to reach the nuts and bolts its having them change all the time and not being compatible from version to version.

[Eugen]

I know. The SDK consists of many parts, as I have listed.
You might be referring to the changes in the Rendering API, which delayed VRay quite a bit, for example.
Yet, these changes were necessary, and we can be grateful that they happened.

My point is, there are many places where the SDK need changes - meaning improvements. It does not necessarily mean breaking plugin compatibility, though. Depends.
We like to leave that update-hell to the 3ds Max folks...

Yet if necessary, ideally they should happen with one version increment, and not Salami-wise, of course, that's true.

[Lord Futzi Voldemort]

I know. The SDK consists of many parts, as I have listed.
You might be referring to the changes in the Rendering API, which delayed VRay quite a bit, for example.
Yet, these changes were necessary, and we can be grateful that they happened.

My point is, there are many places where the SDK need changes - meaning improvements. It does not necessarily mean breaking plugin compatibility, though. Depends.
We like to leave that update-hell to the 3ds Max folks...

Yet if necessary, ideally they should happen with one version increment, and not Salami-wise, of course, that's true.

I completely agree, but in a feature frozen SI version, those changes could be made when 3rd parties demand it.

[luceric]

It's not really important, but the rendering API sdk were worked on at Avid, but eventually finished and shipped in Autodesk. This sdk fantasy is bit a waste of time, imho. the facts are that Maya and Max are already great SDK platform and you can build all the tools you need in there. You could have built your NURBS tools in Maya 15 years ago. You could potentially change the UI of maya to be like XSI if that's what you wanted. I believe the third party developers for XSI will evaporate within two years, because the user base is evaporating, including the big japan studios who are in the process of switching to Maya for the next gen consoles. There are quite a few places that have one or two seats of XSI, but are they still on subscription, do they want to spend money on upgrades, and buying plugins - you'd have to ask the third parties. I've heard horror stories of third parties making plug-ins for XSI and then being really angry that they ended up having lots of mentions and demo download but no sales. In any case you may or may not agree with any of this, but if a team was sent to work on the XSI SDK, then lets say there are great enhancements in two years, is there anyone in two years that will want to start to develop for these new APIs no really knowing how many users will actually have this version of XSI - I don't think so.

[Eugen]

I completely agree, but in a feature frozen SI version, those changes could be made when 3rd parties demand it.

... which is just about now!
*clearingthroat* I demand the operator SDK gets improved!





Hm.




Anyone listening? ; )

I requested that SDK stuff already in the last beta anyway. Let's hope there's some wise decision making this time at AD.

[Letterbox]

It's not really important, but the rendering API sdk were worked on at Avid, but eventually finished and shipped in Autodesk. This sdk fantasy is bit a waste of time, imho. the facts are that Maya and Max are already great SDK platform and you can build all the tools you need in there. You could have built your NURBS tools in Maya 15 years ago. You could potentially change the UI of maya to be like XSI if that's what you wanted. I believe the third party developers for XSI will evaporate within two years, because the user base is evaporating, including the big japan studios who are in the process of switching to Maya for the next gen consoles. There are quite a few places that have one or two seats of XSI, but are they still on subscription, do they want to spend money on upgrades, and buying plugins - you'd have to ask the third parties. I've heard horror stories of third parties making plug-ins for XSI and then being really angry that they ended up having lots of mentions and demo download but no sales. In any case you may or may not agree with any of this, but if a team was sent to work on the XSI SDK, then lets say there are great enhancements in two years, is there anyone in two years that will want to start to develop for these new APIs no really knowing how many users will actually have this version of XSI - I don't think so.

Really that's a fatuous argument, because it's situation of autodesks own creation. If Japanese studios are moving to maya, and now there are less 3RD party developers, it's not like autodesk couldn't see that coming...and it's not like their actions haven't caused it.

If Maya never had it's extension packs, if it never had a DX 11 viewport, if it had it's best support staff fired, if it never had it's ui redesigned, if it had great developers let go - or moved onto other projects, if it's development was handed over to newbies - irrespective of how good they are, if it had NO marketing, if 3rd party developers then because of those actions abandoned it, (and the list goes on ) .... Then...

Then you'd be saying Maya is dying wouldn't you, and you'd be right. So the reason(s) for SI's demise, is clearly in autodesk's hands, it's not like autodesk didn't know what it was doing, and couldn't predict what the outcome would be.

Maybe autodesk will do the same to Maya or Max or anything else at some date, if it suits, as with SI.

Considering that m&e is really a side line for autodesk, if it shut down wholesale not one institutional investor would bleat. I'm sure Bass could spin it as... "We're going back to fundamentals, cad. M&E it's not autodesk's core business".

Whereas for the Foundry and SideFX it IS their core business. So they have significantly less interest in canning any products now or in the future, all they'll do is improve on them. So any "Users" investment and skill sets is better served in learning product from those companies.

Given all of the above... why switch to another autodesk product - Why take the risk?

[SreckoM]

Unless that was AD idea from beginning. Buy direct competitor, take developers, tech, patents?, and just let it die eventually. And invest into Maya, because Maya is easier to extend, larger userbase or whatever reason...

[Nizar]

Given all of the above... why switch to another autodesk product - Why take the risk?

Amen

SideFX announced Houdini 13, the point release was a good release, so a major release can be potentially really good.

Also next year Modo 801 will be the first release developed totally under Foundry, so I expect major improvement in workflow and consistency (and rendering is just quite good).

Blender going better every release.

No reason for give money to ADSK