none of BioWare's engines can handle the number of connections needed for an MMO
This is obvious - those engines were written with 8 and 64 simultaneous players respectively in mind, and most times the target number was considerably less than that. However, the Infinity engine was also written to host that same game on a Pentium 200MHz (if you were lucky!) and there was no dedicated server app. The Aurora engine had a dedicated server app, but if you look at its interface (at least the one originally supplied by Bioware) it was really intended for a small group to play a single module at a time if they happened to have a spare box sitting around rather than running a full-fledged Persistent World. Having said that, the NWN PW server I occasionally play on often has upwards of 60 simultaneous users with no noticeable lag, and that hardware is not enterprise-grade by any means.
So perhaps my "minor tweaks" comment should be re-written as "write a new network layer with optimizations for dedicated hardware and massive-scale multiple connections" but this could likely be done on a coffee break or two. ;)
Now, as for your idea of HeroEngine being exclusively a graphics/database/networking engine, I would refer you to their website where among other things is mentioned their "revolutionary HeroBlade™ tool" which "allows for large teams to collaborate in real-time" and "give your GameMasters the power to ... run creatures, become NPCs and even orchestrate events in a live game", their "Next Gen Windows Client", and "HeroScript™, a proprietary language built from the ground up for the unique challenges of MMO development".
Of course they are promising "complete creative control", but how could they? The client is pre-written (and Windows-only), which limits the UI to a particular style of gameplay. For example, consider the UI differences between DDO and Eve Online - both MMORPGs, but one very fast-paced, visceral and action-oriented and the other more slow-paced, strategic and cerebral. The toolkit for both content management and interaction - supposedly based on the game client - is also pre-written, which limits the ways your support team and GMs can interact with the game environment. Lastly, the scripting language is pre-defined, and anyone familiar with Aurora modding can testify to the enormous capabilities that were opened up every time Bioware came out with a new patch to NWN with just a few new scripting commands. It is possible that HeroScript™ can be easily extended with new functions, but it will still limit the developer to a certain mode of thinking that may not be a good fit for the game in question.
Aurora had some issues that had to be worked around to run a PW, yes - but that was largely because Aurora was not intended as an engine for a PW. The community did fantastic work with hacking together database bindings and dedicated server improvements, not to speak of the actual custom content added to the game itself. Bioware was late in introducing the server-hopping capabilities that CoPaP, among others, now make use of to mitigate the 64-simul-players issue. If server-hopping and database bindings had been present from the start, there's no telling where Aurora and NWN would have gone in the MMO/PW space, or how that would have affected the industry.
That's the sort of revolution I was hoping for from Bioware with this new project, and which I suspect will not happen partly because of this 3rd-party lock-in process. I hope they prove me wrong.
Re: Bioware takes the shortcut
This is obvious - those engines were written with 8 and 64 simultaneous players respectively in mind, and most times the target number was considerably less than that. However, the Infinity engine was also written to host that same game on a Pentium 200MHz (if you were lucky!) and there was no dedicated server app. The Aurora engine had a dedicated server app, but if you look at its interface (at least the one originally supplied by Bioware) it was really intended for a small group to play a single module at a time if they happened to have a spare box sitting around rather than running a full-fledged Persistent World. Having said that, the NWN PW server I occasionally play on often has upwards of 60 simultaneous users with no noticeable lag, and that hardware is not enterprise-grade by any means.
So perhaps my "minor tweaks" comment should be re-written as "write a new network layer with optimizations for dedicated hardware and massive-scale multiple connections" but this could likely be done on a coffee break or two. ;)
Now, as for your idea of HeroEngine being exclusively a graphics/database/networking engine, I would refer you to their website where among other things is mentioned their "revolutionary HeroBlade™ tool" which "allows for large teams to collaborate in real-time" and "give your GameMasters the power to ... run creatures, become NPCs and even orchestrate events in a live game", their "Next Gen Windows Client", and "HeroScript™, a proprietary language built from the ground up for the unique challenges of MMO development".
Of course they are promising "complete creative control", but how could they? The client is pre-written (and Windows-only), which limits the UI to a particular style of gameplay. For example, consider the UI differences between DDO and Eve Online - both MMORPGs, but one very fast-paced, visceral and action-oriented and the other more slow-paced, strategic and cerebral. The toolkit for both content management and interaction - supposedly based on the game client - is also pre-written, which limits the ways your support team and GMs can interact with the game environment. Lastly, the scripting language is pre-defined, and anyone familiar with Aurora modding can testify to the enormous capabilities that were opened up every time Bioware came out with a new patch to NWN with just a few new scripting commands. It is possible that HeroScript™ can be easily extended with new functions, but it will still limit the developer to a certain mode of thinking that may not be a good fit for the game in question.
Aurora had some issues that had to be worked around to run a PW, yes - but that was largely because Aurora was not intended as an engine for a PW. The community did fantastic work with hacking together database bindings and dedicated server improvements, not to speak of the actual custom content added to the game itself. Bioware was late in introducing the server-hopping capabilities that CoPaP, among others, now make use of to mitigate the 64-simul-players issue. If server-hopping and database bindings had been present from the start, there's no telling where Aurora and NWN would have gone in the MMO/PW space, or how that would have affected the industry.
That's the sort of revolution I was hoping for from Bioware with this new project, and which I suspect will not happen partly because of this 3rd-party lock-in process. I hope they prove me wrong.