Captive Biter Spawner Max Output Bufffer Scale With Quality

[DamnPhotons]

Higher quality captive biter spawners should be able to hold more than 100 eggs in their output slot

What?

All biter spawners should be able to hold 200 seconds' worth of production in their output slot.
Currently all quality biter spawners can hold a maximum of 100 eggs in their output slot, which takes 200 seconds for a normal quality spawner at 0.5/s. Higher quality biter spawners only get a faster crafting speed, higher health, and longer spoil time. A legendary biter spawner creates 1.25 eggs/s which means it fills its output in only 80 seconds and stops producing new eggs. It should be able to hold 250 eggs, which would take 200 seconds to produce

normal_spawner.PNG (505.99 KiB) Viewed 660 times

A normal spawner at max capacity

uncommon_spawner.PNG (551.48 KiB) Viewed 660 times

An uncommon spawner at max capacity

Why?

For all other production machines, the automatic insertion limit scales with the machine's crafting speed. This insertion limit (combined with recipe output count, productivity bonuses, and inserters overshooting the limit) determines that maximum number of items that can be in the output slot of a machine.

rails.PNG (963.37 KiB) Viewed 660 times

The limit on inserted stone is 9 without speed bonuses. The output can hold at least 18 rails.

speed_rails.PNG (964.54 KiB) Viewed 660 times

The limit on inserted stone is 45 with speed modules. The output can hold at least 90 rails.

The biter spawner is the only production entity that does not scale it's maximum output buffer with its crafting speed. This is very unintuitive, even though it makes sense in a more technical way. The spawners only take in fuel and since the recipe has no input ingredients, there is no insertion limit of ingredients to scale with crafting speed.

For players who are consistently using a steady stream of biter eggs, higher quality spawners provide a clear benefit. For players who mostly use their eggs in quick bursts when a ship full of promethium asteroids arrives at Nauvis, the crafting speed bonus is irrelevant. The count of buffered eggs is what matters when you are trying to launch thousands of eggs all at once. When you are waiting 30+ minutes for the ship to make a round trip past the edge of the solar system, it is no difference if it takes 80 or 200 seconds to make 100 eggs.

Biter spawners are one of the most challenging items in the game to create at legendary quality and at scale. There should be a significant reward for making the investment in doing so.

Changing the biter spawner to hold more output at higher qualities would make its behavior more consistent with other production entities and more intuitive for players. It will also give players more incentive to make the effort to create legendary biter spawners.

[Alfonse215]

Given 2.1's interplanetary signals and the ability to do platform-to-platform transfers, players will be able to make promethium science production more consistent. So while this would be useful, I don't think this feature will be as necessary for making use of legendary spawners for science production as you make it out to be.

The biter spawner is the only production entity that does not scale it's maximum output buffer with its crafting speed.

That's because it doesn't *need to*. The output scaling is there to ensure that inserters actually have the time to remove items. It takes time for inserters to swing and the like, so the machine may need to buffer more if it's really fast in order to allow continuous production with inserter-based output.

It takes 40 seconds for a legendary captive spawner to fill up its buffer. A single base quality non-fast inserter can keep up with that.

[DamnPhotons]

Given 2.1's interplanetary signals and the ability to do platform-to-platform transfers, players will be able to make promethium science production more consistent. So while this would be useful, I don't think this feature will be as necessary for making use of legendary spawners for science production as you make it out to be.

Platform-to-platform transfers are good for small to medium amounts of promethium science but do not scale up to megabase sizes.
The interplanetary signals just make this even more relevant, since you can easily send a signal to hold the eggs until the promethium ship is ready to receive them at Nauvis.
My Promethium ships consumes 2.16 million eggs in 25 minutes. I want to launch that all at once, instead of waiting for it to trickle in. So I need 21.6k spawners, regardless of the quality. If legendary spawners held more eggs, I could choose between 21.6k normal spawners and 8,640 legendary spawners. There is no reason for me to make legendary spawners when 21k normal spawners are just as effective as 21k legendary spawners.

The biter spawner is the only production entity that does not scale it's maximum output buffer with its crafting speed.

That's because it doesn't *need to*. The output scaling is there to ensure that inserters actually have the time to remove items. It takes time for inserters to swing and the like, so the machine may need to buffer more if it's really fast in order to allow continuous production with inserter-based output.

It takes 40 seconds for a legendary captive spawner to fill up its buffer. A single base quality non-fast inserter can keep up with that.

A point which I explicitly acknowledged as the technical reason, but which still doesn't change the counterintuitive nature of this behavior. It is the only entity that behaves this way, which makes it unintuitive, even though there is a specific reason for it.
Also, it's 80 seconds for a legendary spawner to fill, since the egg recipe is 10 seconds to make 5 eggs, so 0.5 eggs/sec for a normal quality spawner with a crafting speed of 1.

[Alfonse215]

Platform-to-platform transfers are good for small to medium amounts of promethium science but do not scale up to megabase sizes.

Why not?

Let's take your promethium ship example. It's starting to get full, so it calls for eggs. Egg ship 1 picks up a bunch of eggs (let's say 500k eggs) and flies out to meet it. Egg ship 2 starts picking up eggs once ES1 departs.

ES1 meets the promethium ship at the edge, where it transfers eggs for nearly 7 minutes, then returns carrying science. By that time, ES2 should be there and start transferring eggs. ES 3 is loading up.

This continues until the promethium ship is out of resources and heads back out to collect more.

Why doesn't this scale?

A point which I explicitly acknowledged as the technical reason, but which still doesn't change the counterintuitive nature of this behavior.

I don't understand what is unintuitive about it. Not every machine+recipe scales the output buffer when it gets faster; you just used an example of a case that does. You can give a machine that makes cargo landing pads a crafting speed of 5, and its output buffer size will not scale up.

[DamnPhotons]

Platform-to-platform transfers are good for small to medium amounts of promethium science but do not scale up to megabase sizes.

Why not?

Let's take your promethium ship example. It's starting to get full, so it calls for eggs. Egg ship 1 picks up a bunch of eggs (let's say 500k eggs) and flies out to meet it. Egg ship 2 starts picking up eggs once ES1 departs.

ES1 meets the promethium ship at the edge, where it transfers eggs for nearly 7 minutes, then returns carrying science. By that time, ES2 should be there and start transferring eggs. ES 3 is loading up.

This continues until the promethium ship is out of resources and heads back out to collect more.

Why doesn't this scale?

You're only reducing the problem by a factor of 4. Scaling problems are usually much larger in size, like a order of magnitude or more. You haven't significantly changed the problem in question. 500k eggs at once is 5k normal spawners and 5k normal spawners are just as good as 5k legendary spawners. Why not have 2k legendary spawners be just as effective as 5k normal spawners? Legendary captive biter spawners are an endgame item, even compared to other legendary entities. Why should their effective performance in an endgame scenario be identical to the performance of their normal quality version?

A point which I explicitly acknowledged as the technical reason, but which still doesn't change the counterintuitive nature of this behavior.

I don't understand what is unintuitive about it. Not every machine+recipe scales the output buffer when it gets faster; you just used an example of a case that does. You can give a machine that makes cargo landing pads a crafting speed of 5, and its output buffer size will not scale up.

You just need to go higher than a crafting speed of 5.

cargo.PNG (450.85 KiB) Viewed 595 times

speed_cargo.PNG (455.54 KiB) Viewed 595 times

[Alfonse215]

You're only reducing the problem by a factor of 4. Scaling problems are usually much larger in size, like a order of magnitude or more. You haven't significantly changed the problem in question. 500k eggs at once is 5k normal spawners and 5k normal spawners are just as good as 5k legendary spawners. Why not have 2k legendary spawners be just as effective as 5k normal spawners?

They can be if the next egg ship starts requesting eggs immediately after the previous one leaves. Remember: you're not waiting for the spawners to completely fill up before requesting the next cycle. The whole point of this is to ship continuously.

Also, since you're relying on continuous requests, you don't need to load 500k eggs "at once". You just lead to load them faster than they're being consumed. If it takes 7 minutes to consume 500k eggs, then you have less than 7 minutes to load them. At 30 eggs per minute for base quality, with say 6 minutes of loading time, you're looking at 2.8k base quality spawners. For legendary, that's 1.1k.

You just need to go higher than a crafting speed of 5.

Captive biter spawners can only go up to a speed of 2.5 via quality.

[DamnPhotons]

You're only reducing the problem by a factor of 4. Scaling problems are usually much larger in size, like a order of magnitude or more. You haven't significantly changed the problem in question. 500k eggs at once is 5k normal spawners and 5k normal spawners are just as good as 5k legendary spawners. Why not have 2k legendary spawners be just as effective as 5k normal spawners?

They can be if the next egg ship starts requesting eggs immediately after the previous one leaves. Remember: you're not waiting for the spawners to completely fill up before requesting the next cycle. The whole point of this is to ship continuously.

Also, since you're relying on continuous requests, you don't need to load 500k eggs "at once". You just lead to load them faster than they're being consumed. If it takes 7 minutes to consume 500k eggs, then you have less than 7 minutes to load them. At 30 eggs per minute for base quality, with say 6 minutes of loading time, you're looking at 2.8k base quality spawners. For legendary, that's 1.1k.

1. There should be multiple ways to design a base in a sandbox game like this, not a one size fits all approach. You're assuming that doing ship-to-ship transfers of eggs at the edge of the solar system is the only way.

2. The eggs start spoiling as soon as you remove them from the spawner, but not before. You don't want 6 minutes of loading time. You want <1 minute of loading time.

3. You're also assuming my ship keeps bouncing between the edge of the solar system and the shallowest part of the promethium asteroids, with the lowest collection rate. My ship makes about 6.3 million promethium science every two hours. That rate requires an average of 44k promethium asteroids per minute. If the ship gets eggs and fuel from ship-to-ship, then it will need to reach the point where it can collect 88k per minute, which will average out to the correct rate. That takes about 48 minutes one-way (and I'm using thruster stacking, which should not be required). So I do need to get all of my eggs at once and bouncing off the edge of the solar system for consistent resupplies of eggs is detrimental to the performance.

You just need to go higher than a crafting speed of 5.

Captive biter spawners can only go up to a speed of 2.5 via quality.

I was responding to your comment about crafting cargo landing pads with a machine that has a speed of 5. You said they proved that not all recipes scale with crafting speed. I was showing you that they follow the same trend. See my attached screenshots in my previous post.

I don't understand what is unintuitive about it. Not every machine+recipe scales the output buffer when it gets faster; you just used an example of a case that does. You can give a machine that makes cargo landing pads a crafting speed of 5, and its output buffer size will not scale up.

[Alfonse215]

1. There should be multiple ways to design a base in a sandbox game like this, not a one size fits all approach. You're assuming that doing ship-to-ship transfers of eggs at the edge of the solar system is the only way.

Shouldn't "multiple ways to design a base" have different downsides? I mean, that's generally how it works. Some ways of designing your base can make legendary captive spawners useful, while others make them less useful.

I don't see the problem there: the way you've chosen to make promethium science does not make legendary captive spawners very useful. Just like if you build all of your basic science production on Fulgora from scrap products, legendary Foundries aren't a particularly high priority item for you.

2. The eggs start spoiling as soon as you remove them from the spawner, but not before. You don't want 6 minutes of loading time. You want <1 minute of loading time.

Promethium science does not spoil, so it does not inherit the freshness of the egg it is crafted with. As such, it doesn't matter if the egg has 100% freshness or 0.1% freshness; as long as it is consumed while it's still an egg, you get 100% of the science value from it. So as long as the eggs get consumed within 30 minutes of the point when the platform started to load, you are fine.

That 30 minutes is not something to be afraid of. It is a budget that can be invested however you like. So, how do you want to invest those 30 minutes? Wouldn't it make sense to do it in a way that takes advantage of the time instead of squandering it?

3. You're also assuming my ship keeps bouncing between the edge of the solar system and the shallowest part of the promethium asteroids, with the lowest collection rate. My ship makes about 6.3 million promethium science every two hours. That rate requires an average of 44k promethium asteroids per minute. If the ship gets eggs and fuel from ship-to-ship, then it will need to reach the point where it can collect 88k per minute, which will average out to the correct rate. That takes about 48 minutes one-way (and I'm using thruster stacking, which should not be required). So I do need to get all of my eggs at once and bouncing off the edge of the solar system for consistent resupplies of eggs is detrimental to the performance.

You don't need to get all of your eggs "at once"; you need to get them "in 25 minutes", because that's how long you said it took to consume your promethium: "My Promethium ships consumes 2.16 million eggs in 25 minutes." It doesn't matter if if you get those eggs in one titanic rocket launch or multiple serieses of launches that last several minutes. So long as those 2.16 million eggs all find their way to the ship within that 25 minute timeframe (obviously accounting for crafting time), you'll get the same result.

Your platform full of promethium gets to the edge. There's a ship with 540k eggs on it waiting, which your platform requests. It spends the next 6.25 minutes eating those eggs, and sending science to the other ship. After that, it requests another 540k eggs; these are serviced by a new ship that has arrived in that 6.25 minutes. Repeat until you're out of promethium, then go back to harvest for another 96 minutes.

This seems like a production process that does in fact scale. And it makes good use of legendary captive spawners.

I was responding to your comment about crafting cargo landing pads with a machine that has a speed of 5. You said they proved that not all recipes scale with crafting speed. I was showing you that they follow the same trend.

They scale as needed based on the crafting time. That's my point; if the recipe + machine speed is too slow for scaling to kick in, then it doesn't. So there's no reason for a person to expect that legendary captive spawners would be fast enough for scaling to kick in.

It's not inconsistent; it's the system doing what it always did. You're asking for an exception.

[DamnPhotons]

1. There should be multiple ways to design a base in a sandbox game like this, not a one size fits all approach. You're assuming that doing ship-to-ship transfers of eggs at the edge of the solar system is the only way.

Shouldn't "multiple ways to design a base" have different downsides? I mean, that's generally how it works. Some ways of designing your base can make legendary captive spawners useful, while others make them less useful.

I don't see the problem there: the way you've chosen to make promethium science does not make legendary captive spawners very useful. Just like if you build all of your basic science production on Fulgora from scrap products, legendary Foundries aren't a particularly high priority item for you.

Yes, of course there are different downsides. However the biter spawner, and important late game entity, is suffering benefiting from quality only in niche situations. Your examine is a very unusual situation: making basic science on Fulgora out of scrap. My situation is very common: launching a lot of eggs at once to make promethium.

2. The eggs start spoiling as soon as you remove them from the spawner, but not before. You don't want 6 minutes of loading time. You want <1 minute of loading time.

Promethium science does not spoil, so it does not inherit the freshness of the egg it is crafted with. As such, it doesn't matter if the egg has 100% freshness or 0.1% freshness; as long as it is consumed while it's still an egg, you get 100% of the science value from it. So as long as the eggs get consumed within 30 minutes of the point when the platform started to load, you are fine.

That 30 minutes is not something to be afraid of. It is a budget that can be invested however you like. So, how do you want to invest those 30 minutes? Wouldn't it make sense to do it in a way that takes advantage of the time instead of squandering it?

So you want the ship to sit at the edge of the solar system for 20-25 minutes not collecting asteroids? And you want to throw in synchronizing the schedules of 5 ships on top of that? I take advantage of my time it by not having any downtime in my egg processing. You have downtime by having a second ship carry the eggs without processing them. And then you have the promethium ship spending 25 minutes sitting at the edge of the solar system, not collecting any asteroids.

Let's look at how much you're nitpicking about a niche "optimal" scenario.

Here's the timeline for one roundtrip of my ship:
1. Nauvis -> Edge of Solar System: 15 minutes
2. EoSS -> Shattered Planet -> EoSS: 1h 20 minutes
3. EoSS -> Nauvis: 15 minutes
4. Load/Unload at Nauvis: 25 minutes

Total time: 2h 15minutes
Output: 5.2-6M asteroids, 6.3-7M science

I am bottlebecked by loading science into the cargo bay. If I add more inserters loading, I lose inserters unloading.
28 inserters:
10 leg white loading science
16 dedicated leg white unloading eggs
2 flex leg white unloading eggs and buffering quantum chips and fusion cells onto belts out past EoSS
The loading inserters only have about 10 minutes of downtime where they are not constantly active. Once I trim down the loading stage at Nauvis, they will be at 100% uptime.
10 legendary white inserters loading science for 800 sci/s is 6M science in 125 minutes.
If I remove inserters unloading eggs, then the number of eggs I can process before they spoil goes down.
If I make my hub asymmetric, I can fit in 2 more inserters to load science, which gets me another +20% capacity or +1.2M for 7.2M science loaded in 125 minutes. I could already do this, but my Fulgora science won't keep up atm, so I haven't done it yet.
I built the ship to hold 5.2M asteroids and aimed for 25 minutes of science as a general goal, but it can actually fit 6M asteroids (+15%) and 29 minutes of science production is +16% from 25 minutes. So I am not wasting any time with processing my eggs once this thing is going to max capacity.
So if I can launch 2.4M eggs in one salvo of 480 rockets, I can be in and out of Nauvis in <2 minutes, giving me a round-trip-time of about 1h 52 minutes, with 6.45M science loaded at 960 sci/s nonstop.

Now this is one ship, no crazy synchronization schedule that could potentially desync after many hours, and it makes 4 stacked green belts of promethium science per second. Do you think that your design is somehow more optimized and that once people get to the point of making a steady 960 promethium science per second, they should go "oh let me try to overhaul everything and make 5 platforms move on a synchronized schedule with spoiling items so I can make 10-20% more of something and then I will finally have a use to make one of the most important late game items at higher quality"? I think what I've done is gone farther than 99% of players will, but it's also the most simple direct approach that most players would use. I can't imagine many players trying to synchronize multiple ships like that, when you can still fit so much capacity into a single large ship, without risking some kind of desync or deadlock that only occurs in space once every 10-20 hours. Your use case is hyper niche and it does not justify such an important late game production entity having almost no practical value at higher level qualities for the most common use cases. The way I am doing it is the only way to do it in 2.0 and is still the simplest and most direct way to do it in 2.1. I am confident it's also the most common.

I was responding to your comment about crafting cargo landing pads with a machine that has a speed of 5. You said they proved that not all recipes scale with crafting speed. I was showing you that they follow the same trend.

They scale as needed based on the crafting time. That's my point; if the recipe + machine speed is too slow for scaling to kick in, then it doesn't. So there's no reason for a person to expect that legendary captive spawners would be fast enough for scaling to kick in.

It's not inconsistent; it's the system doing what it always did. You're asking for an exception.

The logic is not inconsistent. The end behavior the users sees is.
The devs have made exceptions for much less. They made ice asteroids reprocessing take half the time of carbon and metal simply because there's so much ice at Aquilo and beyond. "There's a lot of ice here, so you should make more crushers that process ice asteroids" is a much more obvious and intuitive thing for a player to consider, and yet the devs still made an exception for that.
If you read the dev posts for FFF and the updates, they clearly value the game having a consistent, intuitive feel for how entities should behave. I am pointing out that this behavior is unintuitive, because it appears to not do what all other production machines do. I acknowledged in the very original post that there is a technical reason for this. I never disputed that. I am simply saying that the current behavior is unintuitive and unexpected. The average player is accustomed to entities buffering more input and output items as its crafting speed increases. They will intuitively expected the biter spawner to do the same and be surprised when it doesn't.

[coffee-factorio]

Wouldn't it be better to simply change the stack size (and possibly with a mod) and not mess with how one exceptional machine works?

Input Follows Auto Insertion. Output Follows Stack Size.

Screenshot 2026-08-23 172001.png (130.69 KiB) Viewed 349 times

If for no other reason that to see if it gets you what you want?

Sorry. But you can pop open a machine and test this, and I was kind of inspired to because I recalled furnaces filled up differently than how I can jam parts into them. A furnace will only fill up to 100 output plates, but you can jam well over that in ores into the input side.

[DamnPhotons]

Wouldn't it be better to simply change the stack size (and possibly with a mod) and not mess with how one exceptional machine works?

If for no other reason that to see if it gets you what you want?

Sorry. But you can pop open a machine and test this, and I was kind of inspired to because I recalled furnaces filled up differently than how I can jam parts into them. A furnace will only fill up to 100 output plates, but you can jam well over that in ores into the input side.

I'm not sure what you're suggesting here. How would changing the stack size of anything affect how many eggs legendary captive biter spawners can produce compared to normal quality spawners, whether the eggs are being consumed at a constant rate or in bulk? Unloading the eggs is not the bottleneck. Having eggs to be unloaded is.

[coffee-factorio]

You want more eggs in a safe space.

Outputs don't scale like inputs, outputs scale to 1 stack per machine only over all machines. Inputs are effected by variable speed as you described.

And that's something that's easily verified by trying to jam items in the output slot of the machine rather than the input slot.

So what we do, to get you more eggs, to see if say, this causes problems when someone makes everything legendary. Is just change the stack size of the item with a mod.

One way of evaluating the merits of an idea is to say "Yes, we'll try it. And..."

... personally I think you have an emotional investment in a ship and want to get the most out of it. I did, in fact sit down and think through what would happen to a large scale bioscience operation if I got to have 250 biter eggs at arbitrary ips rates.

[DamnPhotons]

If you change the stack size of eggs themselves then they will stack up that high in even normal biter spawners.

I'm also not worried about getting the most out of my ship, because I can easily build and supply enough normal quality biter spawners to supply it.

What I want is a meaningful reason to make the effort to produce hundreds or even thousands of legendary biter spawners.

[coffee-factorio]

viewtopic.php?t=3394

There is a post here, there is a quote from Alton Brown.

It says the only uni-tasker is the fire extinguisher.

I understand that it would make salvoing biter eggs seems easier when there is a ton of latency. I have a hunch you're meant to solve that with engineering and automation. It feels like combinators would go away to make room for a solution to add more sand to the sandbox (get more UPS) and not add more toys to the sandbox. But to be frank, if you're salvoing that many eggs at once the rocket launches are going to do so much that you should probably be looking for a way to load eggs into chests.

If I'm not convinced by that one use case I'll poke around to see if someone has thought this through or if this would be up for arbitrary consideration due to things which, I'd call logical red flags. Right now, it feels like arbitrary consideration because "It's about how it feels" and the developers have opinions I am not privy to and wouldn't pretend to be privy to.

[Alfonse215]

Here's the timeline for one roundtrip of my ship:
1. Nauvis -> Edge of Solar System: 15 minutes
2. EoSS -> Shattered Planet -> EoSS: 1h 20 minutes
3. EoSS -> Nauvis: 15 minutes
4. Load/Unload at Nauvis: 25 minutes

Remove the Nauvis->Edge and back, and you've reduced your total cycle time by 30 minutes. Instead of "Load/Unload at Nauvis", it's "Load/Unload at Edge". But it's the same 25 minutes.

That is what I'm suggesting. Those 15 minute trips to/from Nauvis don't need to happen.

Now this is one ship, no crazy synchronization schedule that could potentially desync after many hours

You talk about schedule synchronization as if it's some dark art. 2.1 gives you inter-surface communication and the ability to set platform requests via circuits. There's nothing fiddly about it anymore; the platforms just talk to each other.

When your promethium ship gets X% full, it sends an egg signal along a particular channel. The current egg ship at Nauvis starts loading up and departs. The next ship detects this departure and moves into place to load up. Etc.

This is not deeply complex circuit stuff here.

Do you think that your design is somehow more optimized and that once people get to the point of making a steady 960 promethium science per second, they should go "oh let me try to overhaul everything and make 5 platforms move on a synchronized schedule with spoiling items so I can make 10-20% more of something and then I will finally have a use to make one of the most important late game items at higher quality"?

It seems weird that you're so dismissive of a change that would give you a flat 23% improvement to your promethium SPM regardless of the utility of legendary captive spawners.

I would expect players to look at 2.1's new tools and go "hey, with inter-surface communication, let me reassess the most efficient way to make promethium science. Maybe the way I've been doing it can be improved."

[mmmPI]

What I want is a meaningful reason to make the effort to produce hundreds or even thousands of legendary biter spawners.

While i understand increasing the stacksize of eggs everywhere would not achieve the desired effect , i'm unsure why you'd think increasing the output buffer of quality spawner would change things positively. If you can buffer more eggs per spawner, you will effectively require less of them to "keep ready" the same quantity of eggs no ?

So if I can launch 2.4M eggs in one salvo of 480 rockets, I can be in and out of Nauvis in <2 minutes, giving me a round-trip-time of about 1h 52 minutes, with 6.45M science loaded at 960 sci/s nonstop.

Let's take it to the extreme, if you had 2.4 M captive spawner, they wouldn't even need a buffer ( just 1) , if you had 240K of them, you could do with a buffer of only 10 eggs, if you have 24000 of them, then you need a outputbuffer of 100. The more you increase the output buffer ( or stack size) the less need you have a for a large number of spawner at the same time to store the payload. It seems to me.

It's not the way i would reason usually, more like "how much machine running full time do i need to fulfill the task", 960 sci/s this is accounted for using their production rate, and if you can't achieve 100% uptime, you need to account for "more machine", if you ever increase the ouput buffer of captive spawner, it would just appears to me as easier to achieve 100% uptime, it would give me "more time" for my ships to travel without my spawner clogging.

You may see it as a "reason" to make legendary spawner, because they would allow the user more flexibility, that would make them "easier to use" an "additionnal benefit to quality", but other people may see the opposite, a larger allowance, and thus a reduced need for scaling numbers. A real reason ( for me ) to make "legendary" over "normal" spawner is their production speed, if you reduce the output buffer to 10 or 1 egg, then you really force the players to have "more spawner" or "better quality", you can no longer just use a "normal spawner that buffer a lot of time" to make up for the production rate of a "legendary". So overall i'm not too convinced by the proposition, as to me it would have the opposite effect, having larger buffer on quality spawner would cause me to feel the need for a smaller number of them at the same time because it would be easier to keep only the required amount for science running at all time.

The way I am doing it is the only way to do it in 2.0 and is still the simplest and most direct way to do it in 2.1. I am confident it's also the most common.

I feel it is the simplest and so probably the most comon, but not the only way as you can recycle overgrown soil for fresh eggs if you want. And maybe this is also one argument that works the other way, maybe 2.1 is meant to introduce new ways to do, not "simple" , not necessarily comon but not as niche as the overgrown soil recycling.

I am bottlebecked by loading science into the cargo bay. If I add more inserters loading, I lose inserters unloading.

I don't think that's a problem players comonly encounter ^^ and i don't think the proposal from this thread would adress that. Instead you should make another ship, if your bottle neck is putting science on the platform hub, that's a different area of bottleneck imo.