Alpha 3.16 Postmortem
Undefined Undefined NoneContent
English
Alpha 3.16 Postmortem
02/09/2022 - 9:00 AM
On December 22, 2021, we launched Alpha 3.16: Return to Jumptown, which introduced a number of new features and changes, including the release of the Dynamic Event Jumptown 2.0, the grav-lev rework, and derelict ships with traps to avoid and valuables to secure.
This was a unique patch cycle. As we mentioned in a Roadmap Roundup back in December, Star Citizen Alpha 3.15 took longer to get out the door than we had initially planned, which limited the amount of time we had to stabilize the 3.16 code base. For this reason, we opted to branch from the 3.15 development stream to avoid risking overall stability (which has been the best we've had in years). Taking this approach meant we'd operate on the same code base that's currently on the live servers, while manually integrating 3.16 features (specifically those we deem low risk to integrate).
The following is a postmortem from the senior developers themselves, detailing what was delivered and their thoughts on how it went.
Vehicles
John Crewe, Vehicle Director
The Vehicle Pillar had a relatively small delivery for Alpha 3.16, with the Vehicle Feature and Experience teams primarily focusing on delivering the grav-lev rework.
What went well
The prior state of grav-lev left a lot to be desired, both from the internal side and the public-facing side, as it never quite behaved how we wanted. Numerous issues were present with it both technically and visually that the rework tackled and almost entirely fixed. The experience now is much more visceral, allowing players to select their height and make much more dramatic banked turns without the risk of rolling. Crash recovery and auto descending after dismount also fixed long-standing issues with the system where bikes would often be left unreachable or immobile.
What didn’t go so well
Overall, the new system is significantly better but still has occasional issues detecting surroundings, typically when entering more confined vehicles. While many vehicles were never intended to have grav-lev bikes inside, we want to make the experience better for those who give it a try.
What we’ll do in the future
On the Vehicle Content side, we released the Drake Cutlass Steel. This debuted alongside additional work from the Actor Feature team who expanded upon their mounted-gun tech to allow it to support animated deployment states. The Steel provides an alternative dropship to the popular Anvil Valkyrie, rivaling it in guns and troop transport capacity with a smaller and more nimble chassis. In future patches, we’ll look at its handling and armor values to help improve its survivability coming into dropzones, where it’s currently not quite hitting the desired goals.
AI
Francesco Roccucci, AI Director
The AI team didn’t have any major deliveries for Alpha 3.16. Our main focus was on improvements to the Bartender experience.
What went well
In general, the release branch has been very stable for us, allowing us to close down a lot of issues, submit a lot of improvements, and work well together towards a very specific goal.
What didn’t go so well
We had several cases of submissions stomping on other changes, especially on the Mannequin animation database (ADB) side. This caused previously fixed bugs to resurface, which led to some frustration.
The current structure we use is good in terms of keeping some branches more stable, but it can also require the team to dedicate more time when integrating fixed bugs in the feature streams.
What we’ll do differently in the future
We are currently focusing on the creation of feature testmaps. These will be the gate when validating submissions as they will make sure that changes don’t have side effects or contain wrong data.
We also want the team to sit down and define a checklist that needs to be followed before doing submits until we can get more automatic validation tools.
Locations
Todd Papy, Star Citizen Live Director
Derelict Ships (MTL)
What went well
The level designer owned all gameplay aspects of the location, including where it was located, the navigation gameplay, the distribution of gameplay ingredients and lootboxes, which mission was available, and the flow within each location.
Internal playtests and data from the PTU provided a lot of meaningful feedback on navigation and readability. The team was extremely impressed with the level of feedback coming from the community, which will help as these types of locations are developed. All interaction by the community was really appreciated by the team.
The collaboration between the Design, Art, and Lighting teams was fun, efficient, and provided impressive results with minimal frustrations.
What didn’t go so well
Getting traction on features supported by other teams was difficult at times (mines, loot, mission debug, scanning).
We didn’t have enough QATR (Quality Assurance Test Request) testers. This resulted in the team having to find and write many of the bugs that affected these locations.
We had some setbacks surrounding the branching to release process (mentioned in the intro above) that ended up condensing the amount of time we had to release. This resulted in a last minute rollback, re-branching to an older version instead, which caused strain on the team, and has been discussed as something we would really like to avoid doing again.
What we’ll do differently in the future
We’ll kick off with a better task breakdown and define the level structure and production workflow before we start building stuff.
An entity sanity pass will be done before use (the caterpillar assets had many issues) and allow more time for polish and iteration. We’ll also aim to get the first playable version at the actual location earlier, and an audit of features and dependencies on other teams should happen earlier.
For testing, we’ll provide QA with early, elaborate design documents to allow their test plans to be ready on time for QATR and be more effective at finding legit game-breaking issues. The testers should also have proper tools for testing missions and forcing missions in certain locations.
Jumptown 2.0
What went well
The original/initial drug lab outpost was quite bland and didn’t provide any visual art for the player, opportunities to explore and loot, or provide any PvP/cover to players. For Jumptown 2.0, we wanted to change things to deliver a better player experience.
We had a strong initial kick-off where we laid out the limitations of what we could do that was agreed on and stuck to. We then gave the outside of the drug labs a visual overhaul, adding options for both ground and roof cover. On the inside, we provided the players with physical cargo dispensers to haul cargo from and new and improved drug lab sections. It was really great to have the time to expand the outposts more than we'd anticipated and I think it's much better for it.
What didn’t go so well
The last-minute addition of the mounted turrets was a bit hairy; ideally this should have been built in from the start.
While it’s great to see these outposts updated for the Dynamic Event, they will ultimately be replaced when we create the v2 of high-tech outposts.
We faced some struggles related to having the team split between Austin and the UK. Moving forwards, we will look into having more thorough handovers and an additional stakeholder based in the EU timezone to improve collaboration.
As mentioned above, having to switch content streams so late in the process resulted in additional strain on the team.
What we’ll do differently in the future
Overall, the new and improved drug lab is significantly better but still has some issues and we believe that if there were more time and support from other teams, the overall experience of the event could have been much better.
One aspect that we would want to improve on is the size of the base and drug lab. However, this will require a more significant time investment. By making the inside bigger, we would have more options for cover and PvP gameplay. Another aspect we could improve on is how players enter the base. By adding extra entrance and exit points, we could avoid players being ambushed by others through the only door.
Finally, it would be good to spread the cover further out from the lab to provide more cover from above from incoming ships and bombers, which would encourage more ground-based PVP action.
02/09/2022 - 9:00 AM
On December 22, 2021, we launched Alpha 3.16: Return to Jumptown, which introduced a number of new features and changes, including the release of the Dynamic Event Jumptown 2.0, the grav-lev rework, and derelict ships with traps to avoid and valuables to secure.
This was a unique patch cycle. As we mentioned in a Roadmap Roundup back in December, Star Citizen Alpha 3.15 took longer to get out the door than we had initially planned, which limited the amount of time we had to stabilize the 3.16 code base. For this reason, we opted to branch from the 3.15 development stream to avoid risking overall stability (which has been the best we've had in years). Taking this approach meant we'd operate on the same code base that's currently on the live servers, while manually integrating 3.16 features (specifically those we deem low risk to integrate).
The following is a postmortem from the senior developers themselves, detailing what was delivered and their thoughts on how it went.
Vehicles
John Crewe, Vehicle Director
The Vehicle Pillar had a relatively small delivery for Alpha 3.16, with the Vehicle Feature and Experience teams primarily focusing on delivering the grav-lev rework.
What went well
The prior state of grav-lev left a lot to be desired, both from the internal side and the public-facing side, as it never quite behaved how we wanted. Numerous issues were present with it both technically and visually that the rework tackled and almost entirely fixed. The experience now is much more visceral, allowing players to select their height and make much more dramatic banked turns without the risk of rolling. Crash recovery and auto descending after dismount also fixed long-standing issues with the system where bikes would often be left unreachable or immobile.
What didn’t go so well
Overall, the new system is significantly better but still has occasional issues detecting surroundings, typically when entering more confined vehicles. While many vehicles were never intended to have grav-lev bikes inside, we want to make the experience better for those who give it a try.
What we’ll do in the future
On the Vehicle Content side, we released the Drake Cutlass Steel. This debuted alongside additional work from the Actor Feature team who expanded upon their mounted-gun tech to allow it to support animated deployment states. The Steel provides an alternative dropship to the popular Anvil Valkyrie, rivaling it in guns and troop transport capacity with a smaller and more nimble chassis. In future patches, we’ll look at its handling and armor values to help improve its survivability coming into dropzones, where it’s currently not quite hitting the desired goals.
AI
Francesco Roccucci, AI Director
The AI team didn’t have any major deliveries for Alpha 3.16. Our main focus was on improvements to the Bartender experience.
What went well
In general, the release branch has been very stable for us, allowing us to close down a lot of issues, submit a lot of improvements, and work well together towards a very specific goal.
What didn’t go so well
We had several cases of submissions stomping on other changes, especially on the Mannequin animation database (ADB) side. This caused previously fixed bugs to resurface, which led to some frustration.
The current structure we use is good in terms of keeping some branches more stable, but it can also require the team to dedicate more time when integrating fixed bugs in the feature streams.
What we’ll do differently in the future
We are currently focusing on the creation of feature testmaps. These will be the gate when validating submissions as they will make sure that changes don’t have side effects or contain wrong data.
We also want the team to sit down and define a checklist that needs to be followed before doing submits until we can get more automatic validation tools.
Locations
Todd Papy, Star Citizen Live Director
Derelict Ships (MTL)
What went well
The level designer owned all gameplay aspects of the location, including where it was located, the navigation gameplay, the distribution of gameplay ingredients and lootboxes, which mission was available, and the flow within each location.
Internal playtests and data from the PTU provided a lot of meaningful feedback on navigation and readability. The team was extremely impressed with the level of feedback coming from the community, which will help as these types of locations are developed. All interaction by the community was really appreciated by the team.
The collaboration between the Design, Art, and Lighting teams was fun, efficient, and provided impressive results with minimal frustrations.
What didn’t go so well
Getting traction on features supported by other teams was difficult at times (mines, loot, mission debug, scanning).
We didn’t have enough QATR (Quality Assurance Test Request) testers. This resulted in the team having to find and write many of the bugs that affected these locations.
We had some setbacks surrounding the branching to release process (mentioned in the intro above) that ended up condensing the amount of time we had to release. This resulted in a last minute rollback, re-branching to an older version instead, which caused strain on the team, and has been discussed as something we would really like to avoid doing again.
What we’ll do differently in the future
We’ll kick off with a better task breakdown and define the level structure and production workflow before we start building stuff.
An entity sanity pass will be done before use (the caterpillar assets had many issues) and allow more time for polish and iteration. We’ll also aim to get the first playable version at the actual location earlier, and an audit of features and dependencies on other teams should happen earlier.
For testing, we’ll provide QA with early, elaborate design documents to allow their test plans to be ready on time for QATR and be more effective at finding legit game-breaking issues. The testers should also have proper tools for testing missions and forcing missions in certain locations.
Jumptown 2.0
What went well
The original/initial drug lab outpost was quite bland and didn’t provide any visual art for the player, opportunities to explore and loot, or provide any PvP/cover to players. For Jumptown 2.0, we wanted to change things to deliver a better player experience.
We had a strong initial kick-off where we laid out the limitations of what we could do that was agreed on and stuck to. We then gave the outside of the drug labs a visual overhaul, adding options for both ground and roof cover. On the inside, we provided the players with physical cargo dispensers to haul cargo from and new and improved drug lab sections. It was really great to have the time to expand the outposts more than we'd anticipated and I think it's much better for it.
What didn’t go so well
The last-minute addition of the mounted turrets was a bit hairy; ideally this should have been built in from the start.
While it’s great to see these outposts updated for the Dynamic Event, they will ultimately be replaced when we create the v2 of high-tech outposts.
We faced some struggles related to having the team split between Austin and the UK. Moving forwards, we will look into having more thorough handovers and an additional stakeholder based in the EU timezone to improve collaboration.
As mentioned above, having to switch content streams so late in the process resulted in additional strain on the team.
What we’ll do differently in the future
Overall, the new and improved drug lab is significantly better but still has some issues and we believe that if there were more time and support from other teams, the overall experience of the event could have been much better.
One aspect that we would want to improve on is the size of the base and drug lab. However, this will require a more significant time investment. By making the inside bigger, we would have more options for cover and PvP gameplay. Another aspect we could improve on is how players enter the base. By adding extra entrance and exit points, we could avoid players being ambushed by others through the only door.
Finally, it would be good to spread the cover further out from the lab to provide more cover from above from incoming ships and bombers, which would encourage more ground-based PVP action.
German
Alpha 3.16 Postmortem
02/09/2022 - 9:00 UHR
Am 22. Dezember 2021 haben wir Alpha 3.16: Rückkehr nach Jumptown veröffentlicht, die eine Reihe neuer Funktionen und Änderungen mit sich brachte, darunter das Dynamische Ereignis Jumptown 2.0, die Überarbeitung der Grav-Elev und verlassene Schiffe mit Fallen, die es zu vermeiden und Wertgegenstände zu sichern gilt.
Dies war ein einzigartiger Patch-Zyklus. Wie wir bereits im Dezember im Roadmap Roundup erwähnt haben, dauerte es länger, bis Star Citizen Alpha 3.15 veröffentlicht wurde, als wir ursprünglich geplant hatten. Aus diesem Grund haben wir uns entschieden, die Entwicklung von 3.15 zu verzweigen, um die Gesamtstabilität nicht zu gefährden (die so gut war wie seit Jahren nicht mehr). Dieser Ansatz bedeutete, dass wir mit der gleichen Codebasis arbeiten würden, die sich derzeit auf den Live-Servern befindet, während wir die 3.16-Funktionen (insbesondere diejenigen, deren Integration wir als risikoarm einstufen) manuell integrieren.
Nachfolgend findest du einen Bericht der leitenden Entwickler, in dem sie beschreiben, was geliefert wurde und was sie darüber denken, wie es gelaufen ist.
Fahrzeuge
John Crewe, Direktor für Fahrzeuge
Die Fahrzeugsäule hat für Alpha 3.16 relativ wenig geliefert. Die Teams für Fahrzeugfunktionen und -erfahrungen haben sich hauptsächlich auf die Überarbeitung der Grav-Levs konzentriert.
Was gut lief
Der vorherige Zustand von grav-lev ließ viel zu wünschen übrig, sowohl intern als auch für die Öffentlichkeit, da es sich nie so verhielt, wie wir es wollten. Es gab zahlreiche technische und visuelle Probleme, die durch die Überarbeitung fast vollständig behoben wurden. Das Erlebnis ist jetzt viel realistischer, da die Spieler/innen ihre Höhe selbst bestimmen und viel dramatischere Kurven fahren können, ohne Gefahr zu laufen, zu überrollen. Die Wiederherstellung bei Abstürzen und das automatische Absteigen nach dem Absteigen beheben auch langjährige Probleme mit dem System, bei denen die Fahrräder oft unerreichbar oder unbeweglich blieben.
Was nicht so gut gelaufen ist
Insgesamt ist das neue System deutlich besser, hat aber immer noch gelegentlich Probleme, die Umgebung zu erkennen, vor allem beim Einsteigen in engere Fahrzeuge. Auch wenn viele Fahrzeuge nie dafür gedacht waren, Grav-Lift-Bikes darin zu transportieren, wollen wir das Erlebnis für diejenigen, die es ausprobieren, verbessern.
Was wir in Zukunft tun werden
Was die Fahrzeuginhalte angeht, haben wir den Drake Cutlass Steel veröffentlicht. Dieser wurde zusammen mit zusätzlicher Arbeit des Actor Feature-Teams vorgestellt, das die Technologie der montierten Kanone erweitert hat, um animierte Einsatzzustände zu ermöglichen. Die Steel ist ein alternatives Dropship zur beliebten Anvil Valkyrie, das mit einem kleineren und wendigeren Chassis in Sachen Bewaffnung und Truppentransportkapazität mit ihr konkurriert. In zukünftigen Patches werden wir uns sein Handling und seine Panzerungswerte ansehen, um seine Überlebensfähigkeit in Dropzones zu verbessern, wo er derzeit nicht ganz die gewünschten Ziele erreicht.
KI
Francesco Roccucci, KI-Direktor
Das KI-Team hatte für die Alpha 3.16 keine größeren Lieferungen. Unser Hauptaugenmerk lag auf der Verbesserung des Bartender-Erlebnisses.
Was gut lief
Im Allgemeinen war der Release-Zweig für uns sehr stabil und ermöglichte es uns, viele Probleme zu lösen, viele Verbesserungen einzureichen und gut zusammenzuarbeiten, um ein ganz bestimmtes Ziel zu erreichen.
Was nicht so gut gelaufen ist
Es gab mehrere Fälle, in denen Einreichungen auf anderen Änderungen herumtrampelten, vor allem auf der Seite der Mannequin-Animationsdatenbank (ADB). Das führte dazu, dass bereits behobene Fehler wieder auftauchten, was zu einiger Frustration führte.
Die derzeitige Struktur, die wir verwenden, ist gut, um einige Zweige stabiler zu halten, aber sie kann auch bedeuten, dass das Team mehr Zeit aufwenden muss, um behobene Fehler in die Feature Streams zu integrieren.
Was wir in Zukunft anders machen werden
Wir konzentrieren uns derzeit auf die Erstellung von Feature-Testmaps. Diese werden bei der Validierung von Beiträgen das Tor sein, da sie sicherstellen, dass Änderungen keine Nebeneffekte haben oder falsche Daten enthalten.
Außerdem möchten wir, dass sich das Team zusammensetzt und eine Checkliste festlegt, die vor der Einreichung von Beiträgen befolgt werden muss, bis wir automatischere Validierungstools bekommen können.
Standorte
Todd Papy, Star Citizen Live Direktor
Verlassene Schiffe (MTL)
Was gut lief
Der Leveldesigner hatte alle Gameplay-Aspekte des Ortes in der Hand. Dazu gehörten der Standort, die Navigation, die Verteilung der Spielzutaten und Lootboxen, die verfügbaren Missionen und der Ablauf innerhalb des Ortes.
Interne Playtests und Daten aus dem PTU lieferten eine Menge aussagekräftiges Feedback zur Navigation und Lesbarkeit. Das Team war sehr beeindruckt von den Rückmeldungen aus der Community, die bei der Entwicklung dieser Art von Orten hilfreich sein werden. Das Team schätzte die Interaktion mit der Community sehr.
Die Zusammenarbeit zwischen dem Design-, dem Kunst- und dem Beleuchtungsteam hat Spaß gemacht, war effizient und lieferte beeindruckende Ergebnisse bei minimalen Frustrationen.
Was nicht so gut gelaufen ist
Es war manchmal schwierig, die von anderen Teams unterstützten Funktionen durchzusetzen (Minen, Beute, Missionsdebugging, Scannen).
Wir hatten nicht genügend QATR-Tester (Quality Assurance Test Request). Das führte dazu, dass das Team viele der Fehler, die diese Orte betrafen, selbst finden und schreiben musste.
Wir hatten einige Rückschläge im Zusammenhang mit dem Branching-to-Release-Prozess (wie oben in der Einleitung erwähnt), die die Zeit bis zur Veröffentlichung verkürzten. Dies führte zu einem Rollback in letzter Minute, bei dem wir stattdessen auf eine ältere Version zurückgriffen, was das Team sehr belastete.
Was wir in Zukunft anders machen werden
Wir fangen mit einer besseren Aufgabenverteilung an und legen die Ebenenstruktur und den Produktionsablauf fest, bevor wir mit dem Bau beginnen.
Wir werden vor der Verwendung einen "Sanity Pass" durchführen (die Raupen-Assets hatten viele Probleme) und uns mehr Zeit für den Feinschliff und die Iteration nehmen. Wir werden auch versuchen, die erste spielbare Version früher zu erstellen, und eine Prüfung der Funktionen und Abhängigkeiten von anderen Teams sollte früher stattfinden.
Für das Testen werden wir der QA frühzeitige, ausführliche Designdokumente zur Verfügung stellen, damit ihre Testpläne rechtzeitig für den QATR fertig sind und sie effektivere Probleme finden können, die das Spiel zerstören. Die Tester/innen sollten auch über geeignete Werkzeuge verfügen, um Missionen zu testen und Missionen an bestimmten Orten zu erzwingen.
Jumptown 2.0
Was gut lief
Der ursprüngliche Drogenlabor-Außenposten war ziemlich fade und bot dem Spieler keine visuelle Kunst, keine Möglichkeiten zum Erkunden und Beutemachen und keine PvP-Deckung für die Spieler. Für Jumptown 2.0 wollten wir die Dinge ändern, um den Spielern ein besseres Erlebnis zu bieten.
Wir hatten einen starken Auftakt, bei dem wir die Grenzen dessen, was wir tun konnten, abgesteckt haben und uns daran gehalten haben. Dann haben wir das Äußere der Drogenlabore optisch überarbeitet und Optionen für die Boden- und Dachabdeckung hinzugefügt. Im Inneren haben wir den Spielerinnen und Spielern physische Ladungsspender zur Verfügung gestellt, von denen aus sie Ladung transportieren können, sowie neue und verbesserte Abschnitte des Drogenlabors. Es war wirklich toll, dass wir die Zeit hatten, die Außenposten weiter auszubauen, als wir erwartet hatten, und ich finde, dass sie dadurch viel besser geworden sind.
Was nicht so gut gelaufen ist
Die in letzter Minute hinzugefügten Geschütztürme waren etwas haarig; idealerweise hätte man sie von Anfang an einbauen sollen.
Es ist zwar toll, dass diese Außenposten für das Dynamische Ereignis aktualisiert wurden, aber sie werden letztendlich ersetzt, wenn wir die 2.
Die Aufteilung des Teams zwischen Austin und dem Vereinigten Königreich hat uns einige Probleme bereitet. In Zukunft werden wir uns um eine gründlichere Übergabe und einen zusätzlichen Stakeholder in der EU-Zeitzone bemühen, um die Zusammenarbeit zu verbessern.
Wie bereits erwähnt, führte der späte Wechsel der Inhaltsströme zu einer zusätzlichen Belastung für das Team.
Was wir in Zukunft anders machen werden
Insgesamt ist das neue und verbesserte Drogenlabor deutlich besser, hat aber immer noch einige Probleme und wir glauben, dass das Gesamterlebnis der Veranstaltung viel besser hätte sein können, wenn es mehr Zeit und Unterstützung von anderen Teams gegeben hätte.
Ein Aspekt, den wir gerne verbessern würden, ist die Größe der Basis und des Drogenlabors. Dies erfordert jedoch eine größere Zeitinvestition. Wenn wir das Innere größer machen würden, hätten wir mehr Möglichkeiten für Deckung und PvP-Gameplay. Ein weiterer Aspekt, den wir verbessern könnten, ist die Art und Weise, wie Spieler die Basis betreten. Wenn wir zusätzliche Ein- und Ausgänge einbauen, können wir verhindern, dass die Spieler/innen durch die einzige Tür in einen Hinterhalt geraten.
Schließlich wäre es gut, die Deckung weiter vom Labor aus zu verteilen, um mehr Deckung von oben vor ankommenden Schiffen und Bombern zu bieten, was mehr bodengebundene PvP-Action fördern würde.
02/09/2022 - 9:00 UHR
Am 22. Dezember 2021 haben wir Alpha 3.16: Rückkehr nach Jumptown veröffentlicht, die eine Reihe neuer Funktionen und Änderungen mit sich brachte, darunter das Dynamische Ereignis Jumptown 2.0, die Überarbeitung der Grav-Elev und verlassene Schiffe mit Fallen, die es zu vermeiden und Wertgegenstände zu sichern gilt.
Dies war ein einzigartiger Patch-Zyklus. Wie wir bereits im Dezember im Roadmap Roundup erwähnt haben, dauerte es länger, bis Star Citizen Alpha 3.15 veröffentlicht wurde, als wir ursprünglich geplant hatten. Aus diesem Grund haben wir uns entschieden, die Entwicklung von 3.15 zu verzweigen, um die Gesamtstabilität nicht zu gefährden (die so gut war wie seit Jahren nicht mehr). Dieser Ansatz bedeutete, dass wir mit der gleichen Codebasis arbeiten würden, die sich derzeit auf den Live-Servern befindet, während wir die 3.16-Funktionen (insbesondere diejenigen, deren Integration wir als risikoarm einstufen) manuell integrieren.
Nachfolgend findest du einen Bericht der leitenden Entwickler, in dem sie beschreiben, was geliefert wurde und was sie darüber denken, wie es gelaufen ist.
Fahrzeuge
John Crewe, Direktor für Fahrzeuge
Die Fahrzeugsäule hat für Alpha 3.16 relativ wenig geliefert. Die Teams für Fahrzeugfunktionen und -erfahrungen haben sich hauptsächlich auf die Überarbeitung der Grav-Levs konzentriert.
Was gut lief
Der vorherige Zustand von grav-lev ließ viel zu wünschen übrig, sowohl intern als auch für die Öffentlichkeit, da es sich nie so verhielt, wie wir es wollten. Es gab zahlreiche technische und visuelle Probleme, die durch die Überarbeitung fast vollständig behoben wurden. Das Erlebnis ist jetzt viel realistischer, da die Spieler/innen ihre Höhe selbst bestimmen und viel dramatischere Kurven fahren können, ohne Gefahr zu laufen, zu überrollen. Die Wiederherstellung bei Abstürzen und das automatische Absteigen nach dem Absteigen beheben auch langjährige Probleme mit dem System, bei denen die Fahrräder oft unerreichbar oder unbeweglich blieben.
Was nicht so gut gelaufen ist
Insgesamt ist das neue System deutlich besser, hat aber immer noch gelegentlich Probleme, die Umgebung zu erkennen, vor allem beim Einsteigen in engere Fahrzeuge. Auch wenn viele Fahrzeuge nie dafür gedacht waren, Grav-Lift-Bikes darin zu transportieren, wollen wir das Erlebnis für diejenigen, die es ausprobieren, verbessern.
Was wir in Zukunft tun werden
Was die Fahrzeuginhalte angeht, haben wir den Drake Cutlass Steel veröffentlicht. Dieser wurde zusammen mit zusätzlicher Arbeit des Actor Feature-Teams vorgestellt, das die Technologie der montierten Kanone erweitert hat, um animierte Einsatzzustände zu ermöglichen. Die Steel ist ein alternatives Dropship zur beliebten Anvil Valkyrie, das mit einem kleineren und wendigeren Chassis in Sachen Bewaffnung und Truppentransportkapazität mit ihr konkurriert. In zukünftigen Patches werden wir uns sein Handling und seine Panzerungswerte ansehen, um seine Überlebensfähigkeit in Dropzones zu verbessern, wo er derzeit nicht ganz die gewünschten Ziele erreicht.
KI
Francesco Roccucci, KI-Direktor
Das KI-Team hatte für die Alpha 3.16 keine größeren Lieferungen. Unser Hauptaugenmerk lag auf der Verbesserung des Bartender-Erlebnisses.
Was gut lief
Im Allgemeinen war der Release-Zweig für uns sehr stabil und ermöglichte es uns, viele Probleme zu lösen, viele Verbesserungen einzureichen und gut zusammenzuarbeiten, um ein ganz bestimmtes Ziel zu erreichen.
Was nicht so gut gelaufen ist
Es gab mehrere Fälle, in denen Einreichungen auf anderen Änderungen herumtrampelten, vor allem auf der Seite der Mannequin-Animationsdatenbank (ADB). Das führte dazu, dass bereits behobene Fehler wieder auftauchten, was zu einiger Frustration führte.
Die derzeitige Struktur, die wir verwenden, ist gut, um einige Zweige stabiler zu halten, aber sie kann auch bedeuten, dass das Team mehr Zeit aufwenden muss, um behobene Fehler in die Feature Streams zu integrieren.
Was wir in Zukunft anders machen werden
Wir konzentrieren uns derzeit auf die Erstellung von Feature-Testmaps. Diese werden bei der Validierung von Beiträgen das Tor sein, da sie sicherstellen, dass Änderungen keine Nebeneffekte haben oder falsche Daten enthalten.
Außerdem möchten wir, dass sich das Team zusammensetzt und eine Checkliste festlegt, die vor der Einreichung von Beiträgen befolgt werden muss, bis wir automatischere Validierungstools bekommen können.
Standorte
Todd Papy, Star Citizen Live Direktor
Verlassene Schiffe (MTL)
Was gut lief
Der Leveldesigner hatte alle Gameplay-Aspekte des Ortes in der Hand. Dazu gehörten der Standort, die Navigation, die Verteilung der Spielzutaten und Lootboxen, die verfügbaren Missionen und der Ablauf innerhalb des Ortes.
Interne Playtests und Daten aus dem PTU lieferten eine Menge aussagekräftiges Feedback zur Navigation und Lesbarkeit. Das Team war sehr beeindruckt von den Rückmeldungen aus der Community, die bei der Entwicklung dieser Art von Orten hilfreich sein werden. Das Team schätzte die Interaktion mit der Community sehr.
Die Zusammenarbeit zwischen dem Design-, dem Kunst- und dem Beleuchtungsteam hat Spaß gemacht, war effizient und lieferte beeindruckende Ergebnisse bei minimalen Frustrationen.
Was nicht so gut gelaufen ist
Es war manchmal schwierig, die von anderen Teams unterstützten Funktionen durchzusetzen (Minen, Beute, Missionsdebugging, Scannen).
Wir hatten nicht genügend QATR-Tester (Quality Assurance Test Request). Das führte dazu, dass das Team viele der Fehler, die diese Orte betrafen, selbst finden und schreiben musste.
Wir hatten einige Rückschläge im Zusammenhang mit dem Branching-to-Release-Prozess (wie oben in der Einleitung erwähnt), die die Zeit bis zur Veröffentlichung verkürzten. Dies führte zu einem Rollback in letzter Minute, bei dem wir stattdessen auf eine ältere Version zurückgriffen, was das Team sehr belastete.
Was wir in Zukunft anders machen werden
Wir fangen mit einer besseren Aufgabenverteilung an und legen die Ebenenstruktur und den Produktionsablauf fest, bevor wir mit dem Bau beginnen.
Wir werden vor der Verwendung einen "Sanity Pass" durchführen (die Raupen-Assets hatten viele Probleme) und uns mehr Zeit für den Feinschliff und die Iteration nehmen. Wir werden auch versuchen, die erste spielbare Version früher zu erstellen, und eine Prüfung der Funktionen und Abhängigkeiten von anderen Teams sollte früher stattfinden.
Für das Testen werden wir der QA frühzeitige, ausführliche Designdokumente zur Verfügung stellen, damit ihre Testpläne rechtzeitig für den QATR fertig sind und sie effektivere Probleme finden können, die das Spiel zerstören. Die Tester/innen sollten auch über geeignete Werkzeuge verfügen, um Missionen zu testen und Missionen an bestimmten Orten zu erzwingen.
Jumptown 2.0
Was gut lief
Der ursprüngliche Drogenlabor-Außenposten war ziemlich fade und bot dem Spieler keine visuelle Kunst, keine Möglichkeiten zum Erkunden und Beutemachen und keine PvP-Deckung für die Spieler. Für Jumptown 2.0 wollten wir die Dinge ändern, um den Spielern ein besseres Erlebnis zu bieten.
Wir hatten einen starken Auftakt, bei dem wir die Grenzen dessen, was wir tun konnten, abgesteckt haben und uns daran gehalten haben. Dann haben wir das Äußere der Drogenlabore optisch überarbeitet und Optionen für die Boden- und Dachabdeckung hinzugefügt. Im Inneren haben wir den Spielerinnen und Spielern physische Ladungsspender zur Verfügung gestellt, von denen aus sie Ladung transportieren können, sowie neue und verbesserte Abschnitte des Drogenlabors. Es war wirklich toll, dass wir die Zeit hatten, die Außenposten weiter auszubauen, als wir erwartet hatten, und ich finde, dass sie dadurch viel besser geworden sind.
Was nicht so gut gelaufen ist
Die in letzter Minute hinzugefügten Geschütztürme waren etwas haarig; idealerweise hätte man sie von Anfang an einbauen sollen.
Es ist zwar toll, dass diese Außenposten für das Dynamische Ereignis aktualisiert wurden, aber sie werden letztendlich ersetzt, wenn wir die 2.
Die Aufteilung des Teams zwischen Austin und dem Vereinigten Königreich hat uns einige Probleme bereitet. In Zukunft werden wir uns um eine gründlichere Übergabe und einen zusätzlichen Stakeholder in der EU-Zeitzone bemühen, um die Zusammenarbeit zu verbessern.
Wie bereits erwähnt, führte der späte Wechsel der Inhaltsströme zu einer zusätzlichen Belastung für das Team.
Was wir in Zukunft anders machen werden
Insgesamt ist das neue und verbesserte Drogenlabor deutlich besser, hat aber immer noch einige Probleme und wir glauben, dass das Gesamterlebnis der Veranstaltung viel besser hätte sein können, wenn es mehr Zeit und Unterstützung von anderen Teams gegeben hätte.
Ein Aspekt, den wir gerne verbessern würden, ist die Größe der Basis und des Drogenlabors. Dies erfordert jedoch eine größere Zeitinvestition. Wenn wir das Innere größer machen würden, hätten wir mehr Möglichkeiten für Deckung und PvP-Gameplay. Ein weiterer Aspekt, den wir verbessern könnten, ist die Art und Weise, wie Spieler die Basis betreten. Wenn wir zusätzliche Ein- und Ausgänge einbauen, können wir verhindern, dass die Spieler/innen durch die einzige Tür in einen Hinterhalt geraten.
Schließlich wäre es gut, die Deckung weiter vom Labor aus zu verteilen, um mehr Deckung von oben vor ankommenden Schiffen und Bombern zu bieten, was mehr bodengebundene PvP-Action fördern würde.
Chinese
Alpha 3.16 Postmortem
02/09/2022 - 9:00 AM
On December 22, 2021, we launched Alpha 3.16: Return to Jumptown, which introduced a number of new features and changes, including the release of the Dynamic Event Jumptown 2.0, the grav-lev rework, and derelict ships with traps to avoid and valuables to secure.
This was a unique patch cycle. As we mentioned in a Roadmap Roundup back in December, Star Citizen Alpha 3.15 took longer to get out the door than we had initially planned, which limited the amount of time we had to stabilize the 3.16 code base. For this reason, we opted to branch from the 3.15 development stream to avoid risking overall stability (which has been the best we've had in years). Taking this approach meant we'd operate on the same code base that's currently on the live servers, while manually integrating 3.16 features (specifically those we deem low risk to integrate).
The following is a postmortem from the senior developers themselves, detailing what was delivered and their thoughts on how it went.
Vehicles
John Crewe, Vehicle Director
The Vehicle Pillar had a relatively small delivery for Alpha 3.16, with the Vehicle Feature and Experience teams primarily focusing on delivering the grav-lev rework.
What went well
The prior state of grav-lev left a lot to be desired, both from the internal side and the public-facing side, as it never quite behaved how we wanted. Numerous issues were present with it both technically and visually that the rework tackled and almost entirely fixed. The experience now is much more visceral, allowing players to select their height and make much more dramatic banked turns without the risk of rolling. Crash recovery and auto descending after dismount also fixed long-standing issues with the system where bikes would often be left unreachable or immobile.
What didn’t go so well
Overall, the new system is significantly better but still has occasional issues detecting surroundings, typically when entering more confined vehicles. While many vehicles were never intended to have grav-lev bikes inside, we want to make the experience better for those who give it a try.
What we’ll do in the future
On the Vehicle Content side, we released the Drake Cutlass Steel. This debuted alongside additional work from the Actor Feature team who expanded upon their mounted-gun tech to allow it to support animated deployment states. The Steel provides an alternative dropship to the popular Anvil Valkyrie, rivaling it in guns and troop transport capacity with a smaller and more nimble chassis. In future patches, we’ll look at its handling and armor values to help improve its survivability coming into dropzones, where it’s currently not quite hitting the desired goals.
AI
Francesco Roccucci, AI Director
The AI team didn’t have any major deliveries for Alpha 3.16. Our main focus was on improvements to the Bartender experience.
What went well
In general, the release branch has been very stable for us, allowing us to close down a lot of issues, submit a lot of improvements, and work well together towards a very specific goal.
What didn’t go so well
We had several cases of submissions stomping on other changes, especially on the Mannequin animation database (ADB) side. This caused previously fixed bugs to resurface, which led to some frustration.
The current structure we use is good in terms of keeping some branches more stable, but it can also require the team to dedicate more time when integrating fixed bugs in the feature streams.
What we’ll do differently in the future
We are currently focusing on the creation of feature testmaps. These will be the gate when validating submissions as they will make sure that changes don’t have side effects or contain wrong data.
We also want the team to sit down and define a checklist that needs to be followed before doing submits until we can get more automatic validation tools.
Locations
Todd Papy, Star Citizen Live Director
Derelict Ships (MTL)
What went well
The level designer owned all gameplay aspects of the location, including where it was located, the navigation gameplay, the distribution of gameplay ingredients and lootboxes, which mission was available, and the flow within each location.
Internal playtests and data from the PTU provided a lot of meaningful feedback on navigation and readability. The team was extremely impressed with the level of feedback coming from the community, which will help as these types of locations are developed. All interaction by the community was really appreciated by the team.
The collaboration between the Design, Art, and Lighting teams was fun, efficient, and provided impressive results with minimal frustrations.
What didn’t go so well
Getting traction on features supported by other teams was difficult at times (mines, loot, mission debug, scanning).
We didn’t have enough QATR (Quality Assurance Test Request) testers. This resulted in the team having to find and write many of the bugs that affected these locations.
We had some setbacks surrounding the branching to release process (mentioned in the intro above) that ended up condensing the amount of time we had to release. This resulted in a last minute rollback, re-branching to an older version instead, which caused strain on the team, and has been discussed as something we would really like to avoid doing again.
What we’ll do differently in the future
We’ll kick off with a better task breakdown and define the level structure and production workflow before we start building stuff.
An entity sanity pass will be done before use (the caterpillar assets had many issues) and allow more time for polish and iteration. We’ll also aim to get the first playable version at the actual location earlier, and an audit of features and dependencies on other teams should happen earlier.
For testing, we’ll provide QA with early, elaborate design documents to allow their test plans to be ready on time for QATR and be more effective at finding legit game-breaking issues. The testers should also have proper tools for testing missions and forcing missions in certain locations.
Jumptown 2.0
What went well
The original/initial drug lab outpost was quite bland and didn’t provide any visual art for the player, opportunities to explore and loot, or provide any PvP/cover to players. For Jumptown 2.0, we wanted to change things to deliver a better player experience.
We had a strong initial kick-off where we laid out the limitations of what we could do that was agreed on and stuck to. We then gave the outside of the drug labs a visual overhaul, adding options for both ground and roof cover. On the inside, we provided the players with physical cargo dispensers to haul cargo from and new and improved drug lab sections. It was really great to have the time to expand the outposts more than we'd anticipated and I think it's much better for it.
What didn’t go so well
The last-minute addition of the mounted turrets was a bit hairy; ideally this should have been built in from the start.
While it’s great to see these outposts updated for the Dynamic Event, they will ultimately be replaced when we create the v2 of high-tech outposts.
We faced some struggles related to having the team split between Austin and the UK. Moving forwards, we will look into having more thorough handovers and an additional stakeholder based in the EU timezone to improve collaboration.
As mentioned above, having to switch content streams so late in the process resulted in additional strain on the team.
What we’ll do differently in the future
Overall, the new and improved drug lab is significantly better but still has some issues and we believe that if there were more time and support from other teams, the overall experience of the event could have been much better.
One aspect that we would want to improve on is the size of the base and drug lab. However, this will require a more significant time investment. By making the inside bigger, we would have more options for cover and PvP gameplay. Another aspect we could improve on is how players enter the base. By adding extra entrance and exit points, we could avoid players being ambushed by others through the only door.
Finally, it would be good to spread the cover further out from the lab to provide more cover from above from incoming ships and bombers, which would encourage more ground-based PVP action.
02/09/2022 - 9:00 AM
On December 22, 2021, we launched Alpha 3.16: Return to Jumptown, which introduced a number of new features and changes, including the release of the Dynamic Event Jumptown 2.0, the grav-lev rework, and derelict ships with traps to avoid and valuables to secure.
This was a unique patch cycle. As we mentioned in a Roadmap Roundup back in December, Star Citizen Alpha 3.15 took longer to get out the door than we had initially planned, which limited the amount of time we had to stabilize the 3.16 code base. For this reason, we opted to branch from the 3.15 development stream to avoid risking overall stability (which has been the best we've had in years). Taking this approach meant we'd operate on the same code base that's currently on the live servers, while manually integrating 3.16 features (specifically those we deem low risk to integrate).
The following is a postmortem from the senior developers themselves, detailing what was delivered and their thoughts on how it went.
Vehicles
John Crewe, Vehicle Director
The Vehicle Pillar had a relatively small delivery for Alpha 3.16, with the Vehicle Feature and Experience teams primarily focusing on delivering the grav-lev rework.
What went well
The prior state of grav-lev left a lot to be desired, both from the internal side and the public-facing side, as it never quite behaved how we wanted. Numerous issues were present with it both technically and visually that the rework tackled and almost entirely fixed. The experience now is much more visceral, allowing players to select their height and make much more dramatic banked turns without the risk of rolling. Crash recovery and auto descending after dismount also fixed long-standing issues with the system where bikes would often be left unreachable or immobile.
What didn’t go so well
Overall, the new system is significantly better but still has occasional issues detecting surroundings, typically when entering more confined vehicles. While many vehicles were never intended to have grav-lev bikes inside, we want to make the experience better for those who give it a try.
What we’ll do in the future
On the Vehicle Content side, we released the Drake Cutlass Steel. This debuted alongside additional work from the Actor Feature team who expanded upon their mounted-gun tech to allow it to support animated deployment states. The Steel provides an alternative dropship to the popular Anvil Valkyrie, rivaling it in guns and troop transport capacity with a smaller and more nimble chassis. In future patches, we’ll look at its handling and armor values to help improve its survivability coming into dropzones, where it’s currently not quite hitting the desired goals.
AI
Francesco Roccucci, AI Director
The AI team didn’t have any major deliveries for Alpha 3.16. Our main focus was on improvements to the Bartender experience.
What went well
In general, the release branch has been very stable for us, allowing us to close down a lot of issues, submit a lot of improvements, and work well together towards a very specific goal.
What didn’t go so well
We had several cases of submissions stomping on other changes, especially on the Mannequin animation database (ADB) side. This caused previously fixed bugs to resurface, which led to some frustration.
The current structure we use is good in terms of keeping some branches more stable, but it can also require the team to dedicate more time when integrating fixed bugs in the feature streams.
What we’ll do differently in the future
We are currently focusing on the creation of feature testmaps. These will be the gate when validating submissions as they will make sure that changes don’t have side effects or contain wrong data.
We also want the team to sit down and define a checklist that needs to be followed before doing submits until we can get more automatic validation tools.
Locations
Todd Papy, Star Citizen Live Director
Derelict Ships (MTL)
What went well
The level designer owned all gameplay aspects of the location, including where it was located, the navigation gameplay, the distribution of gameplay ingredients and lootboxes, which mission was available, and the flow within each location.
Internal playtests and data from the PTU provided a lot of meaningful feedback on navigation and readability. The team was extremely impressed with the level of feedback coming from the community, which will help as these types of locations are developed. All interaction by the community was really appreciated by the team.
The collaboration between the Design, Art, and Lighting teams was fun, efficient, and provided impressive results with minimal frustrations.
What didn’t go so well
Getting traction on features supported by other teams was difficult at times (mines, loot, mission debug, scanning).
We didn’t have enough QATR (Quality Assurance Test Request) testers. This resulted in the team having to find and write many of the bugs that affected these locations.
We had some setbacks surrounding the branching to release process (mentioned in the intro above) that ended up condensing the amount of time we had to release. This resulted in a last minute rollback, re-branching to an older version instead, which caused strain on the team, and has been discussed as something we would really like to avoid doing again.
What we’ll do differently in the future
We’ll kick off with a better task breakdown and define the level structure and production workflow before we start building stuff.
An entity sanity pass will be done before use (the caterpillar assets had many issues) and allow more time for polish and iteration. We’ll also aim to get the first playable version at the actual location earlier, and an audit of features and dependencies on other teams should happen earlier.
For testing, we’ll provide QA with early, elaborate design documents to allow their test plans to be ready on time for QATR and be more effective at finding legit game-breaking issues. The testers should also have proper tools for testing missions and forcing missions in certain locations.
Jumptown 2.0
What went well
The original/initial drug lab outpost was quite bland and didn’t provide any visual art for the player, opportunities to explore and loot, or provide any PvP/cover to players. For Jumptown 2.0, we wanted to change things to deliver a better player experience.
We had a strong initial kick-off where we laid out the limitations of what we could do that was agreed on and stuck to. We then gave the outside of the drug labs a visual overhaul, adding options for both ground and roof cover. On the inside, we provided the players with physical cargo dispensers to haul cargo from and new and improved drug lab sections. It was really great to have the time to expand the outposts more than we'd anticipated and I think it's much better for it.
What didn’t go so well
The last-minute addition of the mounted turrets was a bit hairy; ideally this should have been built in from the start.
While it’s great to see these outposts updated for the Dynamic Event, they will ultimately be replaced when we create the v2 of high-tech outposts.
We faced some struggles related to having the team split between Austin and the UK. Moving forwards, we will look into having more thorough handovers and an additional stakeholder based in the EU timezone to improve collaboration.
As mentioned above, having to switch content streams so late in the process resulted in additional strain on the team.
What we’ll do differently in the future
Overall, the new and improved drug lab is significantly better but still has some issues and we believe that if there were more time and support from other teams, the overall experience of the event could have been much better.
One aspect that we would want to improve on is the size of the base and drug lab. However, this will require a more significant time investment. By making the inside bigger, we would have more options for cover and PvP gameplay. Another aspect we could improve on is how players enter the base. By adding extra entrance and exit points, we could avoid players being ambushed by others through the only door.
Finally, it would be good to spread the cover further out from the lab to provide more cover from above from incoming ships and bombers, which would encourage more ground-based PVP action.
Links
No links available.
Images
0No images available.
Metadata
- CIG ID
- 18549
- Channel
- Undefined
- Category
- Undefined
- Series
- None
- Comments
- 0
- Published
- 4 years ago (2022-02-09T23:00:00+00:00)