Well quite a large jump on the roadmap this week indeed, was expecting FPS AI as it was progressing well but seems it's not going to be done in time.
@Octagon7711 optimization as it's a task for every update is really the main focus now that the update is on testing and feature work is being complete, but as it was announced, the new release format is date-driven, this means that indeed if the patch is needing more optimization by the end of the month, it should still release.
Then they'll make up for it on smaller hotfix patches, 3.2.1 and so; there's really no better way to prevent the schedule to get off rails.
1. What we have seen of some of the planning makes it clear that there are some "common elements" of activity that feed multiple e.g. ship work. You can see this if you go back to some of the "more detailed" plans.
So internally e.g. "do widget X" is a precursor to do "ship 1 activity", "ship 2 activity", "ship 3 activity" etc. And this may be what they are tracking internally.
Do widget X though may well be something that, in isolation, no one cares about. What they want to know about are the ships (say). So what we get in the planning is "do X and ship activity 1", "do X and ship activity 2", "do X and ship activity 3".
If widget X is the last thing to be completed however all the ships (say) will be completed at the same time. Even though all that has happened internally is that they have finished widget X!
2. Partial progress. Always a bugbear. How do you account for work that is partially complete. The most common way - and one that is being used - is to break down top level tasks into sub-tasks. Hence the 11/39 and 23/48 type of numbers. And this does indeed give a better idea of how far things are along.
However you still have partial progress - on the unfinished sub-tasks. So 11/39 means 11 finished and 23 in some partial state of completion. And its always a case of how much effort do you invest in project management to capture that partial progress. Especially as they have broken the work down into what they believe they can deliver in a quarter.
If the sub-tasks - the 39s and 48s and so forth - are simply next level down activities - then 11/39 may be pretty close to complete! If extra effort has been applied to create "short duration" sequential sub-tasks then 11/39 may only be about half done. We simply don;t know - they should of course. And that is the information they will have when they decide to keep a task in the current patch or slip it to the next.
As a result big jumps in %complete don't surprise me. Just a case of having to focus on what they actually deliver in the 3 month window rather than the intervening weeks. And also do they deliver stuff that they slip into the next patch. Missing something by the odd week - it happens; if they slip something though and fail to deliver that is worthy of a question mark. So it will be worth checking whether they finished the slipped 3.1 stuff.
any chance we can do [bind culling] in a 3.2.x PTU build?
CJ: "Not my decision but I don't think you are likely to see Bind Culling in a 3.2.x PTU build. Without Object Container Streaming, Bind Culling causes loading stalls as entities enter the client's range. We're still working on being able to turn on Bind Culling in the PU, but even in the simpler test maps where we do have it working, the loading stalls are pretty frequent and noticeable. That's fine for developers, as it gives us a way to find and fix the bugs that Bind Culling/OCS will introduce while OCS is still being worked on, but it's probably not an experience most backers would want."
Thanks sir, looking forward to the massive performance boost of 3.3 then
CJ: "Bind Culling and OCS should give us performance improvements on clients and that is, after all, one of the big reasons for doing them in the first place. But let's not count our chickens before they hatch. The only way to know for sure how much of difference BC/OCS will make is to finish implementing them and then measure the difference."
I don’t want to be alarmist, however how would performance get to a point where the goals of the game as stated could be met without major improvements from those two methods? What are the other options?
CJ: "There are many ways to skin the performance cat. If we didn't do BC/OCS we'd find other ways to meet our goals. The point I was trying to make is that no one optimization is going to be the silver bullet. It's going to take a lot of different optimizations, each optimizing different areas of the game in different ways. We're confident we'll get there, it just takes a while."
Something I've wondered a lot is what exactly OCS is. I mean, I know that it involves moving entity loading off the main thread and making entity loading asynchronous, but is there more to it than that? I've read a lot of claims, some more plausible-sounding than others, but I'm not aware of anything definitive from CIG on the matter.
CJ: "I don't work on Object Containers so take what I'm about to say with a pinch of salt.
My understanding is that an Object Container is an entity that contains other entities. What makes OCs special is that they are collapsible so we can unload the contents of an OC but keep the OC entity itself in memory. You can think of any large structure in the game (ships and even the solar system) as a hierarchy of OCs. These hierarchies of OCs form the skeletons of these structures and the engine can opt to load or unload the different parts. The skeleton is always there, but the meat on the bones can be stripped back if it isn't needed by a client.
Possibly that was a bit of a gruesome analogy but I hope it helps get the idea across
CIG has now moved most of the stuff to polishing status. Stuff that's not yet in polishing status: -All armor reworks and new clothing -Economy Polish -Groups System -Quantum Linking -Anvil Hurricane -Vanduul Blade -1 new ship weapon and 1 new FPS weapon
Stuff like new armor and clothing doesn't need that much time in polishing status, but Groups System, Quantum Linking and new ships might release a bit broken because soon there isn't enough time to test and fix them.
They need to seriously work on performance. I have one of the fastest gaming computers in the world (5.2 GHz 8700K and overclocked Titan V) and the game runs like doody poo poo.
They need to seriously work on performance. I have one of the fastest gaming computers in the world (5.2 GHz 8700K and overclocked Titan V) and the game runs like doody poo poo.
Don't worry, by the time SC releases, that will be a super low end configuration.....
If you are holding out for the perfect game, the only game you play will be the waiting one.
They need to seriously work on performance. I have one of the fastest gaming computers in the world (5.2 GHz 8700K and overclocked Titan V) and the game runs like doody poo poo.
"We all do the best we can based on life experience, point of view, and our ability to believe in ourselves." - Naropa "We don't see things as they are, we see them as we are." SR Covey
Yes it's to expect after 3.2 releases that was already to be yesterday but some important bug-fixes show to be required beforehand.
3.3 will be reviewed and we'll likely see the Q2 2019 3.6 added, expecting that impact mostly the gameplay category, aside from that as CR said on RTV yesterday, 3.3 is the big OCS long-awaited tech and is a core requirement to expand the game universe so OCS is a must-have so the first proper planet Hurston, Lorville landing zone, its moons and the rest stops can release.
New base tech: -Object Container Streaming and Network Bind Culling: Your client will only load things that are close to you, increased performance -Voice Over IP and Face Over IP integration -Environmental Blending Shader: dirt, ice, sand, etc. will blend into objects in environment
And also: -New ship weapons -1 new FPS weapons: Kastak Arms Karna assault rifle -Virgil TrueDef-Pro Armor -RSI Odyssey Flightsuit Rework -Hurston Clothing sets -Improvements to missions: Scanning added to missions, racing missions -Improvements to service beacons -Improvements to ship AI -Performance optimizations
Hah Network Entity Streaming jumped to completion, I guess one of OSC functions, stuff having to work together with the client bind culling as well. 3.3 surely will take priority on the tech front so that stalled gameplay category will have to face push backs when gets reviewed hmmm
They hadn't even started it last week, so whatever it was it was short job.
Or was already going on and wasn't added to the public roadmap, but might be one of those handlers that are just the backend that are the support, so the culling should be set to use it.
OCS is quite complex, but seen the ship interiors being their own container then streaming makes sense with performance not having to render what and who is inside, gives a more clear idea of what this is all about
The game development being efficent and not wasteful? Project management that doesn't make other software project managers shake their heads in disbelief?
The roadmap is just another device to string along the faithful, and entice the newbies. But hey, they're good at that....
If you are holding out for the perfect game, the only game you play will be the waiting one.
The game development being efficent and not wasteful? Project management that doesn't make other software project managers shake their heads in disbelief?
The roadmap is just another device to string along the faithful, and entice the newbies. But hey, they're good at that....
The game development being efficent and not wasteful? Project management that doesn't make other software project managers shake their heads in disbelief?
The roadmap is just another device to string along the faithful, and entice the newbies. But hey, they're good at that....
Comments
@Octagon7711 optimization as it's a task for every update is really the main focus now that the update is on testing and feature work is being complete, but as it was announced, the new release format is date-driven, this means that indeed if the patch is needing more optimization by the end of the month, it should still release.
Then they'll make up for it on smaller hotfix patches, 3.2.1 and so; there's really no better way to prevent the schedule to get off rails.
1. What we have seen of some of the planning makes it clear that there are some "common elements" of activity that feed multiple e.g. ship work. You can see this if you go back to some of the "more detailed" plans.
So internally e.g. "do widget X" is a precursor to do "ship 1 activity", "ship 2 activity", "ship 3 activity" etc. And this may be what they are tracking internally.
Do widget X though may well be something that, in isolation, no one cares about. What they want to know about are the ships (say). So what we get in the planning is "do X and ship activity 1", "do X and ship activity 2", "do X and ship activity 3".
If widget X is the last thing to be completed however all the ships (say) will be completed at the same time. Even though all that has happened internally is that they have finished widget X!
2. Partial progress. Always a bugbear. How do you account for work that is partially complete. The most common way - and one that is being used - is to break down top level tasks into sub-tasks. Hence the 11/39 and 23/48 type of numbers. And this does indeed give a better idea of how far things are along.
However you still have partial progress - on the unfinished sub-tasks. So 11/39 means 11 finished and 23 in some partial state of completion. And its always a case of how much effort do you invest in project management to capture that partial progress. Especially as they have broken the work down into what they believe they can deliver in a quarter.
If the sub-tasks - the 39s and 48s and so forth - are simply next level down activities - then 11/39 may be pretty close to complete! If extra effort has been applied to create "short duration" sequential sub-tasks then 11/39 may only be about half done. We simply don;t know - they should of course. And that is the information they will have when they decide to keep a task in the current patch or slip it to the next.
As a result big jumps in %complete don't surprise me. Just a case of having to focus on what they actually deliver in the 3 month window rather than the intervening weeks. And also do they deliver stuff that they slip into the next patch. Missing something by the odd week - it happens; if they slip something though and fail to deliver that is worthy of a question mark. So it will be worth checking whether they finished the slipped 3.1 stuff.
CJ: "Not my decision but I don't think you are likely to see Bind Culling in a 3.2.x PTU build. Without Object Container Streaming, Bind Culling causes loading stalls as entities enter the client's range. We're still working on being able to turn on Bind Culling in the PU, but even in the simpler test maps where we do have it working, the loading stalls are pretty frequent and noticeable. That's fine for developers, as it gives us a way to find and fix the bugs that Bind Culling/OCS will introduce while OCS is still being worked on, but it's probably not an experience most backers would want."
CJ: "Bind Culling and OCS should give us performance improvements on clients and that is, after all, one of the big reasons for doing them in the first place. But let's not count our chickens before they hatch. The only way to know for sure how much of difference BC/OCS will make is to finish implementing them and then measure the difference."
CJ: "There are many ways to skin the performance cat. If we didn't do BC/OCS we'd find other ways to meet our goals. The point I was trying to make is that no one optimization is going to be the silver bullet. It's going to take a lot of different optimizations, each optimizing different areas of the game in different ways. We're confident we'll get there, it just takes a while."
CJ: "I don't work on Object Containers so take what I'm about to say with a pinch of salt.
My understanding is that an Object Container is an entity that contains other entities. What makes OCs special is that they are collapsible so we can unload the contents of an OC but keep the OC entity itself in memory. You can think of any large structure in the game (ships and even the solar system) as a hierarchy of OCs. These hierarchies of OCs form the skeletons of these structures and the engine can opt to load or unload the different parts. The skeleton is always there, but the meat on the bones can be stripped back if it isn't needed by a client.
Possibly that was a bit of a gruesome analogy but I hope it helps get the idea across
CIG has now moved most of the stuff to polishing status. Stuff that's not yet in polishing status:
-All armor reworks and new clothing
-Economy Polish
-Groups System
-Quantum Linking
-Anvil Hurricane
-Vanduul Blade
-1 new ship weapon and 1 new FPS weapon
Stuff like new armor and clothing doesn't need that much time in polishing status, but Groups System, Quantum Linking and new ships might release a bit broken because soon there isn't enough time to test and fix them.
If you are holding out for the perfect game, the only game you play will be the waiting one.
Everything except Quantum Linking is now in polishing stage.
"We all do the best we can based on life experience, point of view, and our ability to believe in ourselves." - Naropa "We don't see things as they are, we see them as we are." SR Covey
3.3 will be reviewed and we'll likely see the Q2 2019 3.6 added, expecting that impact mostly the gameplay category, aside from that as CR said on RTV yesterday, 3.3 is the big OCS long-awaited tech and is a core requirement to expand the game universe so OCS is a must-have so the first proper planet Hurston, Lorville landing zone, its moons and the rest stops can release.
New locations:
-Planet Hurston and its moons
-Rest Stop stations (small)
New game mechanics:
-Repairing (manual repair, repair ships and repair bot)
-Refueling and harvesting fuel elements from clouds
-Salvaging
New AI:
-FPS Combat AI
New ships:
-Aegis Hammerhead
-Tumbril Cyclone: TR, RC, AA and RN variants
Ship reworks:
-RSI Constellation Phoenix
-Consolidated Outland Mustang (all variants)
New base tech:
-Object Container Streaming and Network Bind Culling: Your client will only load things that are close to you, increased performance
-Voice Over IP and Face Over IP integration
-Environmental Blending Shader: dirt, ice, sand, etc. will blend into objects in environment
And also:
-New ship weapons
-1 new FPS weapons: Kastak Arms Karna assault rifle
-Virgil TrueDef-Pro Armor
-RSI Odyssey Flightsuit Rework
-Hurston Clothing sets
-Improvements to missions: Scanning added to missions, racing missions
-Improvements to service beacons
-Improvements to ship AI
-Performance optimizations
RoadMap July 13th 2018
OCS is quite complex, but seen the ship interiors being their own container then streaming makes sense with performance not having to render what and who is inside, gives a more clear idea of what this is all about
MAGA
The roadmap is just another device to string along the faithful, and entice the newbies. But hey, they're good at that....
If you are holding out for the perfect game, the only game you play will be the waiting one.