justdrop said: When you tried to say the development was funded by SoE and then changed it to a different argument after they proved you wrong.
You really should read the link vjek posted. EQ was funded by Sony.
Sony Interactive Studios America funded it for the first two years, then shared funding for the last year with Sony Online Entertainment.
But since you don't seem to understand the point I was attempting to make, I understand why you think I 'moved the goalposts'.
vjek said:Savanja said:I mean, we literally have some of those developers working on Pantheon. Making a game now versus then is NOT the same.
Setting aside the client (in this case, Unity), in your opinion, Savanja, which parts of the remaining two primary MMO components ( the server & the persistence layer) should take longer in 2023 than it did in 1997, and why?
Speaking as someone who had to keep the jankiness of EQ running on a daily basis, I can tell you the level of complexity between what EQ was and still is compared to the infrastructure of Pantheon is not even a comparison.
EQ ran on Windows NT Workstation 4.0 boxes. There was no central databases. Everything was done with flat files, including character files. The only thing shared between servers was when they got patched from a central beefed up handbuilt PC which would copy all the game files to a world cluster. Character files were backed up nightly but they never accounted for removing deleted character files which meant the backup continued to grow. That also meant if we ever had to restore an entire world server's characters (which we did several times I was there due to hardware failures) you got all of your deleted characters showing up again on your account from the last time they were backed up. Thankfully, these archaic ways are gone and databases have been embraced for so many useful things in the MMO sphere.
In addition, there was really no separate persistence layer in EQ. It all relied on communication from the zone servers to the world server to save character state. That is why there were so many different ways to dupe items in EQ as the limitations of relying on that connection led to inconsistencies. The first time we used a dedicated persistence server as the authority for game state was with Vanguard. The same goes for a social server.
Comparing development of EQ or any of the first generation of MMO's to anything remotely similar to Pantheon or other current MMO's is disingenious and a non starter. We have learned so much from modern software dev and operational practices that have increased the level of complexity and the time to market.
Artois said: ...Speaking as someone who had to keep the jankiness of EQ running on a daily basis, I can tell you the level of complexity between what EQ was and still is compared to the infrastructure of Pantheon is not even a comparison.
EQ ran on Windows NT Workstation 4.0 boxes. There was no central databases. Everything was done with flat files, including character files. The only thing shared between servers was when they got patched from a central beefed up handbuilt PC which would copy all the game files to a world cluster. Character files were backed up nightly but they never accounted for removing deleted character files which meant the backup continued to grow. That also meant if we ever had to restore an entire world server's characters (which we did several times I was there due to hardware failures) you got all of your deleted characters showing up again on your account from the last time they were backed up. Thankfully, these archaic ways are gone and databases have been embraced for so many useful things in the MMO sphere.
In addition, there was really no separate persistence layer in EQ. It all relied on communication from the zone servers to the world server to save character state. That is why there were so many different ways to dupe items in EQ as the limitations of relying on that connection led to inconsistencies. The first time we used a dedicated persistence server as the authority for game state was with Vanguard. The same goes for a social server.
Comparing development of EQ or any of the first generation of MMO's to anything remotely similar to Pantheon or other current MMO's is disingenious and a non starter. We have learned so much from modern software dev and operational practices that have increased the level of complexity and the time to market.
IMO:
Great info, Artois, thanks for sharing that.
Regarding: [ Setting aside the client (in this case, Unity), in your opinion, Savanja, which parts of the remaining two primary MMO components ( the server & the persistence layer) should take longer in 2023 than it did in 1997, and why? ] It seems like everything you've enumerated would result in Pantheon taking LESS time than EQ1, given you now have have IDEs, Databases with millions of IOPS, hardware with hundreds of times the speed & capacity, pre-built solution frameworks of many kinds, and 25 years of MMO successes to imitate, learn from or build from.
That's why I phrased the question the way I did, because I have seen the hardware and software changes you're referring to over the past 25 years, and now? Everything is faster. Development time, deployment time, everything. The only restriction that's remained the same is the speed of electrons through copper or photons through fibre, and even with those, latency has gone from 250+ms w/ dial-up & EQ1 to far, far less, given the switching time reductions as consumers have gige+ in their homes.
I appreciate the context, yet, the EQ1 team did deliver a complete MMO in less than 3 years. Including everything in the client (which was not the context of my question) and everything not in the client (which is the context of my question). Many teams since 1997 , when EQ1 entered Beta, have designed, implemented and deployed a complete MMO. Some of those teams are 1 or 2 developers, or for a certainty, less than 10. ECO, MnM, Project Gorgon and many, many others. The eqemu project just went open-source. Anyone today, on any GCP, AWS or similar VM or modest hardware, can run their own EQ1 server. There are many other MMOs that offer similar, with client, server, and persistence free or open source.
Yet, this team has had over 3 times that amount of time, and the public disclosure of progress is such that with more people, more tech, and more time, the same milestones cannot be achieved. It's.. extremely disconcerting, to me. I recognize that hindsight provides unique value. I'm also not just throwing shade to see what sticks, I'm trying to find out why months go by with no progress on the server or persistence layers. Objectively, given what YOU know about the server and persistence portions of this project, Artois, that should have been done in less than 2 years, learning from EQ1's mistakes, or lack of tech at the time. Yet.. that's not what has happened.
Can you offer any insight as to why the server and persistence layers of Pantheon have taken so long, compared to EQ1, beyond "operational practices and complexity" ?
vjek: ...that should have been done in less than 2 years...
You cannot say anything like this with any degree of certainty unless you are directly involved with the build of the project itself. That period maybe sensible from an outside view, but unless you know everything about man hours available, equipment, planning and strategy, construction, problems, finances available, etc., etc. then the statement is a wet finger in the air. Personally, I do think it has taken a very long time to get here, but I am not saying that it should have taken much less time, because I just dont know the exact details of the project.
vjek: ...given you now have IDEs,...
Development environments and tools created to make the world of Pantheon are easier and more reliable these days, this is indeed true, however, that ease and efficiency is replaced by the complexity of each individual item created by those systems. Each items interconnectivity with everything else in the game world is much more complex. The IDE maybe less of an issue, but the development cycle switches from battling with an IDE to creating ever more complex game objects; the greater that functionality and complexity, the harder it is to build, integrate and test and the longer it takes to get the item into the world. It is not correct to simply say that the IDE and other systems are better so the project should be easier and quicker to build; it would be if you were building exactly like for like against EQ, but Pantheon's a far more complex beast.
I feel, direct comparison to EQ build time is unbalanced. Pantheon's underlying game play ideals and social aspects are similar to EQ, but that's it. Everything else is nothing like EQ. Landscapes, draw distances, zoneless navigation, NPCs actions, mobs actions, weather, climates, quests dependancies, factions and tons of other stuff I have not listed, are a much more complex than in EQ. I think if you set out to build a like for like EQ world with today's IDEs and tech, you could indeed get it to an alpha state in a couple of years (given the finances and staff), but Pantheon's engine is not like EQ, it could not be done in that kind of time frame.