faq.html

{{define "<q>"}}<section id="{{.}}"><a href="#{{.}}">🔗</a><h3>{{end}}
{{define "</q>"}}</section>{{end}}

<h2>FAQ</h2>
{{template "<q>" "what"}}
	What is this about?
</h3>
<p>
	For now, the aim is to perfectly reconstruct the lost<sup>[citation
	needed]</sup> source code of the first five
	<a href="https://en.wikipedia.org/wiki/Touhou_Project">Touhou Project</a>
	games by <i>ZUN Soft</i> (now <i>Team Shanghai Alice</i>), which were
	originally released exclusively for the NEC PC-9801 platform.
</p><p>
	The original games being:
</p><dl>
	<dt>TH01:</dt>
	<dd>
		<span lang="ja">東方靈異伝　～ </span>
		The Highly Responsive to Prayers (1997)
	</dd>
	<dt>TH02:</dt>
	<dd>
		<span lang="ja">東方封魔録　～ </span>
		the Story of Eastern Wonderland (1997)
	</dd>
	<dt>TH03:</dt>
	<dd>
		<span lang="ja">東方夢時空　～ </span>
		Phantasmagoria of Dim.Dream (1997)
	</dd>
	<dt>TH04:</dt>
	<dd>
		<span lang="ja">東方幻想郷　～ </span>
		Lotus Land Story (1998)
	</dd>
	<dt>TH05:</dt>
	<dd>
		<span lang="ja">東方怪綺談　～ </span>
		Mystic Square (1998)
	</dd>
</dl><p>
	Since we only have the binaries, we obviously can't know how ZUN named any
	variables and functions, and which comments the original code was
	surrounded with. <i>Perfect</i> therefore means that the binaries compiled
	from the code in the <a href="https://github.com/nmlgc/ReC98">ReC98
	repository</a> are <a href="https://github.com/nmlgc/mzdiff">
	indistinguishable</a> from ZUN's original builds, making it impossible to
	<i>disprove</i> that the original code <i>couldn't</i> have looked like
	this. This property is maintained for every Git commit along the way.
</p><p>
	Aside from the preservation angle and the resulting deep insight into the
	games' mechanics, the code can then serve as the foundation for any type
	of mod, or any port to non-PC-98 platforms, developed by the community.
</p>
{{template "</q>"}}



{{template "<q>" "who"}}Who are you?</h3>
<p>
	I created the <a href="https://github.com/thpatch/thcrap">Touhou Community
	Reliant Automatic Patcher</a> and <a href="https://thpatch.net">Touhou
	Patch Center</a> in 2012, and remained a core developer of both before
	retiring in March 2019. Older Touhou fans might also remember me for
	<i>Touhou Music Room</i> (2010/2011) and the <i>Touhou Vorbis
	Compressor</i> (2011).
</p><p>
	Check my <a href="https://github.com/nmlgc">GitHub page</a> as well as the
	<a href="/fundlog">crowdfunding log</a> here for more proof of my track
	record.
</p>
{{template "</q>"}}



{{template "<q>" "three"}}
	What's the split between <code>OP</code>, <code>MAIN</code> and
	<code>MAINE</code>? What things are they responsible for? How do they
	interact with each other?
</h3>
<ul><li>
	<code>OP.EXE</code> contains the game's main menu and all other menus
	reachable from there. In TH04 and TH05, it also contains the <i>ZUN
	Soft</i> animation and the initial sound configuration menu. As a modder,
	you'd care about this file if you want to add anything involving new
	menus, like practice modes or a replay selection.
</li><li>
	As soon as you start the game, you're in <code>MAIN.EXE</code>
	(<code>REIIDEN.EXE</code> for TH01). As the biggest executable, it handles
	everything happening within a stage. Most notably, boss behavior – unlike
	the Windows games, where this is stored in the separate
	<a href="https://en.touhouwiki.net/wiki/User:Mddass/Touhou_File_Format_Specification/ECL">
	ECL script files, found in the main <code>.dat</code> archive</a>.
</li><li>
	<code>MAINE.EXE</code> (<code>FUUIN.EXE</code> for TH01,
	<code>MAINL.EXE</code> for TH03) then contains code for the ending
	cutscenes, the staff roll, and anything after that. In TH03, it also
	contains the stage start and end screens with the character pictures, as
	well as the cutscenes before stages 8 and 9.
</li></ul><p>
	Since only one executable can be active at any given time, there has to be
	some way of sharing data between them. This is done using a resident
	<code>"Config"</code> structure, kept at the top of conventional DOS RAM.
	{{Blog_PostLink "2020-02-23" "As of 2020-02-23"}}, we've
	reverse-engineered the contents of this structure for all 5 games.
	(<a href="https://github.com/nmlgc/ReC98/blob/e0d90dbdc39941d775d5b38cbef33fd56522d846/th01/th01.h#L43">TH01</a>,
	<a href="https://github.com/nmlgc/ReC98/blob/aa56a7cb18881fbc602c0324b727c1578cac825e/th02/th02.h#L224">TH02</a>,
	<a href="https://github.com/nmlgc/ReC98/blob/484730e31969e55eb590dea6051641532dffb7c5/th03/resident.hpp">TH03</a>,
	<a href="https://github.com/nmlgc/ReC98/blob/042b7802bf070aede7f67d9aaec91f9b8e80c971/th04/resident.hpp">TH04</a>,
	<a href="https://github.com/nmlgc/ReC98/blob/042b7802bf070aede7f67d9aaec91f9b8e80c971/th05/resident.hpp">TH05</a>)
</p>
{{template "</q>"}}



{{template "<q>" "pi-what"}}
	What is this "position independence" thing about?
</h3>
<p>
	Position <i>dependence</i> means that a binary's references to global
	variables are expressed as raw number constants, rather than being named
	with identifiers:
</p><pre>mov ax, some_data   ; Position-independent
mov ax, 1234h       ; Position-dependent; assumes that
                    ; [some_data] is at address 1234h
</pre><p>
	If you increase or decrease the number of bytes anywhere in the non-header
	parts of an executable, you'll end up breaking most of these
	position-dependent references, since global variables no longer are where
	the game expects them to be. This will lead to quite some instability.
</p><p>
	Now, why is this such an issue for PC-98 Touhou? 16-bit x86 code has to
	take <a href="https://en.wikipedia.org/wiki/X86_memory_segmentation">
	segmentation</a> into account for all its memory accesses. This means that
	each actual address is built out of two 16-bit values, the <i>segment</i>
	and the <i>offset</i>. Since offsets therefore can only range from 0 to
	2<sup>16</sup>-1, the line between actual memory offsets and numeric
	constants becomes blurred. Most disassemblers I know of that target this
	architecture therefore only do a very superficial attempt at identifying
	data references, and give up once arrays are involved, just leaving a
	numeric constant in place of such a reference. And for good reason: Doing
	this properly effectively requires an emulator, running the game and
	performing control flow analysis. Anything less than that –
	<i>especially</i> anything parsing individual lines of ASM – and you're
	bound to
</p><ul>
	<li>miss addresses that are <i>calculated</i> from other constants</li>
	<li>mistake function pointers for data pointers</li>
	<li>mistake pointers into special memory segments (graphics RAM, BIOS RAM,
	etc.) for data pointers</li>
	<li>or confuse <i>actual</i> numeric constants for memory addresses.
</ul><p>
	And even <i>with</i> an emulator, you're still faced with the fact that on
	the low level of ASM and C,
	<a href="https://www.youtube.com/watch?v=i2fhNVQPb5I&t=43s">the declared
	size of an array is simply advisory anyway</a>. So what do you put,
	especially when being confronted with out-of-bounds array access bugs in
	the original game itself?
</p><p>
	So if you've chosen to deliver quality instead of delivering experimental
	research, the best choice is to give up, don't pretend to be
	position-independent in the first place, and treat <i>every</i> numeric
	constant that falls within the range of any data segment as a possible
	memory reference. Sure, this means that the <i>actual</i> number of memory
	references are lower, and thus, the actual percentage of
	position-independence is higher than the front page may suggest. But
	<i>we can't tell</i>, and erring on the side of caution is, in my opinion,
	better than pretending that the code is more position-independent than it
	actually is, just because it ran through some sort of experimental tool.
</p><p>
	While position-dependent code is still a significant step up from modding
	game binaries via hex-editing, it effectively still suffers from most of
	the same constraints, despite looking like regular source code that you
	can just arbitrarily edit and recompile. So while modding the game in all
	sorts of ways is definitely <i>possible</i> right now, it's definitely
	harder than it needs to be. Once a binary reaches 100% position
	independence though, developing any sort of mod, in either C/C++ or ASM,
	will become trivial.
</p>
{{template "</q>"}}



{{template "<q>" "pi-how"}}
	How is position independence calculated?
</h3>
<p>
	The <a href="/progress">absolute number</a> is the sum of all remaining
	hexadecimal literals in all code segments of a binary's big
	<kbd>.asm</kbd> dump file that fall into all of these categories:
</p><h4>
	1) Matches the regex <code>(-?[0-9][0-9a-fA-F]{1,4})h</code> for
	hexadecimal literals in TASM/MASM syntax
</h4><p>
	IDA dumps all number literals ≥10 as hex by default. Restricting the PI
	calculation to hex numbers allows us to clearly mark false positives by
	simply converting those numbers to decimal. Having to do this manually
	further communicates that every such conversion was a conscious decision,
	based on the newly RE'd context the number is used in.
</p><p>
	This might seem like useless work at first, only necessary because it's
	dictated by some counting algorithm on a website. However, most of those
	false positives turn out to be things like (sub)pixel coordinates, number
	of score points, frame counts… which the typical person <i>does</i> prefer
	and expect to be expressed in decimal. Thus, this conversion turns into
	quite a quality-of-life improvement for anyone reading and modding the
	code. <i>Especially</i> with the fixed-point 12.4 "subpixel" type used for
	playfield-space coordinates in TH03-TH05, which we can abstract away even
	at the ASM level.
</p><h4>
	2) Falls within the <a href="https://github.com/nmlgc/rec98.nmlgc.net/blob/43a9f1d51602eeb224a55ffeb384f2ad12d20798/rec98repo.go#L160">
	data segment ranges occupied by ZUN data</a>
</h4><p>
	This means that <i>structure size independence</i> is an explicit non-goal
	of PI. The reason becomes clear if you look at all the things a 16-bit
	number literal can represent:
</p><figure>
	<embed src="{{call $.StaticFileURL "faq-pi-ranges.svg"}}" alt="Value ranges" />
</figure><p>
	If we don't limit the value range to ZUN data,
	<a href="https://en.wikipedia.org/wiki/Benford%27s_law">all the low
	numbers</a> would vastly drown out the actual memory references we are
	trying to identify, resulting in a number that's even less representative
	of the code's actual position independence. And since structures can have
	any size, we'll necessarily have to leave them to reverse-engineering.
</p><h4>
	3) Not the argument to any x86 instruction unrelated to memory accesses
</h4><p>
	These include
</p><ul>
	<li>the I/O port instructions <code>IN</code> and <code>OUT</code> (which
	take architecture-specific port numbers),</li>
	<li><code>INT</code> (which takes a software interrupt number),</li>
	<li>and <code>ENTER</code> (which takes the size of a function's stack
	allocation for local variables).</li>
</ul>
{{template "</q>"}}


{{template "<q>" "why"}}
	Why crowdfunding?
</h3>
<p>
	I'm trying to steadily grow this project into an actual job so that I can
	spend more time working on it. Even though I don't particularly <i>like</i>
	Touhou these days, I can do a much better job here than in any corporate RL
	software development position, where I am typically limited by people, dumb
	tech stacks, and the fact that ReC98 is just a much more interesting project
	in general. The demand on part of the fandom is also clearly there, as
	evidenced by the success of this crowdfunding and this store being sold out
	for almost the entirety of 2021.
</p><p>
	Another advantage: It's <i>you</i>, the patrons, who then get to choose
	which game I focus on. This has always felt wrong for me to decide, and I've
	never had much of a preference for a specific game to begin with.
</p>
{{template "</q>"}}



{{template "<q>" "automate"}}
	Can't a machine automate all this work? It all seems very blue-collar and
	mechanical.
</h3>
<p>
	Maybe. While it would have been an option to collect lots of money for
	developing an automated decompilation solution, that would have been a
	huge risk, and my previous attempts at it failed spectacularly. In
	contrast, selling small chunks of progress for an hourly wage leads to a
	stream of tiny, but immediate results. It may take longer in the end, but
	even partially reverse-engineered game code can be a tremendous help to
	modders. Also, naming variables, contextualizing numeric constants, and the
	resulting insights into the game mechanics, is something you simply can't
	get out of an automated solution.
</p><p>
	Consider this piece of ASM:
</p><pre>; Somewhere…
	mov	byte_2CEC2, 40

; Somewhere else…
	cmp	byte_25351, 0
	jz	@@return_from_function
	; …
	cmp	byte_2CEC2, 0
	jz	@@down
	cmp	byte_2CEC2, 32
	jbe	@@return_from_function
	mov	byte_2CEC2, 0
	; …
@@down:
	dec	byte_25351
</pre><p>
	Now, I could simply decompile this into
</p><pre>// Somewhere…
	byte_2CEC2 = 40;

// Somewhere else…
	if(byte_25351 == 0) {
		return;
	}
	; …
	if(byte_2CEC2 != 0) {
		if(byte_2CEC2 <= 32) {
			return;
		}
		byte_2CEC2 = 0;
	}
	byte_25351--;
</pre><p>
	However, that doesn't really tell you anything that you couldn't already
	tell from looking at the assembly. After manually reverse-engineering
	the meaning of these variables, we learn that
</p><ul>
	<li><code>byte_2CEC2</code> is the <i>miss countdown</i> – set to 40 if
	you collide with anything, playing an explosion animation while it's
	nonzero, then decremented each frame, and actually removing the life once
	it <i>reaches</i> zero</li>
	<li>and that <code>byte_25351</code> is the number of bombs in stock.</li>
</ul><p>
	And lo and behold, we just proved the existence of an 8-frame deathbomb
	window,
	<a href="https://twitter.com/ReC98Project/status/1176592461155713026">
	ending up with an insight that's immediately valuable to many fans</a>.
	Finally, let's define some symbols:
</p><pre>MISS_FRAMES = 32
DEATHBOMB_WINDOW = 8

; Somewhere…
	mov	_miss_time, MISS_FRAMES + DEATHBOMB_WINDOW

; Somewhere else…
	cmp	_bombs, 0
	jz	@@return_from_function
	; …
	cmp	_miss_time, 0
	jz	@@down
	cmp	_miss_time, MISS_FRAMES
	jbe	@@return_from_function
	mov	_miss_time, 0
	; …
@@down:
	dec	_bombs
</pre><p>
	And suddenly, it becomes both obvious <i>and easily moddable</i> to
	whoever reads the code, even while it's still assembly. <i>This is the
	level I operate at.</i> Decompilation only becomes mere syntactic sugar
	at this point.
</p>
{{template "</q>"}}



{{template "<q>" "duration"}}
	How long is this crowdfunding campaign going to run?
</h3>
<p>
	Indefinitely – and that's the beauty of it. Whenever someone is
	interested, they can insert a coin, and see how that money gets turned into
	tangible progress towards their goal of choice. Effectively, this project
	will run for as long as the market deems it valuable. Maybe we get enough to
	complete one game, maybe we won't. Maybe there will be no interest
	whatsoever for a few months, and then a small number of big transactions.
	Who knows.
</p><p>
	In a way, this is therefore closer to art commissions than it is to your
	typical video game crowdfunding campaign.
</p>
{{template "</q>"}}



{{template "<q>" "prices"}}
	How does pricing work? What is a "push"?
</h3>
<p>
	A push is a reasonably long stretch of work towards a given goal, currently
	sold for {{HTML_PushPrice}} a piece. You can purchase any partial amount of
	that sum of money though – and definitely should, if only just to signal
	your interest in a particular goal to the wider community and maybe drum up
	more support for it. However, any goal requires at least one fully funded
	push first before I start working on it. This approach works well with
	reverse-engineering, as it ensures that I get to concentrate on newly RE'd
	code for a while, leading to more accurate picture of the details and
	interactions and keeping the high standard that this project has developed
	over the years.
</p><p>
	Smaller stretches of work do make sense for modding-related goals and Seihou
	though. For these goals, you can bypass the regular push system via
	<i>microtransactions</i>, and get immediate deliveries of any small piece of
	work without the upfront investment of a full push.
{{template "</q>"}}



{{template "<q>" "dosbox-x"}}
	PC-98 emulation is getting better and better, DOSBox-X even has dynamic
	recompilation now. Are source ports of a single game series even worth it?
</h3>
<p>
	Again, you decide.
</p>
{{template "</q>"}}



{{template "<q>" "recomp"}}
	The N64 community is doing <i>recompilations</i> now. Aren't decompilations outdated?
</h3>
<p>
	I’d just like to interject for a moment. What you’re referring to as "N64 recompilation", is in fact, a combination of five different efforts developed in parallel by the N64 community over many years:
</p><ol start="0">
	<li>Emulators that define the hardware's behavior in software and can accurately run the game</li>
	<li>A renderer that takes N64 display lists and microcode and produces modern HD images</li>
	<li>Bugfix and quality-of-life patches developed as a result of reverse-engineering the game</li>
	<li>A runtime that splits apart the individual subsystems of N64 hardware (i.e., graphics, sound, and peripherals) into standalone modules with their own APIs that can run independently from the CPU and the hardware's interfaces</li>
	<li>The raw decompiler from MIPS assembly instructions into (unreadable) C</li>
</ol><p>
	The bare minimum features required for such a recompilation port are 3) and 4). Recompilation in the pure meaning of the word – raw translation of one assembly language to another – provides exactly one benefit to players, and that's reducing system requirements by cutting out CPU emulation. The CPU is an essential part of a console, but useless by itself; it can only play games in the context of further graphics, sound, and input hardware, provided by the runtime.<br />
	So if all you had was a recompiler and a hardware runtime, you'd get… exactly the original game running at the original resolution and the original frame rate, just maybe with reduced loading times. If all you care about is <i>a</i> native port, recompilation does indeed accelerate the process by freeing port authors from the need to bother with game logic. But from a player's point of view, such a port looks barely any different from the game being wrapped into a standalone emulator. And the further back in gaming history you go, the less value such a recompilation can provide, as typical emulators for older systems have long been able to run these games at full speed on modern hardware anyway.
</p><p>
	Thus, it's actually 1) that delivers most of the perceived value of these ports. So much of it, in fact, that branding them as <i>recompilations</i> takes all the justified hype and excitement about how all these community efforts have been coming together, and redirects it to its least important component. A glance into the comment sections confirms that the average fan now thinks that <i>"recompilation"</i> means <i>"magic instant porting box for a system's entire library"</i> and is already expecting legal action from Nintendo, without understanding that this is legally no different from regular emulation. But how can you blame them if everyone in these videos is saying <q>no emulation</q> as if, again, the N64 solely consisted of its CPU?<br />
	From what I can tell, recompiling the game wasn't even necessary to achieve the high frame rates, which instead are the result of <a href="https://github.com/rt64/rt64/tree/e4af592195bf1903df9d58e2b350e8183fb0e9ef?tab=readme-ov-file#architecture">lots of clever processing, optimization, and interpolation in the renderer itself</a>. This code always runs natively on the host CPU and GPU regardless of whether it's an <a href="https://x.com/dariosamo/status/1532736513183731713">emulator plugin</a> or linked into a recompiled game. Meanwhile, even a recompiled game has to run its logic at the original frame rate to avoid bugs.
</p><p>
	<img
		src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAIAAAADACAMAAAD7uJuVAAABmFBMVEUAAAD//////+///9T//8j39////6r//4vq9+r//3///2T//1Xp6f/391PH6cfo6GTU1P/p6VHp6E204rS/2L/Fxf/ExP/Bwf+/v//U1E2/v7+5uc+/v3+P1I/FxUqqqv+/v0murr+wsFecmuiamuiqqkVnxWeTk79Xv1ecmj5/f/+Df79/f79/f6o7tDuHf39/f39/f3ODfzNzc4AgqiAdox0/iz9VVf8emh0YmBgVkxVRUchVVaJdVX9VVX9ISMcQhxAcfxg/P/8YfxhdVSI/P9RcURsKewoIdggqKv8CawI/Pz89PT8qKtBBNlEAZQAYVRA2NlEQVRAAWxYBWwBBNhYAWwAASz8ARFEWFv82Kj8ESwAqKj8AMn82KhEFRAA0KBAFQwAPD+4zJg8VNgoAIK0AGb8IMgAACegUKggAMgAAAP8IKggBAOgAAOgEAL8KIQAKIAAAAL8FAK0FAKoAIAAMGQAIAH8MFgAEGAAAAH8KAFEMAD8OCQAAAFEOAB8AAD8OABYQAAAMAAAKAAAIAAAFAAAEAAAO47O+AAAAAXRSTlMAQObYZgAACMVJREFUeNrtnP9/00Qcxhd1Oi1sE0XHZFZlRaagODdxMJDhkDmUA8cXtW5TFMqMtYLaqGsFrtb82+aauz5JPvay5Nrmlzy/NOuryfPu++6SNmyM5MmTJ0+ePMY5N3st2/5icTbbfi/tjPuLvJ1tf5GDIJP+IgfBkPsBwNtZ9AMgC4JvigBwnAwIjgPAsW0QZCDglG2DYJgC0A+CDAS8Zcs0PAI3AwGfKoCaB9DKSgAUZCEgAwXniAAoGFY/BAxTAfohYKgK0A8Bw1SABQABZyYmzoBgGGej42EBE4XChE0IhiYAAIRgOAIwBIRgSAKQEycCBO7QBKC/UAgQtIYmAP1elv3tKhQMQwD6C6UOQXWXD0zBuR4CLhZ8gNIWV3EH2E8FHFYA81xmMAqO9BDwUUHk+ddKQQX/DgDgCghCAt7YL+uFAjjo3yjwQL5+m56ESsj8btXhwfQVAAg4DUcAtqq27fQboM0jORu5EK1CALftMEG7n/0gCF+KeZdgi9v9J7j8ujVZ9HK2c/BaQxGcslV4IOqjUbGTSeuVy6b109aU6BcEti0RQNADoMaLkmDKmjZCOGlZ+4oqm7LQkVPR6QIhjZqvoHvS2mdZJ436ramiCgNBB4E7qp4g1EYUwZRlQHDZ23lazcXmNusSNLg+7cibSDsK09i3zXmTdQlqPBnBtIEArMUK83LrLlEQSzBpWZ8ZCUC/RCAK9ATFydE3UwFYXjobLue8zgK5fp+rVPxtcQ1gMuU69+ICwIsZQItzXmYhgm6/3BbXAMaCCC1zgFHLGoUABAD1DpbYEueETYZUAgpm097QPjg6erCHgPuo9wEoARRcm011Sx/quBdMwOou3j0AKMGPIyq2bT8wB8ASrKIeABoCAfDIHAAnoYZak7ANgru3mMpN/yiPHkCAEcAm6tFPCIBw83dZDwEmANxBPb+6EMrKDgcBgvw1Yg7goB79FMEx66ffA454Z4GWvMQ6HP0kV3ExjuTBI4PvAUeuyHpdPwiAYF6vTkB//2yLRPqPjVvdjB0Vz1jWu5zE5FuaOgX/b/8LVjBjPkBPAtdIQRWLD+/fCmVGAlhXzQVQBbtVVY/xHyf9PgDmgUaAwZcSzL/zjH3wYrcfACQf/mDSDgLazwSC6gcAJejjF9OdlQUZJjNjjR9Tz435k5HEGIDWA2DhHTx3dGzsaJ8BkE49SlbExYet0bLT6+unBwLQqYfmHa+fsyUKsM7Y+iAAvlgITbQV7qUCAeGh6TMA+gGwI/ohYOAAPy2EAVb8j8hiPnjGT/tzY6YLsNT5GdcJc4CPIwA7/kdkMR+YHHNve0YBrPk/dwlM69u+AExCKUABMAUmCJhU0EeANvcFYK1LAXIIACAImFRgMARui0fir2/vuHLWcS+qeLvZ3BZPj0sC1sn58CRsJ7h1iXpErm8mEKIATc6ba7g8zzBGCdAfjyBhKQDDkRcVwKLaXpIfUPw5AVQ/58mX9iQEAFAEi5cUwKVFub2mPqKJOQEE2c9IfxICACASANswviTmBIskQT8IEgGAYE3MCR1AO806EGOtA4DwJflzOBXSnxBBjLUeQCGsEQAQoD8xAouk3jkRAaBM4OgeIqlPSCyScuB2Dce2BqCcBgCTkWkUcCgAAE29Df+J+znTKOBQoAMoj7ST9ruy3anuMi9f3qvu1stUQR0KCMB2E7dx0nwlU1/2q8zLPe+R81+IgjIUEIAmbmQlB3BVvxfmRTyKt6pVQJeoQriZXAD6gwB1nQIKoPLHn+m+jDTsEECDrH+qgAIgyQFqYYAaXf9Q0HcACLA3GNuwlYLv5H1iooAAVMwAIMC+s7FxB/9IUi/L+8REAQDQbwRgkzg8nIoPgG0AbNLXuwYAOoI6tgGwSV/PW+YAuFUDgjK2JcB1zz+IiQIjAIpQqQe2fYD71yscztIpcLEItBiIg+fC8E4qBS25CDT5LdwPLAAQgpaBAJoG6fcSAqAErrkApEb6CQAhaPVRAMp4g2ohsqDAWAAFqFEtvWS1zAUgEEDfrK1RYCgAgQD6Zm0DBS4OoE0NAuiMI7AJFLQ0AFQrleXoAVrxAkRSCQDBYwJAFOgF8DQCQPD4Vw1Aay8C6NuK/vKqg3MQCdnvwCecKNALoG8LvzcW30/3O8CJguQCAIB+bqcA4G5iARgC9AMg4RDoFUAAClEsgv4bc3M3SFuP/dTrYxVIAVCORzL/5kqlOdLeYz/1+lgFEEAOhH4zAL0CCKAq0W80BHoFuomFfiRUqgmPJBUA+pGg9r4BOLr+3gCFqQuaHRMAgCC+H0PgEViWBsFJADBfWraRC1NPPVfwsv/Vksz8FgHwMmX5eeKZQiiH5F6rewfYKoHggndc0S8ISiWKEAC1FEFBSxAPIBSU3vtc1QcBKIKNvCQBni5oCOIBpAKBIOpFMAQUAf1fvYx+PUE8gPot5UORieV4bJ4dZFUCXDwsenwAvB5ZBkE8AAgOYWlhIocRVlW9COo1BAkARP+JyPdhgnDbO7hf/+yTqKdZBUEcAO1HPUV4//ZyidbTcBDsAWBVvtAh1QQB8dcNNeZ/QOVdgq3YixH6ST1B0NdLAh9AEczHX47npCj8oYIOQV8v34cTeP33sR9Ibsz5A4U/1dAgaOvRj9fPfUsF0Lv0NI4GxiH1pH/PH0rd3h2Oth/1cf0QoFNAO3oTVHX3zhIJgAJ6HA3BbpXUa/dzk/zLLTxCJqKfL9ivQQQkIpDVlCBmvmC/GhGQ4B9u0QybMf0AtuXPD3k3/6BIj0AJ6Hjq5ov/CPwUf30GjMjBMM6a+RKFRnVCBBDABcZZM18wbCZ/+temB8ajdvwxcTH1jQigFo+68cfSxdQ3I6AHx3zQQUKAOQHVS9Y9eZ2BAExFRVBroCgeqFETj1j75ggPyXzQDAl2wAI0REBQqp+UqO8zAtWP/oHUUwTMB4z7AOspAuYDxn1w9RRDG+MCAwQAZIEA8RkgoD7/n/Ly5MmTJ0+ePHny5MmTPP8BNdQbXdABhZkAAAAASUVORK5CYII="
		alt="Mima's sprite from TH05, 4× upscaled using the classic XBRz algorithm"
		style="float: right;"
	/>
	And <i>that's</i> why we're now seeing all these developments for an early 3D console, and not for an undoubtedly less complex 2D system. If graphics are defined as geometry data and textures, it makes intuitive sense to rasterize it all at higher resolutions and wider aspect ratios, and then go the extra mile towards frame interpolation. But due to the custom way that each 2D system implements its graphics <i>and</i> the already rasterized nature of each game's artwork, you can't just simply flip a switch and have such games render at a wider aspect ratio. The most intuitive game-independent renderer feature for those systems, therefore, is the same old filtered upscaling that emulators have been doing for decades already, such as the classic XBRz algorithm used for the sprite on the right.<br />
	Sure, maybe you might want to develop a fancier spin on the same concept by filling the new pixels with their statistically most likely value according to a training set of similar games (or, as it's more commonly referred to, "AI upscaling"). Even AI frame interpolation could be possible if someone puts in the time to develop a {{Blog_PostLink "2024-04-24" "rollback mechanism"}} that would predictively render multiple frames ahead and somehow detect sprites from logged VRAM writes. But all of this can be done independently from ReC98, within just an emulator. If this had been our goal, we wouldn't ever have had to look at any Touhou game code.
</p><p style="clear: both;">
	This brings us to 2). On the surface, recompilation simplifies modding a lot – just add new C code before compiling, without any ROM size constraints or ASM to worry about, right? Except that the generated code is as readable and unannotated as a raw disassembly, which brings you back to reverse-engineering the game <i>anyway</i> to figure out what's going on. A recompilation might provide a more efficient starting point if no one has done <i>any</i> RE or classic decompilation on the game yet, but doesn't change the fact that <i>some</i> RE will <i>always</i> be required, if only to work around the inevitable issues introduced by 1).<br />
	Note how all the YouTube videos that presented these ports in May 2024 focused on Majora's Mask, a game that had been <a href="https://zelda64.dev/games/mm">94% decompiled</a> by the time its recompiled port was released. Hence, they already knew where the patches needed to go and how they would look like. These videos rarely explore this technology in the context of other games, and if they do, the results are not nearly as impressive.<br />
	In short, the added value of a recompilation will <i>always</i> be proportional to the reverse-engineered knowledge about a game.
</p><p>
	And for now, this has been ReC98's trajectory – reverse-engineering PC-98 Touhou to document every detail about these games and how they run on PC-98 hardware, storing the RE'd knowledge in the form of a perfect decompilation that has yielded identical binaries from Day 1, and facilitating bugfixes and mods way before it's viable to port these games away from PC-98. This was the most immediately useful choice in 2014, and continues to be the most useful choice in 2024 as the PC-98 hacking scene has only significantly grown in the emulation and translation patching departments since then. Besides, both RE and bugfixing are quite tough already if you're faced with an underdocumented platform running engine-less spaghetti code written by a beginner programmer. It's {{Blog_PostLink "2023-03-14" "so easy to get it wrong"}} even <i>with</i> a 100% decompilation, and requires great care to not accidentally modify game logic. After all, ReC98 can only ever earn the trust of the {{HTML_TagInline "gameplay"}} community if its bugfixes retain hypothetical replay compatibility. Recompilation does not help with any of that.
</p><p>
	That's not to say that ReC98 won't <i>ever</i> work on those other prerequisites of recompiled PC-98 ports. In fact, I hinted towards that in {{Blog_PostLink "2022-08-15" "the announcement blog post for TH01's 100% decompilation"}}:
</p><blockquote>[…] maybe you can just merge in a PC-98 emulator core and get started with something halfway decent in a short amount of time.</blockquote><p>
	If we wanted to have the PC-98 version of 3), this is the route we should have followed. Building a PC-98 runtime that can be easily linked into a decompiled PC-98-native code base is the hard part – once we've solved <i>that</i>, we could probably leverage <a href="https://github.com/M-HT/SR">an existing x86 recompilation project</a> to apply this runtime to other PC-98 games beyond Touhou. Instead, my backers have invested exactly {{HTML_Currency 100}} into portability in the 21 months after TH01's 100% decompilation. So it would seem that people agree with the original goal of first getting <i>all</i> 5 games to 100% before starting any porting work?<br />
	In the end, it's all down to the priorities of each individual backer, including you.
</p>
{{template "</q>"}}



{{template "<q>" "uth05win"}}
	Isn't <a href="http://m.newsmth.net/article/TouHou/single/11992">
	uth05win</a> already what you wanted to achieve? We even have
	<a href="https://github.com/Wintiger0222/uth05win">source code</a> for it.
</h3>
<p>
	Initially, I thought the same, and had the impression that uth05win's
	source code release would immediately obsolete ReC98.
</p><p>
	Fast-forward to 2022 though, and TH05 has been one of the most requested
	games among ReC98 backers. uth05win <i>did</i> legitimately reverse-engineer
	most of TH05, and it definitely was a tremendous help during the initial
	reverse-engineering phases of not only TH05, but also TH04 and, to a lesser
	extent, even the previous three games. However, the final port has taken
	quite some liberties, ranging from
	{{Blog_PostLink "2020-08-16" "fanfiction fixes for even just minor inconsistencies within ZUN's original code"}}
	to
	{{Blog_PostLink "2022-04-30" "flat-out wrong code in certain boss scripts"}}.
	It's completely understandable why die-hard PC-98 Touhou fans immediately
	dismiss it as <i>"not the real thing"</i>. Which, ironically, led to ReC98's
	approach of a provably legit source code reconstruction being appreciated
	<i>more</i>, not less, among this group of people.
</p><p>
	Also, the obvious reason I don't restrict myself to just one game.
</p>
{{template "</q>"}}



{{template "<q>" "identical-code"}}
	Why do pushes that are geared towards one specific game also tend to come
	with progress in other games? Aren't you wasting time there by not focusing
	100% on what your patrons wanted you to do?
</h3>
<p>
	If the same function appears in more than one game, more or less
	unchanged, I'd only be wasting time re-familiarizing myself with all the
	involved concepts months later. I think it makes more sense to immediately
	cover identical functionality in all games. It's basically free progress
	for everyone else.
</p><p>
	Then again, the more progress is made, the more infrequently this will
	happen, as the amount of not yet reverse-engineered code shared between
	the games approaches zero.
</p>
{{template "</q>"}}



{{template "<q>" "refunds"}}
	Do you have a refund policy?
</h3>
<p>
	Yes! You can request refunds for every push I haven't started working on
	yet. I <i>will</i> keep the money after having delivered a push though.
</p>
{{template "</q>"}}



{{template "<q>" "mod-bugs"}}
	I found a bug in one of your mod releases!
</h3>
<p>
	Please tell me! I will release a bugfix for free, together with a short
	explanatory blog post, if the bug is
</p><ul>
	<li>a clear regression (i.e., I broke something that worked fine in the
	original game), and</li>
	<li>not actually related to a bug or missing feature in the original game
	that would be addressed in a future push or microtransaction.</li>
</ul><p>
	An example: The <a
	href="https://github.com/nmlgc/ReC98/releases/tag/P0234">P0234 build of the
	TH01 Anniversary Edition</a> had a bug during the transition between
	Sariel's two forms that caused certain pellets to be rendered as white
	streaks. This was due to a quirk I didn't keep in mind when writing the new
	pellet rendering code that this release was all about, so I
	{{Blog_PostLink "2023-03-14" "fixed it shortly after I received the bug report, for free"}}.
</p><p>
	Two counterexamples:
</p><ul>
	<li>The same P0234 build broke the <code>s</code> command-line parameter
	that would start TH01 with just a stage selection and no debug features. A
	quick investigation showed that it only ever worked in the original game if
	a specific byte in memory was written to before launching the game, which
	turned the issue into a ZUN bug. But since it was easy to fix, I prioritized
	the issue after the bug report, and shipped the fix in the <a
	href="https://github.com/nmlgc/ReC98/releases/tag/P0239">subsequent P0239
	build.</a></li>
	<li>The <a href="https://github.com/nmlgc/ssg/releases/tag/P0226">P0226
	release of Shuusou Gyoku</a> added 32-bit rendering support to allow the
	game to run at full speed on Windows ≥8, but did not include a windowed mode
	or any scaling support. This is a separate feature that requires separate,
	dedicated attention, and <a
	href="https://github.com/nmlgc/ssg/issues/7">already had an entry in the
	issue tracker</a>.</li>
</ul>
{{template "</q>"}}



{{template "<q>" "tsa-takedown"}}
	Can't Team Shanghai Alice take down this project and crowdfunding at any
	time?
</h3>
<p>
	While I can't promise that they <i>won't</i>, the same kind of source code
	reconstruction has been done for the
	<a href="https://github.com/pret/pokered#see-also">Generation I-III
	Pokémon games</a>, <a href="https://github.com/n64decomp/sm64">Super Mario
	64</a>, <a href="https://github.com/zeldaret/oot">Ocarina of Time</a>, and
	<a href="https://github.com/diasurgical/devilution">Diablo</a>,
	all of which still generate revenue for their rights holders. PC-98 Touhou,
	on the other hand, is both no longer sold <i>and</i>
	unlikely to be ever sold again in its original form due to
	{{Blog_PostLink "2019-11-06" "various copyright infringements in the games themselves"}}.
	This fact should make PC-98 Touhou an inherently safer choice for a decompilation project than any of the ones above.
</p><p>
	Despite all that, Team Shanghai Alice <a
	href="https://github.com/github/dmca/blob/a695b94/2022/04/2022-04-11-touhou-project.md">did file a DMCA takedown against full downloads of the PC-98 games back in April 2022</a>. At the time, this appeared to officially rob the games of their perceived abandonware status, but I've heard about no other similarly high-profile takedowns in the years since. Hence, it still remains unclear
	whether they plan to extend their copyright enforcement to the source code
	and research level this project exists at. Without a precedent inside
	the Touhou scene, ReC98 does seem safe for the time being – especially since
	it has never included any asset data from the original games, and is
	unusable without supplying that data from existing game copies. If anything, this takedown has made ReC98 <i>safer</i> to operate in the years since. As this project was left alone at a time when the pirates weren't, the unspoken agreement seems to be that this project can continue as long as it doesn't increase the number of functional game copies out there.<br />
	Takedowns of decompilation projects <i>have</i> happened outside Touhou
	though, most notably with <a
	href="https://github.com/github/dmca/blob/3ce1c94/2021/02/2021-02-19-take-two.md">Take-Two's
	DMCA claim against the GTA 3 / Vice City project</a>. Unfortunately, <a href="https://storage.courtlistener.com/recap/gov.uscourts.cand.384429/gov.uscourts.cand.384429.47.0.pdf">that case was settled out of court</a>, which still leaves decompilation projects as a whole without a clear legal precedent. Thus, Team Shanghai Alice might very well try to do the same, cementing its current trajectory of ever more corporate conduct and tighter control.
</p><p>
	That said, it would take quite a bit more than a simple DMCA claim to GitHub
	to take down this project. Everything about it has always been self-hosted
	outside the US, and the GitHub presence of both the game code and website
	repositories only fulfills four reasons:
</p><ol>
	<li>Providing a nice code and commit browser</li>
	<li>Offering another expendable place for issues to be reported, in addition to my usual Internet presence</li>
	<li>Easy discoverability</li>
	<li>Participating in the ⭐ star count popularity contest</li>
</ol><p>
	Yes, no "web hosting" on this list, and no essential reliance on GitHub
	infrastructure anywhere else. 1) could even be implemented as part of this
	website in a push or two, if you all consider that a worthwhile thing to
	have.
</p><p>
	There is certainly an argument in favor of taking down the project at the
	first sign of resistance. Why continue working in the hostile environment
	that is canon Touhou if the perceived "free culture spirit" has been nothing
	but a misunderstanding from the very beginning, and the rights holders are
	more corporate, controlling, and distant than actual big corporations? The
	rational choice would definitely be to leave the sinking ship. I'd love if
	there was enough money in the non-Touhou parts of the PC-98 scene for a
	follow-up project, it would be a shame to let my experience on the platform
	go to waste. <a href="https://www.mobygames.com/game/pc98/rusty">Meme</a> <a
	href="https://www.mobygames.com/game/pc98/possessioner">games</a>, <a
	href="https://www.mobygames.com/game/pc98/sex-2">anyone</a>?
</p><p>
	Or maybe ReC98 will just be all about Seihou then. Its fans have been <a href="/blog/tag/seihou">incredibly generous</a> in the years since the takedown, pushing Shuusou Gyoku from its aging and fully Windows-native codebase to {{Blog_PostLink "2025-01-25" "a fully native Linux port"}} with barely any compromises on quality. Thanks to pbg licensing all Seihou source code under MIT, everything about this is legal, thus making Seihou an objectively much more sustainable investment of your money.
</p><p>
	But that's not what the non-Seihou backers signed up for. I
	have several ideas for transforming the Touhou side of the project after a takedown notice
	while still keeping its essence. I <i>will</i> keep it running as long as
	possible – even if that will someday mean that I have to manually send out
	the source code to people. And the blog – arguably the main attraction while
	development is still ongoing – should be, in theory, even safer than that.
	Let's wait and see how far Team Shanghai Alice will actually escalate this.
	There <i>are</i> risks, and you should be aware of them and invest
	responsibly, but I'm far from panicking.
</p><p>
	Until then (and let's all hope we'll never reach that point): Always keep in
	mind that <i>the product is both the code and the documentation, in the form
	of new commits in a Git repository. Nothing more, nothing less</i>.
	Perform a <code>git clone</code> after I pushed the commits you bought, and
	you now have a DRM-free digital copy of the progress you paid for. Even if I
	have to start manually sending out the source code to people, rest assured
	that nothing I produce will ever be put behind a paywall. (The only thing
	that <i>is</i> behind a paywall is the time it takes to make it all happen.)
</p><p>
	And finally, because it seems to be frequently misunderstood: I have
	<i>never</i> sold the promise of a finished Windows/Linux/phone release of
	any of the games, and <i>still</i> don't sell the promise of any Team
	Shanghai Alice release in relation to the completion of ReC98. Sure, in a
	fair world, Team Shanghai Alice would leave this project alive until it
	reached its core goal, and then acquire and commercially exploit it. They
	have every right to do this, and it would be fine by me, as I will have been
	paid my fair share at that point. But throughout all its Windows history,
	Touhou has always been a poster child of the <a
	href="https://youtu.be/qW61xJNJ9m8?t=20">"not invented here" mentality
	prevalent in Japanese business</a>. Even the currently unlikely event of a
	takedown is much, <i>much</i> more likely than them acknowledging or even
	using my work. Heck, I remember hearing about offers for professional
	localization from people who are much closer to ZUN than I am<sup>[citation
	needed]</sup>, and they've all been turned down…
</p>
{{template "</q>"}}



{{template "<q>" "pull-requests"}}
	Can I still help out with the reverse-engineering by contributing to the
	ReC98 repository?
</h3>
<p>
	The amount of time I spend on raw reverse-engineering and decompilation
	almost pales in comparison to the deeper research and documentation work
	that this project evolved towards. Unless you can deliver at a similar
	level, I would spend almost as much time reviewing your changes as if I just
	did everything myself, if not more. Frequent RE pull requests also reduce
	the chance for me to turn this into my only job, which I would very much
	like to do.
</p><p>
	If you still want to help by coding, I've got a bunch of other
	{{HTML_TagInline "contribution-ideas"}}. These are slightly out of scope of
	the main project, but interesting for the big picture nonetheless. They come
	with a lower barrier to entry, offer more freedom than regular
	reverse-engineering work, <i>and</i> I would actually appreciate your help
	there.
</p><p>
	In all honesty though, spending your time contributing to any other project
	would probably bring you much further in life than anything related to the
	main Touhou series ever will.
</p>
{{template "</q>"}}



{{template "<q>" "website-contributions"}}
	What about contributing code to the website?
</h3>
<p>
	There is a wide array of potential features that could be added to this
	site. Better accessibility for progress tables and syntax highlighting for
	code snippets are two examples of features that I already have in mind for
	future website pushes. Other features might be cool to have, but are maybe
	too expensive relative to their usefulness, such as {{Blog_PostLink
	"2022-10-31" "porting the ZMBV codec to the Web for efficient lossless video support on the blog"}}. And of course, there might be features that I
	haven't even thought about so far.<br />
	Unlike the core reverse-engineering business, improving the website is
	something I only get to do maybe once or twice a year, as a side project.
	That's why I would also highly appreciate you helping out in that regard.
</p><p>
	Note that any code contributions to the website will be licensed under the
	<a href="https://github.com/nmlgc/rec98.nmlgc.net/blob/master/LICENSE.md">AGPL.</a>
</p>
{{template "</q>"}}



{{template "<q>" "cap-why"}}
	Why a cap?
</h3>
<p>
	The cap corresponds to the maximum time I can healthily allocate to this
	project within the next 4 weeks. It is meant to
</p><ul>
	<li>protect all of <i>you</i> from throwing more money at me than I can
	reasonably convert into progress, and to</li>
	<li>ensure that at any point, I will be at most 4 weeks behind any
	commissioned pushes. Therefore, it will also protects <i>me</i> from
	burning out.</li>
</ul>
{{template "</q>"}}



{{template "<q>" "cap-too-low"}}
	With the current rate of progress, and the cap being at the low level that
	it is, the project is never going to finish!
</h3>
<p>
	If you all manage to regularly sell out the store at higher and higher
	prices per push, I will be able to increase the cap ever so slightly by
	reducing those pesky RL work hours. If those reach zero and I can turn ReC98
	into my only job, I can remove the cap entirely and go for a proportional
	"bidding war" model instead, allocating my constant amount of time relative
	to how much money comes in for a particular goal.
</p><p>
	But as of December 2021, that's still far off. Meanwhile, the steadily
	increasing amount of care and documentation I put into this project has
	proved highly popular, while no one has ever requested me to compromise on
	that and instead rush towards 100% RE as fast as possible. So, getting
	everything 100% done within the foreseeable future doesn't actually seem to
	be much of a concern for my existing audience.
</p>
{{template "</q>"}}



{{template "<q>" "slow-2018"}}
	Some of the 2018 pushes were delivered months or even years after they were
	paid…
</h3>
<p>
	Back then, I not only didn't have a cap, but also vastly undersold myself,
	while also offering crowdfunded features for thcrap in parallel. That's
	why the latter are sometimes referred to in the old blog posts here. But
	compare that to now:
</p><ul>
	<li>I've resigned from thcrap</li>
	<li>The aforementioned cap is in place</li>
	<li>The build system for this project is now pretty close to optimal,
	providing quick turnaround times, minimizing frustration, and keeping me
	nicely focused.</li>
</ul><p>
	However, if you <i>absolutely request</i> me to prioritize an element of a
	game that requires a ton of not yet reverse-engineered knowledge to fully
	grasp, and you <i>absolutely</i> don't accept your money going to anything
	else, I <i>will</i> have to put that on the back burner. It will be made
	clear in the backlog whenever that happens, though.
</p>
{{template "</q>"}}



{{template "<q>" "policenauts"}}
	I'd like to see PC-98 Policenauts (or any other DOS program compiled using
	Borland/Turbo C++) decompiled. What's in it for me?
</h3>
<p>
	The ReC98 repository includes
	<a href="https://github.com/nmlgc/ReC98/blob/master/Research/Borland%20C%2B%2B%20decompilation.md">
	a currently incomplete file with the ASM→C++ patterns, as well as
	information about the limits of decompilability</a>. This file will be
	continuously updated with new insights. So while you probably wouldn't
	want to support this project until the very end, it might be worth
	supporting ReC98 for just a bit – at least until it becomes obvious that I
	completely figured out Turbo C++, and that other decompilation project
	you wanted to see made significant progress.
</p><p>
	And who knows, maybe we <i>will</i> see a somewhat automated decompilation
	solution come out of this.
</p>
{{template "</q>"}}



{{template "<q>" "replays"}}
	I want replays! What's in it for me?
</h3>
<p>
	As of January 2022, I also offer to develop PC-98-native replay mods, if you
	don't want to wait for your favorite game to get 100% decompiled and ported
	to a modern system first. There's a separate option in the order form just
	for that goal.
</p>
{{template "</q>"}}



{{template "<q>" "translations"}}
	I want translations into languages with non-ASCII characters! What's in it
	for me?
</h3>
<p>
	In 2023, {{DB_CustomerByID 9}} commissioned the basic feature set that would allow such translations, and I expect to deliver it during 2024, after completing the <a href="https://github.com/nmlgc/ssg/issues/42">Shuusou Gyoku Linux port</a>. However, you probably want more than the basic features, so feel free to further support this goal by donating to their {{HTML_Emoji "opencollective"}} <a href="https://opencollective.com/thpatch">Open Collective</a> page. There's no cap there, and every bit will help to improve the end product. {{Blog_PostLink "2023-07-28" "Check my announcement blog post for details and feature ideas"}}.
</p>
{{template "</q>"}}



{{template "<q>" "netplay"}}
	I want TH03 netplay! What's in it for me?
</h3>
<p>
	{{Blog_PostLink "2024-04-24" "This blog post details all possible approaches for implementing it"}}, together with their advantages, disadvantages, and prerequisites. I will start dedicated funding campaigns shortly after we've met the first of these prerequisites.
</p>
{{template "</q>"}}



{{template "<q>" "ads"}}
	Do you sell ad space on this site?
</h3>
<p>
	Every contributor, no matter how much they paid, has the option to have
	their name be turned into a link to a URL of their choice. So if you
	consider that to be advertising, then yes. If you had more than that in
	mind, hit me up, and we might make it happen. No JavaScript or remote
	content, though!
</p>
{{template "</q>"}}



{{template "<q>" "take-my-money"}}
	Alright! I have understood what this project is about, and am convinced
	that I want to support it. Take me to the order form!
</h3>
<form action="/order">
	<p class="row">
		<input type="checkbox" id="consent_takedown" required>
		<label for="consent_takedown">
			I have understood the <a href="#tsa-takedown">potential risk of
			takedown claims by Team Shanghai Alice against this project</a>.
			While the amount of progress I am about to purchase will either be
			delivered or refunded, there is no guarantee that the project will
			stay alive for further progress beyond that. (This does not apply to
			Seihou-related goals.)
		</label>
	</p><p class="row">
		<input type="checkbox" id="consent_gdpr" required>
		<label for="consent_gdpr">
			I agree that the data of my contribution will be processed, stored,
			and made publicly available as seen in the <a
			href="/fundlog">crowdfunding log</a>, under a <a
			href="https://github.com/nmlgc/rec98.nmlgc.net/blob/master/LICENSE.md">public-domain
			license</a>.<br />
			I will have the option to stay anonymous though.
		</label>
	</p><p>
		<button>Order</button>
	</p>
</form>
{{template "</q>"}}