🎉 Celebrating 25 Years of GameDev.net! 🎉

Not many can claim 25 years on the Internet! Join us in celebrating this milestone. Learn more about our history, and thank you for being a part of our community!

Game engine companies?

Started by
32 comments, last by Drakon 22 years, 6 months ago
quote: Original post by Simagery
I don''t think middleware companies (the category game engines fall into) are going to move toward the unlimited use pricing model, it''s just not fiscally smart.


But in fact NDL does have such a pricing option for their NetImmerse engine, . And they encourage licensees to choose this option. (At least they encouraged us to choose it.) They are profitable. And they sell pricey support agreements each year to supplement new license revenues. (I believe we did or will soon sign a purchase order to renew our NetImmerse support, but we didn''t buy the unlimited use license.)

quote: Original post by Simagery
This model seems to work as Criterion is far and away the most successful (outside of the genre specific boys: Quake and Unreal).


Are they profitable? (That is one measure of success. And because of my interactions with certain companies, I know that not all game industry middleware companies are profitable. I won''t say which ones I''m aware of since that info was given in confidence.)

Graham Rhodes
Senior Scientist
Applied Research Associates, Inc.
Graham Rhodes Moderator, Math & Physics forum @ gamedev.net
Advertisement
Yes, Criterion is profitable.

Yes, I know NetImmerse does use the unlimited use model. But I still don''t think it is fiscally smart. Look at it this way: wouldn''t you rather purchase middleware (which is supposed to make your life easier) that you could eventually get good enough at that you don''t have to rely on support? I would assume that most serious PC game developers have reached a point where they wouldn''t think it necessary to pay $50K for a year''s OpenGL/DirectX support. Do you really want to buy a product that you''re going to continually have to rely on support for for years to come?

I''ll admit that there''s a certain comfort in having that support, and I''m sure it includes updates and such. But I think a product should be good enough to sell, not good enough to give away and problem-ridden enough to only sell support.

Now, I realize this is a gross generalization, just my personal opinion. If I was cutting corners on a budget, I would have a tough time paying for support on a middleware product if I felt I was familiar with the product (like DirectX or OpenGL). Currently, this is a no-brainer because middleware is very new. But ten years from now you''d hope that it wasn''t still changing so much as to warrant yearly support contracts.

I''m not saying that the unlimited use model + support fees is not a great idea... I personally would prefer the option for myself. I just think it is not fiscally smart, i.e. it''s not good for the company''s bottom line (and in my mind for the perception of their product).

How do you find NDL? Are you using it only on the PC or are you using it as a multi-platform SDK?
Simagery, as you mention normally support includes all the updates that come out during the duration of your license, so it seems to me this in itself would be worth it, aside from any other technical support you would get.

R.
_________________________The Idea Foundry
Correct, but I want to be clear that I actually prefer the unlimited license from a developer''s perspective. But from a business, bottom-line, may the stock holders happy perspective, I believe that serious middleware solutions will be on a per title (per platform) basis.

But then again, I''m not a sales/business person, so I don''t really concern myself with those issues. (And here I am in the business forum sticking my nose where it don''t belong!)
To comment on Simagery''s reply to my reply to his reply (what?):

I actually agree with your opinion that the unlimited use pricing model is not fiscally smart, for the reasons you mention. At least not in *this* industry, in the long term. (Short term it is fine.) I think the long term thinking behind unlimited use engines is based on the idea that eventually game studios who use the product will not need any 3D graphics programmers on staff---only artists---thus avoiding the possibility that a studio could use and tweak the engine internally without paying for support. I''m not sure that''ll ever happen in this industry, although it is an appealing concept for some reasons (saving money by removing staff who were rendered unnecessary by using good middleware). At least for now, even if you license an engine you still need 3D programmers who are smart enough to dig into engine source code since there are always issues. And so support from the engine company is not enough.

As Tacit mentioned, support fees also pay for engine upgrades and these are always nice. For example, NetImmerse 3.1 had a decent character animation subsystem, but NetImmerse 4.0 has a MUCH improved character animation subsystem. Not only does it manage motion-blended animation better, but it also supports hardware vertex shaders for hardware skinning. And the upcoming NetImmerse 4.1 will apparently have more improvements (e.g., possibly making it easier to customize the way pelvis motion is blended for a particular game). Artist tools are improving version after version. They are also exposing hardware shaders in the new versions, so the engine is keeping up (with reasonable lag) with hardware and rendering technology advances.

As for NDL, we''ve been very pleased. NetImmerse is quite capable, although we haven''t really exploited it fully in our work yet. We use the PC version only. Their source code is very very solid and clean. (Although since they support PS/2, PC, Xbox, and Gamecube the code isn''t as efficient and clean as it could be.) Their support is as good as we''ve experienced anywhere, but it isn''t perfect. Since the people actively developing the engine are also the sole support staff, the answers tend to be a bit short and sometimes lacking technical details. Questions must be fairly infrequent or they would never progress the engine. On the other hand, when we can prove that there is a bug, they fix it damned quick. (Sometimes we give them the fix.) The longest time we''ve waited for a bug fix was perhaps less than a week. (There have been a couple of exceptions---things that would require changing the engine architecture just don''t really happen at all for a few versions.)

Graham Rhodes
Senior Scientist
Applied Research Associates, Inc.
Graham Rhodes Moderator, Math & Physics forum @ gamedev.net
By the way, NetImmerse is a privately held company. I don''t know if they have an employee stock ownership plan or incentive stock option plan. I know this is off topic, .

Graham Rhodes
Senior Scientist
Applied Research Associates, Inc.
Graham Rhodes Moderator, Math & Physics forum @ gamedev.net
The other thing about the unlimited use model is that, realistically, if you''re a game development company, how many games will you really make using the same engine? If we factor in an average development cycle of 2 years, by the time you''ve finished your first project, the engine is already somewhat out of date. Assuming you still go ahead with the same engine for a second title (hey, it''s been done), by the time that one is finished you have in your possession a very advanced 4 year old engine that just won''t pass muster for the latest games. So...the solution: either pay up and get another ''unlimited'' license or by a new, limited license. Either way they make good money off you. And that''s not even counting the maintenance fees.

Just one idea.

R.
_________________________The Idea Foundry
quote: Original post by Tacit
The other thing about the unlimited use model is that, realistically, if you''re a game development company, how many games will you really make using the same engine?


Yes, absolutely. There''s multiple goals with the unlimited use model, and another one is to try and get multiples of the single use license fee, anticipating that a given studio won''t use the game for more than one or two games.

quote: Original post by Tacit If we factor in an average development cycle of 2 years, by the time you''ve finished your first project, the engine is already somewhat out of date. Assuming you still go ahead with the same engine for a second title (hey, it''s been done), by the time that one is finished you have in your possession a very advanced 4 year old engine that just won''t pass muster for the latest games.

BUT, the annual support model does manage to avoid this situation to a significant degree. NetImmerse 3.1 for example supported OpenGL and DirectX 7, using the fixed function pipeline for T&L. NetImmerse 4.0, which came out basically around January 2001, supports DX8 and the programmable hardware pipeline. There was an interim update around April I think that added Gamecube support. Plus, they do add updated algorithms (improved inverse kinematics for example), drop outdated algorithms and features (software only Bezier patch geometry for example), etc., in a darned good attempt to keep up with the state-of-the-art. It ain''t perfect, and it isn''t as up-to-date as a studio with a Sweeney or Carmack can manage, but its a hell of a lot rosier than the picture you painted, , .

You would never need to buy a second unlimited use license. The annual support (which is at least an order of magnitude less than an unlimited use license) keeps your engine as up-to-date as it gets from that engine vendor.


Graham Rhodes
Senior Scientist
Applied Research Associates, Inc.
Graham Rhodes Moderator, Math & Physics forum @ gamedev.net
Righto. Well, thanks for chatting about it. Been interesting.

R.
_________________________The Idea Foundry
BTW Graham, do you fall into the Physics stream at ARA or engineering?
_________________________The Idea Foundry

This topic is closed to new replies.

Advertisement