Nice! I like these last few paragraphs

uniformly distributed over the world

Limitation: discrepancies between single-node vs Distributed Worlds in computation

Remove stack plot.

Go from observation -> conclusion more clearly

Add caption to figure

wdym by populous? Be more precise

I realise it's not quite done. But your conclusions should also have a limitations and further work section

What does that tell me?

Something about accuracy of computation

This section is a nice add-on but leaves me wanting more.

Sure, the # of object stayed the same. Can you somehow verify that the objects paths/velocities behaved as if it were a single node?

Why the word data? Surely just 'Distributed Worlds' Or FPS

Also why does the blue line dip BEFORE the red line?

Again, might be nice to have the graph also show a single node server's performance to compare and contrast in the same graph.

Also, it's very similar to the # object graph. This could be noted. Or could be side by side to show similarity?

Haven't defined

It doesn't effectively quantify the 'duplicated space'.

It quantifies the duplicated space ...

No need for effectively or inverted commas for no reason.

(or I'd prefer quantifies the multiplier on space associated with...)

This relationship is unclear to me. Is it just 1 - acc ? Please define it much more clearly and not in parenthesis

It also seems to me like you've inverted the implied defn from above. If 1.0 is perfect accuracy then why is 0.1 good? Unless these are different accuracies (I think they are) in which case you should keep it consistent.

This could also be in the main section. I would define it briefly in the main paper (e.g. We use the increased amount of space covered by nodes associated with using Distributed Worlds on a square grid as an estimate of the increased computational cost.) Then link to a detailed formula here. It was too vague in the main section

This is just the duplicated area? Stating that this corresponds to a computational cost multiplier would assume that objects are distributed uniformly in space. This assumption should be stated - and that therefore this is just an estimate of the true computational cost multiplier at any given time.

A tidier version of this could be somewhere in the main paper I feel. It's really cool to actually see it.

where

You need to state more about the experiment. I couldn't reproduce it from this. I'd need to know node boundary shapes - squares? or hexagons? or something else? and # of nodes, # of boids, boid params etc. This could go in the appendix of you like, or even a link to a Github repo - but it's got to be in there somewhere.

How are you measuring this? I know you define it later but this should be up front and clear.

Also - does it approach 1? Surely at 100% node size, the boids are only aware of all boids in this node or neighbouring nodes - not ALL boids.

not sure this is the right phrasing

This could be more precise

New sentences

Maybe a figure showing boids in a world earlier in the paper?

Shouldn't the percentages always add up to 100?

Ah caught your expln later, but it'd be nice to have it mentioned earlier - particularly as what the y-axis represents depends on your defn of 'processed'. I think you mean that an object is visible rather than just owned - but this is not clear until I have to figure it out from the context later

Could have this and prev one be the same plot for one-to-one comparison? Only if it clearly demonstrates the desired point of course

Again, percentages should add up to 100?

I feel like the effect is harder to see as the processing on different servers spikes at the same time seemingly? If you could show just relative portion of total processing (i.e. it always adds up to 100%), then we'd see that the nodes are not stable in their share of total processing power? This plot makes it hard to see

Isn't a sentence

I like simplicity, but it doesn't tell the reader much? And isn't it more a 'Distributed World', as the idea is that there's only one world?

Why does it jump from 1-2-4-6-9?

Whatever graphics need to be shown,

Is it really? I'd rather you introduce this as a new ideal goal in order to maintain even and stable performance across nodes.

You don't need to redefine and reintroduce notions you've spent the previous chapter explaining

clunky sentence

why italics?

All benchmarks were run on the Apple M1 chip as it is....

Again, don't need to state. Just state

In this chapter, we study the empirical performance of the Distribute Worlds system to demonstrate its viability as a novel game engine architecture.

But!

Such sentences shouldn't be part of your diss

as shown in Appendix A1

Rework long sentence

remove will

an

concatenating the list of known objects ...

can be variable? given that you then make it fixed for this implementation?

Can you show this empirically?

requests?

that the

source? how do you know that it is generally considered so?

some distance? not just 'a radius'

boundary? margin?

this sentence should be two. And reworked slightly

and

??

this may be true but I'm not sure you've quite shown it.

You've just stated that hexagons minimise are and have 6 neighbours. Why does it directly follow that they minimise # of connections needed to cover a fixed area? I feel like it'd be kinda tough to prove that.

Surely this depends on the size of the margin and boundaries?

be more specific. hexagons minimise the perimeter of tile in order to cover a fixed area.

Funny! Not sure it's necessary but up to you

too, many, commas!

initialisation

can't be just comma. I'd split to 2 sentences

note that? important to note that?

You've previously mentioned websocket and are only now just defining it?

Perhaps your original Boids Implementation could go here?

a

More meaningful discussion of alternatives? If it doesn't really matter that you picked TS, then that's fine too

This should be at the start of chap 4, not here

before

it is safe to remove the player object from this node?

not from the entire game world. I don't think you defined game world in the way you mean here anywhere.

I want more explanation rather than merely deferring to a figure

I think this figure would be more powerful side by side with the earlier one somehow. With a direct comparison in the caption.

Maybe have this much earlier? It makes the ideas you later introduced a bit more concrete.

new sentence

Perhaps mention that this logic could be abstracted away if it were a sort of game engine, where the engine manages duplicated message sending and the user (game dev) just manages input message sending and position handling.

Does it need to be extended to allow for ? Not sure it's the right wording

Need a second comma here

or rework sentence to start with 'Whilst more complex...'

latter is better I think

Haven't defined the player snapshot

Don't think you can use delayed like this

and the

the then makes it seem like it's a if A then B sentence but it's not, the then is part of the A

doesn't make sense. are delayed in arriving at a node?

Don't like starting the para with a connector

Think there's a more succinct and clear way of stating this.

Would it not suffice to send input to any nodes which the player is very close to? Rather than EVERY connected node? This distance could be dependent on the maximum velocity in the game

would be

doesn't quite track to only have a comma here. Need a new sentence or a connector

if the client were to only send input data to ...

node it is connected to

sends user input

that the

Worth having a para on the fact that in 2D systems, you can use congruent squares as in Figure 4.1 but that your architecture generalises to any configuration e.g. hexagons, cubes in 3D or any shapes.

Also worth mentioning that the entire game world must be covered by nodes?

Such that your architecture allows for any collection of nodes whose union of boundaries covers the entire game world. And that the boundaries must be pairwise disjoint.

For example, a single node architecture fits into this conception.

wdym between? Perhaps you mean takes place on both the client and any node ...

To make this tighter, I'd define the two (?) step process of player creation requests more explicitly. Then say something like,

In traditional, single-node systems the player creation request is done a single step.

Also, you could have stated somewhere earlier that traditional architectures can be thought of as a single node. I.e. your architecture is, in a way, more general. It's a nice conceptual connection.

This isn't obvious to me. Is this the malicious message you're on about below?

The paragraphs in this section felt a bit disconnected. Maybe some inline bold subtitles could help? Or qualify that you cover it chronologically at the beginning of the section?

I'd first define the security risk then talk about mitigating it, rather than defining it as an aside

it should connect to

This feels a little like a redefinition of the goals previously stated in terms of client FPS

But

software isn't used to run computers?

??

comma?

again

wdym by strictly?

This seems like a secondary reason to me? Don't you want to do this to maintain that nodes own ONLY the object within their boundaries? I.e. localisation of ownership.

We? All of a sudden?

remove

Syntax highlighting if it's gonna be much closer to actual code than pseudocode?

Again eqn

You've been quite flexible with your defn of margin and boundary. Are they the limits of the area or the area itself? Are they distances?

Not sure this is the right phrase

Repeated this is because

No need for parenthesis

I would connect the boundary defn more directly to the previous paragraph. In fact, I'd define boundary in the previous paragraph without using needing to go via the word section - it's unnecessary. And then define the margin in one paragraph.

Bolding margin twice seems unnecessary to me.

I like this diagram. The node-to-node connection arrow seems a little vague to me though.

would have

Is this true? What if the counter were updated independently on each machine? Obviously, they might not align perfectly with more complicated calculations and isn't practical, I'm not just sure that what you're asserting is necessarily true.

Given that you then use this as an experiment, I would make that clear upfront. Why do you talk about this? Either now or later I'd a sentence or two justifying why you used boids as a simple experiment to demonstrate limitations. Why not another model?

Otherwise, it feels a little out of place.

This doesn't quite do it for me - It's just a sentence. It should link to what you can see in the image a bit more directly, and then how you come to that conclusion.

This isn't strictly true. You could vertically scale. I think you could substantiate the second sentence more precisely to reflect this.

This only makes sense to me because I already know roughly how it works. This line might leave me confused if I didn't already know

This sentence is a bit meaningless to me. Isn't the scaling of of any game world contained within its system? I think I see what you're trying to say but could reword this. I think my confusion mostly comes from the term Distributed Worlds system. It isn't well defined to me.

Again, equation number or something

The core idea is for Distributed Worlds to operate on an ...

seen in Section 2.4 or similar

Object for a Distributed Worlds System?

Also, don't you define the objectives at the start of the next chapt? It's a bit confusing

No but

Requirements?

I'd try to more directly relate the sentences that follow to the word elastic. Not just take advantage, but adapt to the computational demands etc

Rephrase:

This work focuses on proving that ... but does not implement a video game.

why perceived? Aren't you measuring 60FPS, fairly objective?

Bring it to prev page somehow

What do you mean by a performance penalty?

remove comma

no comma needed

is limited in how much it can process in 16.66ms. Recurring?

and even the

remove

always

I? We? This work?

Can't start this sentence with but

Cap?

I think this section would be made more concrete with an example of a video game where it is common practice for players to do this

Cap?

del

clunky

comma

Above? Specify, reference

Again

Haven't defined UPS? Also the updates per second doesn't increase, the update duration does. The updates per second actually goes down?

Earlier you put 16.66 recurring. I think this could be fused neatly with the prev sentence

New sentence

Figure 3.2 compares

Ref earlier section?

remove

Capital?

This section could just be one tighter paragraph I feel.

I would just make it one paragraph about 3.2. It also feels like introducing the notion of 60FPS being the gold standard for your testing shouldn't be here either.

remove

So? Sentence feels a little out of place. Make it make sense in this para

no capital

Slightly too repetitive. Maybe 'We interpret this as meaning that the game world can handle...'

approximately 1200 objects

Figure 3.2 shows that ...

What does the figure tell me? A description is fine but I need you to draw a conclusion from it.

Section 3.2? Again, ref and label

No need for a 'this' here. I would just say : Therefore, the requirement for all testing conducted is that the game must run at at least 60 FPS to be considere ...

Missing parenthesis? You can use latex \ref{section} to reference a previous section which has a \label{section}, creating a clickable link and automatically inserting the correct number

You've already defined FPS, just stick to it or only use frame/s

what do you mean 'a decision about each frame'. Do you mean a decision about its action/behaviour in that frame?

earlier you used present tense ('is created') and now future

I don't know what you mean by this

Full stop?

Too many repeated words from prev sent

This sentence is hard to read

This whole sentence doesn't quite track to me. The 'but' particularly throws me off

disjoint

You didn't mention load balancing above. You just mentioned using multiple servers. I only state this to create a more directly link.

Also, a comment from Amos on the little feedback he's given to me is that 'as above' is lazy.

Perhaps a reference here would be good.

Note that

don't like the phrase

this is your first use of bold which feels out of place. Maybe italics again if this is meant to be a defn

remove even

Remove Well

Any system that needs to efficiently identify objects ... benefits significantly from spatial partitioning. Such system include ...

wherever they are playing from around the globe?

Got a citation for this?

different

italics

Does a boid program need the render part?

Not sure intense is the right word here

is often? Or keep the can be and change to

'can be ... if the graphics are intense'

?

Do you need to put frames/s here when you've just defined FPS

what threshold?

Maybe italicise as it is a new defn

This feels a like a figure/pseudocode. To me, it should have a caption or a number like an equation and be referenced in the text

comma? Or reword to be less clunky. Maybe just While the game is the running, the game loop ...

clunky

haven't defined massively multiplayer yet

What is a continuous section? I'm not sure the word continuous should really be used here?

the?

This shouldn't be how you start the diss. Something more like

The fundamental limitations of traditional multiplayer game architectures ...