Friday Facts #298 - Demo upgrade for Stable

[Ext3h]

This breaks a fundamental rule that was decided back in FFF-128 when Loaders were being considered. This rule is that there is only one way to move items between structures, by using Inserters. Even though the items were never actually transferred, and our intent was that the Loader consumes them, it is enough that the player might think they were transferred.

That sums up the loader issue quite nicely.

If you think about it, what is a "loader" to start with? It's effectively just a specialization of an inserter, with the added requirement of terminating a belt in order to achieve full throughput.

That "belt terminal" part of the loader is the only missing fragment there. No direct insertion capabilities or anything, just an "end of belt" or "begin of belt" terminator which has barely enough internal buffer capacity to fully scale with inserter stack size.

We have mods for "belt boxes", which explored belt integrated storage but ignored the "terminal" constraint (and all have visual issues, due to not actually integrating a belt), and we have mods exploring the original "loader" design, as an inserter-free solution. Both approaches have bypassed the "only inserters transfer to different inventory type" constraint though.

If you were to revisit the idea of a loader, as merely the combination of regular, vanilla inserter + belt terminal piece, it should be fitting the established constraints quite well. Visualization wise, a variation of the "underground belt entrance" with a "blocky / chest like" instead of a "curved" cover actually feels like a natural choice for such a belt terminal.

[Qon]

One thing I truly hate about Destroyer capsules and despite them being awesome even pre-balance is the fact that by the time you deploy your max follower count of them - you literally have about a minute before they start exploding.

I agree, even Defenders are loads of fun once you auto deploy them. I made a mod until this gets added to the base game.
https://mods.factorio.com/mod/auto-deploy-robots

The mod is bugged and crashes the game. I reported it 2 months ago but no update...
https://mods.factorio.com/mod/auto-depl ... 000cf0b20c

Factorio 0.17.24
SADR 0.0.2

The mod Simple auto deploy robots caused a non-recoverable error.
Please report this error to the mod author.

Error while running event auto-deploy-robots::on_tick (ID 0)
auto-deploy-robots/autodeploy.lua:50: attempt to index global 'autodeploy' (a nil value)
stack traceback:
auto-deploy-robots/autodeploy.lua:50: in function '?'
auto-deploy-robots/event_handler.lua:7: in function <auto-deploy-robots/event_handler.lua:4>

Please fix?

[abregado]

The mod is bugged and crashes the game. I reported it 2 months ago but no update...

Lawl... some developer I am...

[abregado]

In which tread do you want us to place our play-trough screenshots?

Just email them to me. The address is in a text file in the script-output/campaign-screenshots directory.

[Deadlock989]

We have mods for "belt boxes", which explored belt integrated storage but ignored the "terminal" constraint (and all have visual issues, due to not actually integrating a belt), and we have mods exploring the original "loader" design, as an inserter-free solution. Both approaches have bypassed the "only inserters transfer to different inventory type" constraint though.

Beltboxes don't "integrate belts" because that isn't actually possible, unless you try hack-y runtime scripting which merely places yet more invisible non-colliding loaders and very loosely associates them with some other entity type. There were mods which tried that kind of thing: see Klonan's Belt Buffer (0.16 only). They inevitably break when some other mod does something ill-advised with player/robot build events. So entities with "built-in" belt feeders can only fake it, and are unreliable at best in a game with nearly 2000 mods made by well-meaning amateurs available, because no such belt-fed entity exists in vanilla and never will.

Visual integration of belts was made considerably more difficult in 0.17 because belts are now larger than a tile. Modded entities which overlap tile boundaries immediately run into layering issues because modders have only a fraction of the layering control that the devs have in the C++ engine. However those issues are not unique to modded entities: see various combinations of neighbouring splitter and stone wall for more examples. There is never going to be a perfect solution to these issues because this is a 2D sprite-driven game pretending to be 3D, and it works just fine 99.8% of the time.

All the loader mods which actually use the loader entity haven't "bypassed" anything. They inherit everything from the basic loader prototype that was abandoned over a year ago but oddly made it into the campaign. There is one exception with a mod which presents the player with a thing which looks like a loader but is actually just a pile of invisible mega-mega-mega-inserters in the same place, which unsurprisingly behaves more like an inserter.

Personally I'm a little disappointed by the news that loaders are finally being put on the shelf, but I'm not actually surprised. I would have liked to have seen them powered by electricity and feeding specific mass inventory types only, with inserters playing a "smarter", more precise role. But realistically, that ship sailed more than a year ago.

[Ext3h]

All the loader mods which actually use the loader entity haven't bypassed anything. They inherit everything from the basic loader prototype that was abandoned over a year ago but oddly made it into the campaign. There is one exception with a mod which presents the player with a thing which looks like a loader but is actually just a pile of invisible mega-mega-mega-inserters in the same place, which unsurprisingly behaves more like an inserter.

Personally I'm a little disappointed by the news that loaders are finally being put on the shelf, but I'm not actually surprised. I would have liked to have seen them powered by electricity and feeding specific mass inventory types only, with inserters playing a "smarter", more precise role. But realistically, that ship sailed more than a year ago.

The reason why it's being put on shelf is that its role collides with inserters though. Even more so if it was "powered by electricity" and "not smart", which effectively describes a boosted vanilla stack inserter. That already made the role of the original, "fully integrated" loader design redundant.

If you strip the parts of the design with are already provided by the inserter part, all that remains is:

• Inserting / retrieving at full (stack) inserter / belt throughput rate
  Picking items from a belt individually, inline, is a nice visual effect, but other than visuals, it more a source of bugs and annoyances due to unpredictable performance.
• Balanced insertion on both belt lanes from single item stack
  Balanced insertion is a major pain in most designs when placing something back onto a belt. An inline or terminal "belt modifier" to adjust the default insertion behavior is missing.

For inserting, that boils down to the old "belt hopper" / "funnel" design idea.
For retrieving it boils down to a (slightly buffering) terminal piece.

Both parts improve belt<->inserter interaction rather than attempting to substitute it. Both parts still require engine support to actually function properly.

[Quarnozian]

Now that Loaders are gone, we needed a new 1x1 Consumer entity which the player must use inserters to fill.

What about a delivery drone? Give it input slots like the ones in assemblers, where it can stack as many items as needed. (for example, concrete stacks to 100, but the assembler can stack it over 1000 for building the rocket silo) Once it has the requested items it can fly off as a way of "consuming" the items.

[Deadlock989]

The reason why it's being put on shelf is that its role collides with inserters though. Even more so if it was "powered by electricity" and "not smart", which effectively describes a boosted vanilla stack inserter. That already made the role of the original, "fully integrated" loader design redundant.

Not entirely. Loaders don't swing. Their transfer rate is constant and they never miss. You need to boost a single stack inserter well beyond vanilla limits for it to cope with a full belt. But I don't disagree with you. Loaders were the initial iteration of the stack inserter idea - they were first discussed in March 2016 and then stack inserters arrived the following June. I think with hindsight the "mistake", if you can call it that, was leaving loaders in the game half-done but moddable in the first place, tantalising us.

The devs clearly believe that inserters, with all their eccentricities, provide more interesting design puzzles and I don't really disagree with that either.

[eradicator]

It was never intended to stay like that. Have a go with the new Compilatron chest.

Btw, when i first read that i thought you were saying that for the duration of that quest compilatron *becomes* the chest. I.e. the player would be giving items directly to compilatron. :D

[youtakun]

The introduction campaign is bugged. It's totally stuck with no way to proceed and no idea if I did something wrong.

I need green circuits to make an assembly machine, but I need an assembly machine to make green circuits.

[Rseding91]

The introduction campaign is bugged. It's totally stuck with no way to proceed and no idea if I did something wrong.

I need green circuits to make an assembly machine, but I need an assembly machine to make green circuits.

You have an assembling machine in your inventory.

[youtakun]

Ok I barfed this. I'm too expert for this introduction. I though the existing assembly machines weren't blue or yellow so I tried to pipet them and couldn't so I thought I didn't have any. And stuff appearing in my inventory out of thin air is also apparently a thing in the scripted introductions.

[ownlyme]

I like the way the lasers changed, but would it be possible to also let them light up their surroundings e.g. the biters they hit and the ground around it? This would make it look even better.

this.
actually i kinda hoped you'd make them light everything around the beam like streams

[Jap2.0]

Well, I had a go at the tutorial. Impressions:

- In the first level, I got the solid fuel from the wreck (and then started mining coal), but compilatron went crazy for a little while wanting me to mine wood.
- Compilatron's dialogs go over the inventory GUI (and everything else for that matter - which can be annoying when it obscures a significant part of the screen). This was later annoying when he would always put his "out of ammo" comment right on top of my first ammo chest.
- The coal is set up so that making any more than two drills point in a loop is nigh impossible
- Why can't you rotate the belt in front of the inserter/chest (or just interact with it at all, for that matter?)
- Having a slightly longer delay before he starts complaining about things requiring coal might be nice (I see the out of fuel icon, no need to immediately pester me)
- Why no toolbelt?
- It gave me hints (in both the panel at top left and from compilatron) for things I had already used (namely control-click)
- Compilatron's speed always says 43.2km/h
- If compilatron is telling you to fuel one item and you place another, he will go to the other structure, give the fuel warning, and return to the first when that is fueled; if he is telling you to fuel one item and another runs out of fuel, he will stay at the first. Interesting observation, and probably intended (it seems reasonable to go to whatever the player is doing).
- Actually, sometimes he'll randomly change which structure he's giving a fuel warning about, so I'm not quite sure about how that works.
- It appears that he will only give fuel warnings about mining drills and furnaces (not inserters)
- It would be really nice to be able to ctrl+click items into the chests. If that's not feasible, I'd settle for circuit network control of inserters.
- "Status" on the tooltip is cool. Is that new?
- The small, scattered resources made me feel like the level design was trying to prohibit automation at large scale - especially the small amount of copper and the small scattered nature of the coal patch (and for any significant amount of burner infrastructure you need a ton of coal)
- Mining productivity is the only (initially visible) use of green science? I now feel no motivation to go towards green science.
- The early tech tree is rather linear and has a lot of (imo) nearly useless technologies
- The icon for electric inserters shows two inserters which are not unlocked by that technology
- Come on, I just built my burner-level perfect ratio green circuit setup and now I have to evacuate?
- But by evacuate, you mean "systematically deconstruct everything you have," right? Cool.
- Shoot, my inventory's full. Let's see if I can craft some stuff to compress my inventory a bit. Final inventory count at evacuation: 5 wooden chests, 2 iron chests, 199 belts, 50 burner inserters, 22 small electric poles, 4 boilers, 8 steam engines, 21 burner miners, 1 offshore pump, 5 stone furnaces, 24 assembly machine 1s, 3 labs, 2107 coal, 311 stone, 700 iron, 132 copper, 4 solid fuel, 3 copper wire, 80 iron gears, 212 electronic circuits, 21 red science packs, 1 radar.
- Oh hey, a toolbelt.
- Huh, if you have the tooltip follow the cursor, and pause the game while a tooltip is on, it will continue to follow the cursor in the foreground on the pause screen.
- PLEASE let me research splitters.
- Very nice hidden stuff around the map
- All I ever wanted to do was expand my iron, but now I have to deal with constant waves of biters
- I'm building my iron anyway.
- "New crafting recipes" shows up after researching things such as gun turret upgrades that don't give new recipes
- One time, a couple gun turrets ran out of ammo (I hadn't automated reloading that side yet), and I lost probably 20 minutes of progress. To be honest... it was rather disheartening. It'd be nice if it still autosaved; all the lack of autosaves taught me was that I needed to save manually.
- I'm trying to automate a whole lot of stuff at once... this would be so much less stressful with underground belts.
- The biter AI is a bit strange. It always targets a few specific areas, and won't stray from those targets - biters will run right by the turrets shooting at them, the player has to physically chase them, and half the time they completely ignore if you hit them with a pickaxe to the head.
- Did I mention that the nonexistant pickaxe is as good of a weapon as the pistol? Yeah, it's about as good.

Final score: 3:40:07, 24418 biter kills

Oh hey, now I can research logistics.


Abregado: there are a few things in here which I think might be considered bugs, should I report them or are you taking note/already know?

Neither the inserter not the chest are filtered. Looks like i can just stuff iron ore in there until no space left? (This is not a hidden request for filterable chests :p)

Just try it and see what happens

I did actually. And when the chest was "full" with iron ore i noticed that i had accidentially put one of each type of the target items, so the test was in vain. But due to this bug i don't want to try again just now. I'll just interpret that smiley as "it's been taken care of".

I checked and it appears to void items when it's a slot short of full (I put about 4k ore into it and it was stuck at 1.5k in the chest). It also works if you try to fill it with multiple other items. The excess does not appear to be returned.

[eradicator]

- PLEASE let me research splitters.
- I'm trying to automate a whole lot of stuff at once... this would be so much less stressful with underground belts.

I found the "you have only belts!" aspect strangely fascinating. A whole new problem to solve! I've sent you a picture of my solution :p.

- biters will run right by the turrets shooting at them

You must construct additional pylons walls pipes...

I think in general you tried to do too much too early. It's supposed to be aimed at new players. They won't try to mass-produce green chips *before* they need any. And besides the first "map" bit is so tiny. But then again, why not, he did tell us to break it...

there are a few things in here which I think might be considered bugs, should I report them

I think the answer to that is generally "yes".

[youtakun]

Some slight playability issues too, like what I already mentioned that stuff appears in your inventory out of thin air. There was also a time when it said, "you need to research". I looked in my inventory if there appeared science centers. nope. I checked if they are now craftable. nope. Then I was confused a bit until I noticed they appeared in the world out of thin air (at least I didn't see the compilatron built them)

[Unknow0059]

I'm surprised beams weren't visible at night before. This is one of those things that are really basic.

I never noticed because i haven't got to a point in the game where i need multiple lasers, so i'm not around them to see them in action.

[abregado]

Nice feedback, I guess if you put in so much effort Id better do so as well.

- Compilatron's dialogs go over the inventory GUI (and everything else for that matter - which can be annoying when it obscures a significant part of the screen). This was later annoying when he would always put his "out of ammo" comment right on top of my first ammo chest.
- If compilatron is telling you to fuel one item and you place another, he will go to the other structure, give the fuel warning, and return to the first when that is fueled; if he is telling you to fuel one item and another runs out of fuel, he will stay at the first. Interesting observation, and probably intended (it seems reasonable to go to whatever the player is doing).
- Actually, sometimes he'll randomly change which structure he's giving a fuel warning about, so I'm not quite sure about how that works.
- It appears that he will only give fuel warnings about mining drills and furnaces (not inserters)

Yeah Compilatron will still get a lot of polish. I tried to have it show things near the player, so when you place a burner structure it reevaluates which one to show, but isnt nessesarily the one you placed. Showing burner inserters is an issue because it raises the chances of Compi demolishing one of your buildings.. until we solve that, he wont show burner inserters.

- The coal is set up so that making any more than two drills point in a loop is nigh impossible
- The small, scattered resources made me feel like the level design was trying to prohibit automation at large scale - especially the small amount of copper and the small scattered nature of the coal patch (and for any significant amount of burner infrastructure you need a ton of coal)

You just need to stop having "expert blindness". Your brain is tricking you into thinking it is harder because your pre-optimized strategy no longer works.

The spacing on the resource patches really helps new players realize they need more than one miner. The human desire to fill all slots is very strong!

- Why can't you rotate the belt in front of the inserter/chest (or just interact with it at all, for that matter?)

To improve the chances that a new player puts on both sides of the belt

- Having a slightly longer delay before he starts complaining about things requiring coal might be nice (I see the out of fuel icon, no need to immediately pester me)

Yep, that will certainly be changed. Currently there is a bug which resets Compi's wait timer for this action... I will find it eventually.

- It gave me hints (in both the panel at top left and from compilatron) for things I had already used (namely control-click)

This is new. New players sometimes press something by accident and then dont know how to reproduce the effect. Now it gives you the hint if you do something, rather than after Compi tells you how. It does mean for veterens that it seems the game is telling you what you already did, which sucks and breaks one of the design constraints (If you dont need help you dont get it)

- It would be really nice to be able to ctrl+click items into the chests. If that's not feasible, I'd settle for circuit network control of inserters.

Easy does not mean productive. If i make it harder for the player here, it pays off in the next 40 minutes (more movement, more reward because of automating, and more production overall). it is only 50 items so why not just automate it and then do something productive with your time while you wait?

- Mining productivity is the only (initially visible) use of green science? I now feel no motivation to go towards green science.

We are still deciding what we wanted to put in the demo. Green science will be in for sure, and we needed one thing at least to spend it on. Maybe Defender robots?

- The early tech tree is rather linear and has a lot of (imo) nearly useless technologies
- The icon for electric inserters shows two inserters which are not unlocked by that technology

The early tech tree is not trying to be cool, like the freeplay tech tree. It is a tool for demonstrating the Technology Screen, research process, ect. It also helps the player see how far through the scenario they are.

- Huh, if you have the tooltip follow the cursor, and pause the game while a tooltip is on, it will continue to follow the cursor in the foreground on the pause screen.

This is a bug and will be fixed when the new tooltips come in.
Tooltip on the cursor is great for new players, but vets seem to hate it. Ideas?

- "New crafting recipes" shows up after researching things such as gun turret upgrades that don't give new recipes

Bug. That window is being replaced with a whole new one eventually.

- One time, a couple gun turrets ran out of ammo (I hadn't automated reloading that side yet), and I lost probably 20 minutes of progress. To be honest... it was rather disheartening. It'd be nice if it still autosaved; all the lack of autosaves taught me was that I needed to save manually.

The autosaves should be on! Only in the final wave are they disabled because if you died, you probably made a mistake at the start of the wave.

- The biter AI is a bit strange. It always targets a few specific areas, and won't stray from those targets - biters will run right by the turrets shooting at them, the player has to physically chase them, and half the time they completely ignore if you hit them with a pickaxe to the head.
- Did I mention that the nonexistant pickaxe is as good of a weapon as the pistol? Yeah, it's about as good.

Yeah, this is to promote having enough turrets and walls, otherwise you can just have 1 turret surrounded by 5 layers and it attracts all the biters. Some biters are sent to kill turrets, others are sent to attack miners/furnaces/steam engines.

Final score: 3:40:07, 24418 biter kills

Crikey... Thanks for playing!

I checked and it appears to void items when it's a slot short of full (I put about 4k ore into it and it was stuck at 1.5k in the chest). It also works if you try to fill it with multiple other items. The excess does not appear to be returned.

Yeah, if you do something crazy wrong like putting 1.5k ore into the chest it just voids a full stack of the most numerous item. Since all the important items are in small quantities, I dont see a need to worry about them getting voided.


Thanks for your feedback

[Jan11]

More important "Have lasers now a sound?"

[Jap2.0]

Yeah, a lot of that was definitely me wanting to do things how I've always done it. I did have a bit of fun and go completely overboard, though. (I spent... I don't know, maybe 15 hours in the last demo playing around with things in the time before I was able to buy the game?) I do find no underground belts to be a bit of a fun challenge - you have to design and nest things carefully if you want to automate a bunch of stuff with no way for one belt to cross over another (although there were a few places I couldn't do that - see for example the left coal line in the screenshot). I ultimately went for "mass produce all the things" (I probably could've finished it in half an hour less and a quarter the kills if I went straight for the objective). Here's a screenshot of my final base (direct link):

Base.png (2.69 MiB) Viewed 6772 times

- PLEASE let me research splitters.
- I'm trying to automate a whole lot of stuff at once... this would be so much less stressful with underground belts.

I found the "you have only belts!" aspect strangely fascinating. A whole new problem to solve! I've sent you a picture of my solution :p.

- biters will run right by the turrets shooting at them

You must construct additional pylons walls pipes...

I think in general you tried to do too much too early. It's supposed to be aimed at new players. They won't try to mass-produce green chips *before* they need any. And besides the first "map" bit is so tiny. But then again, why not, he did tell us to break it...

Yeah, it is a really intriguing puzzle. Yours is a really nice setup.
And yes, I definitely tried to do way too much, way too early - I should have done a little less, and scaled it up more evenly (for example, I had my defense on the left completely built up when at the same time I only had a couple turrets set down on the right).

I came in with the full intention of breaking it

In the end, I found that splitters aren't actually that much of an issue - they're nice, but you can do as much with inserters as you can with them. I should've split some of my production into multiple parts - it would've helped with a lot of my issues moving materials past each other.

- Compilatron's dialogs go over the inventory GUI (and everything else for that matter - which can be annoying when it obscures a significant part of the screen). This was later annoying when he would always put his "out of ammo" comment right on top of my first ammo chest.
- If compilatron is telling you to fuel one item and you place another, he will go to the other structure, give the fuel warning, and return to the first when that is fueled; if he is telling you to fuel one item and another runs out of fuel, he will stay at the first. Interesting observation, and probably intended (it seems reasonable to go to whatever the player is doing).
- Actually, sometimes he'll randomly change which structure he's giving a fuel warning about, so I'm not quite sure about how that works.
- It appears that he will only give fuel warnings about mining drills and furnaces (not inserters)

Yeah Compilatron will still get a lot of polish. I tried to have it show things near the player, so when you place a burner structure it reevaluates which one to show, but isnt nessesarily the one you placed. Showing burner inserters is an issue because it raises the chances of Compi demolishing one of your buildings.. until we solve that, he wont show burner inserters.

- The coal is set up so that making any more than two drills point in a loop is nigh impossible
- The small, scattered resources made me feel like the level design was trying to prohibit automation at large scale - especially the small amount of copper and the small scattered nature of the coal patch (and for any significant amount of burner infrastructure you need a ton of coal)

You just need to stop having "expert blindness". Your brain is tricking you into thinking it is harder because your pre-optimized strategy no longer works.

The spacing on the resource patches really helps new players realize they need more than one miner. The human desire to fill all slots is very strong!

- Why can't you rotate the belt in front of the inserter/chest (or just interact with it at all, for that matter?)

To improve the chances that a new player puts on both sides of the belt

- Having a slightly longer delay before he starts complaining about things requiring coal might be nice (I see the out of fuel icon, no need to immediately pester me)

Yep, that will certainly be changed. Currently there is a bug which resets Compi's wait timer for this action... I will find it eventually.

- It gave me hints (in both the panel at top left and from compilatron) for things I had already used (namely control-click)

This is new. New players sometimes press something by accident and then dont know how to reproduce the effect. Now it gives you the hint if you do something, rather than after Compi tells you how. It does mean for veterens that it seems the game is telling you what you already did, which sucks and breaks one of the design constraints (If you dont need help you dont get it)

- It would be really nice to be able to ctrl+click items into the chests. If that's not feasible, I'd settle for circuit network control of inserters.

Easy does not mean productive. If i make it harder for the player here, it pays off in the next 40 minutes (more movement, more reward because of automating, and more production overall). it is only 50 items so why not just automate it and then do something productive with your time while you wait?

All reasonable enough

- Mining productivity is the only (initially visible) use of green science? I now feel no motivation to go towards green science.

We are still deciding what we wanted to put in the demo. Green science will be in for sure, and we needed one thing at least to spend it on. Maybe Defender robots?

Yeah, I see how that would be an issue. You don't want to add too much to the campaign/demo (both so as to not overwhelm new players and so that they might actually buy the game sometime), and adding too complicated of things (eg. trains) might not be a good idea; you'd need one or two simple technologies that are exciting but also understandable to new players (for example, someone new might not know why I'm at all excited by electric energy distribution 1). Quick looking through the tech tree, I'm not seeing any magic bullet; combat robots might be a decent idea. Trains are cool, but complicated; cliff explosives solve a logistical challenge (cliffs in the middle of your base), but at that point cliffs are great natural defenses; maybe some decent weaponry (the smg, grenades)? The thing with that is that it also falls under your logic for only giving the pistol (although at that point they will have had to have built good defenses already). Another thing is that a lot of that is mostly good offensively, and (I think) you've said that destroying nests doesn't do anything in that level?

- The early tech tree is rather linear and has a lot of (imo) nearly useless technologies
- The icon for electric inserters shows two inserters which are not unlocked by that technology

The early tech tree is not trying to be cool, like the freeplay tech tree. It is a tool for demonstrating the Technology Screen, research process, ect. It also helps the player see how far through the scenario they are.

Fair enough. The story was pretty decent (and I was kind of ignoring that in that comment), although one thing that initially confused me looking ahead was that the technology removing the hand crafting restrictions looked like it completely unlocked the items with no distinction that they were already unlocked and this was just for hand crafting, which was a bit confusing (although I figured it out relatively quickly, and the scenario explained it).

- Huh, if you have the tooltip follow the cursor, and pause the game while a tooltip is on, it will continue to follow the cursor in the foreground on the pause screen.

This is a bug and will be fixed when the new tooltips come in.
Tooltip on the cursor is great for new players, but vets seem to hate it. Ideas?

Yeah... I'm not sure how I feel about that. I'm using a 1280x1024 monitor, so the default GUI size is 75%, which is tiny. I changed it to 100%, which almost works (there's a bit of overlap occasionally), and when I was doing so Bilka suggested trying that to free up some space. I haven't changed it back, although I don't prefer it to having the information on the side (I'm not sure if it's worse or not).

- "New crafting recipes" shows up after researching things such as gun turret upgrades that don't give new recipes

Bug. That window is being replaced with a whole new one eventually.

Do you need reports for those or are you already tracking them for the replacement?

- One time, a couple gun turrets ran out of ammo (I hadn't automated reloading that side yet), and I lost probably 20 minutes of progress. To be honest... it was rather disheartening. It'd be nice if it still autosaved; all the lack of autosaves taught me was that I needed to save manually.

The autosaves should be on! Only in the final wave are they disabled because if you died, you probably made a mistake at the start of the wave.

Yeah, it was after it gave me the warning about autosaves. I dragged the last wave on for a long time; I wanted to build up more infastructure and automation, so I researched a few levels of turret shooting speed before I did green science. I would like to note that at 16 labs, 3 (+ a bit) layers of walls, and 2 layers of turrets, the last wave's not much of a challenge (so much for difficulty scaling...). I didn't have to touch my defenses the entirety of that wave.

- The biter AI is a bit strange. It always targets a few specific areas, and won't stray from those targets - biters will run right by the turrets shooting at them, the player has to physically chase them, and half the time they completely ignore if you hit them with a pickaxe to the head.
- Did I mention that the nonexistant pickaxe is as good of a weapon as the pistol? Yeah, it's about as good.

Yeah, this is to promote having enough turrets and walls, otherwise you can just have 1 turret surrounded by 5 layers and it attracts all the biters. Some biters are sent to kill turrets, others are sent to attack miners/furnaces/steam engines.

Okay. I guess that kind of makes sense (although it was a bit annoying to have to chase the biters to the middle of the base - but I guess that adds a bit to the difficulty). I noticed that sometimes later on it's nearly impossible to defeat biters faster than they're coming in with the pistol, so if a turret runs out of ammo you really need to fill it up and then go kill the biters if you want to ever have a chance of ending the wave, even if it causes a bit more damage.

Another note: I should've automated more little things than I did. Completely hand-crafting dozens of turrets is really annoying, and there are several places that it would've been very easy to quickly automate things - setting up a gear assembler, or a wall assembler, for example, and I think that's something the scenario did a good job of pushing me toward (even if I entirely had my own plans) - I think with the examples in the first parts, it does a pretty good job of encouraging automating at least some.

Final thoughts: it's a different type of scenerio from how I usually play the game, but I enjoyed it (even if it was stressful at times). While people are saying that it isn't representative of the game - and I agree that the focus is somewhat different than the base game - the biters are a great source of motivation and goals, it's at least as good of a scenario as the last one (even if that one is a bit more nostalgic), and overall it's a better scenario.
Do you still want my folder of screenshots?