How I Wrote the Screenplay for “Arrival” and What I Learned Doing It
thetalkhouse.com
thetalkhouse.com
Also, the note about the limitations of screenwriting software is spot on. The fact that Final Draft costs $250 and can't even support image embedding is ridiculous.
Edit: fixing my solidus direction.
My primary problem with FD is no support for UTF at all. Real fun writing in other languages. Sometimes even impossible. Fade In saves the day.
For the reason noted in TFA.
Will I ever use this method in another screenplay? Maybe not.
There are tools to do that differently. Often times, I need to 'decorate' a screenplay (TV) with drawings for producers and as a selling tool. It's like an illustrated screenplay. I do it in... Word. Of all things. It's not a production screenplay, so a simple copy/paste between two applications works best. Paste Courier 12pt text into Word, style it a bit and embed images. Done.
If you want to have pictures follow your text, do it in Word or Scrivener (better). Scripts have a certain layout and rules which are there for a reason(s). Reasons which are important.
Can you explain why one would need special software for this? What does Final Draft give the writer, that word or open office doesn't?
You could, and people do, write screenplay in Word or whatever. Trouble arises later when you need to go through revisions and work with other people that work in those production-oriented apps. For that reason alone it's better to do it all in a specialised app.
Not to mention that lots of popular IDEs are written in unnecessarily wasteful/slow environments (Eclipse and Idea which use Java, etc).
As for implementation language, I'm not sure java is significantly slower in the case of the project size we are talking about here.
I find dialogue in variable width font easier to read, so I'm wondering if there's really some compelling reason why monospaced is better.
One page in a standard format screenplay translates into roughly one minute of screentime. There are exceptions to this rule, but the first thing someone in the business does when they get a new script is check the last page number.
Further, when a screenplay goes into production, the work for the day is organized into eighths of a page, which is used as a shorthand for how long each piece of work is going to take.
So while changing the tabulation and type spacing might make a screenplay more readable, it would make it much much less produceable. At the end of the day, a screenplay isn't meant to be read, and the formatting reflects that.
In general though, if format isn't what it is you wouldn't be able to gauge how long a script lasts and you wouldn't be able to break it down for production easily.
Beyond that, there are versioning tools (screenplay versioning can get a bit esoteric), alternate dialogue capabilities, scene numbering and reporting tools, and a lot more. It also has a lot of capabilities related to production, such as omit functionality, watermarking, script locking, etc.
Final Draft gets bagged on a lot for being expensive and ugly, but it really does have a ton of specialized tools that are incredibly powerful if used correctly.
I'm curious. What features to writers need that git couldn't provide?
The gist of it is that each version gets distributed to various channels of pre-production and production. Once locked, each page has its number and if the page changes, it stays in (!), but new one is added according to certain rules. In a way it's like a commit history printed out, but isn't. It's weird, but it works.
Nothing in that article supports the assertion that screenplays have a different concept of versioning than software developers. Screenwriters just have a different preferred format for displaying diffs, a different convention for version numbering (but that's completely irrelevant to a tool like git), and a paper-saving way of handling insertions and removals, when printing out on paper (again, irrelevant to git).
Advice like "If it's only one or two words that are different, consider just whiting out the existing script and writing it in, rather than reinventing the wheel and generating a new color." makes it very clear that this industry hasn't even tried to use technology to make these problems go away.
I wonder how long until this whole system gets replaced by tablets that always automatically show the latest revision, without requiring the use of esoteric paper colors.
Screenplay is a tool, a blueprint that serves various departments throughout pre-production, production and post-production. That blueprint is also in flux as others are using it. I guess, well I believe one could modernise the whole screenplay pipeline. Trouble is you would have to make a significant advance in order for people to start using it. Every dept. except acting/blocking could start using new system(s) right away. Acting and blocking would still prefer paper.
Take into consideration that it's easy to criticise a system as an outsider. You would have to witness first-hand an organised chaos that film and/or TV production is in order to appreciate every single idiosyncrasy (at a first glance) that makes it all work.
Prime example are cameras. Usually people (who are for the first time on a set) ask about camera. Why this camera, why not that, etc. It works and doesn't stop the production train. If you can avoid slowing down anything on set (or before or after) and THEN introduce significant benefits, people will listen and welcome you immediately. This is an industry that takes in new technologies all the time, no hesitation. Provided they bring significant boost to productivity.
I think the key insight that's missing is that the content of the screenplay is separate from the presentation. It sounds like there's no reason why a screenplay couldn't be edited as plain text in a font more readable than Courier with changes tracked by an off the shelf version control system. The editor shouldn't have to pay heed to concerns of pagination or paper coloring or version numbering, because all of that can be rather trivially automated by inspecting the version history and rendering a printable PDF as needed. Some actors may prefer looking at an archaic presentation of the information, but that requirement doesn't need to influence how anybody else interacts with the information.
Regarding font - Courier 12pt is standard monospace used. People have (and do, but not often) use different monotypes. Using a monotype with certain margins and along with several other rules yields you a consistent overview of the script where the result of using all of that is approximately 1 minute of screen time per one page of script. That's why it's used. Non-mono font wouldn't yield same results.
Further on, people do not interact with scripts on paper anymore. You do print yourself a copy if you prefer to read it like that. Also, actors (and some people on set) do have a physical copies due to the nature of table reads (I've seen tablets on table reads), and blocking (moving around the set with script in hand, rehearsing).
In reality, no one has to even think about it. Reason is because there is one (or select few) producer and multiple consumers of the script. One persons job is to make sure everyone has the latest revision that they base their work on and that's it. It would be an issue if everyone had commit rights, but they don't. They just checkout latest commit and have the person that checks it out for them! Sounds dire, but is actually painless.
I'm certain that an automated word count could easily provide a more accurate heuristic and free up the graphical presentation to be more readable. This heuristic is exactly the kind of narrow-minded unwillingness to consider modern solutions that is so appalling. And the conventions were obviously not constructed for this purpose; they're relics of long-gone technological limitations and the 1 minute per page rule of thumb was probably not formulated based on a thorough measurement.
I'll relay another joke from set:
A new Assistant Director comes on to help the second unit on Gone with the Wind. He finds the script supervisor.
AD: How much are we shooting today?
SS: One eighth of a page.
AD: You're kidding! We'll be home in an hour.
SS: I don't think so.
AD: What do you mean? What are we shooting?
SS: Atlanta burns.
The point is that it's an inherently fuzzy process, and heuristics work a lot of the time, but nobody cares about a system that works a lot of the time.
If you design something better, I'll be glad to use it (hell, I'll help you design it), but please don't be so quick to judge. It's a business of edge cases, and it's been built as such.
If it makes you feel any better, a lot of multicam shows and studio shows have switched over to all digital script distribution that's automatically distributed. But there is absolutely no replacement for a paper script for a lot of really compelling reasons.
That's true of every single system that evolves organically, including software engineering. Or do you think all the cruft that accrues in every programming language is due to rational design?
The question isn't whether some new convention would be better, it's whether the benefit would outweigh the transition cost of either getting everyone to adopt it all at once or having crossed wires as people use different conventions at the same time.
Switching to a more readable font is a change that is trivially implemented now that nobody actually uses mechanical typewriters, and it would break nothing. Using a more readable font doesn't force you to include any more lines of text per page, so the dubious heuristic about page count and screen time wouldn't necessarily even be a casualty. There are no other substantive reasons to reject improvements to readability.
I certainly understand that there will be oddities of the film production process that are not worth fixing. But people who go out of their way to defend the pointless use of bad typography on printed documents that are meant to be read are in no position to be making credible arguments in defense of the less obvious flaws in their standard practices.
Inertia and information signalling are substantive reasons in an industry that is governed by convention and lots of networking.
Using a different font is a signal to the reader that you are ignorant of industry conventions, and hence less likely to have written a script that will follow acceptable norms--or perhaps worse, that you're a special snowflake who is willing to distract the reader from the content of the screenplay in favor of fiddling with form.
This is the same reason why most people don't make stunningly bold typographic choices when writing a resume -- they're not in a position of power, and they're either not confident enough in their own taste or the reader's taste to take any risks. Guides for aspiring scriptwriters basically suggest that you stick to the 'standard format' and not deviate from it [1].
Certainly, if you were writing and submitting scripts, feel free to implement your own suggestions.
[1] https://www.writersstore.com/how-to-write-a-screenplay-a-gui...
> Using a more readable font doesn't force you to include any more lines of text per page, so the dubious heuristic about page count and screen time wouldn't necessarily even be a casualty
So the heuristic would be a casualty, unless you jumped through hoops to make sure it still worked. Does this still sound like a cost-less switch to you?
Unless somebody has actually done an analysis of how well script page count correlates to screen time with both fixed width and proportional fonts, the most you can assert is that the heuristic of unquantified accuracy might require adjustment. Conventional wisdom that hasn't been subjected to rigorous analysis shouldn't be presumed to be precise. The film industry certainly hasn't given itself much opportunity to accidentally discover it if proportional fonts turn out to yield a better estimator.
The typographic conventions for scripts were not constructed with the purpose of being used for this kind of estimation, thus it would be completely unsurprising if a study were to find it to be 30% off on average; the rule of thumb itself is the product of at least some rounding already. And there's no reason to believe that a re-calibrated heuristic based on more readable typography would need to have higher variance than one based on typewriters, either.
Computing the typewriter page count of a document is such a simple task that the feature could be added to any relevant piece of software in an afternoon. It would make sense to do so even if you're still going to print it out with typewriter formatting, because glancing at a footnote with those statistics is easier for a person to accomplish than counting physical pages.
Every organically evolved system has details in it that, in a vacuum, could be individually isolated, analyzed, and then improved. You're almost certainly right that Courier New 12-pt is not the optimal font for a screen play -- but is it worth the time and effort to 'fix' it?
> the most you can assert is that the heuristic of unquantified accuracy _might_ require adjustment.
Sure, but the onus is on the person suggesting the change. Nobody is going to make sweeping changes to long-established conventions based on little evidence that it will make a big improvement. And the industry is not going to do the study you suggested, for the simple reason that someone will guess that the cost of doing that study, plus switching to the new font, is probably less than the total benefits of having a new font. Is that an airtight 100% logical proof? No. Does that matter? No, because humans make subjective judgments when the cost of acquiring better information is perceived to be high .
I mean, what you're suggesting is pretty similar to going around and telling every team that should refactor their code base for improved readability.
Are there clear benefits to having better readability? Yes. Do they always exceed the cost of refactoring? Emphatically no. Is it easy to put a definite number on the cost or the benefit of refactoring? No. Therefore, we're left with heuristics and guesses. Sometimes it's worth it to refactor, but only when the code starts to really really smell. And neither of us are in a position to judge whether the process in the film industry has started to 'smell', because we don't work there.
'If it ain't broke, don't fix it' is an excellent meta-heuristic. Most organically evolved systems are sitting at a local fitness maximum. Finding and moving to a better local maximum is usually costly.
Screenplays are not meant to be read, they're meant to be understood and produced. Your argument is akin to saying that blueprints are too complicated for the average person, therefore they should be changed. They're not meant to be understood by the average person, they're meant to be understood by industry professionals. Just because you find it less than readable doesn't make it so. Those of us who know what we're doing with it think it's just fine.
I'm sorry you don't find them easily readable, but they're really not meant for you.
That's nonsensical, unless you mean that they're meant to be mechanically analyzed—page counts, etc. that are obviously easier to accomplish without first printing them out. Even if they are not typically read in the manner of a novel, they're still intended solely for being read by human eyeballs, and for that purpose the typography is obviously deficient in ways that can be improved with no downside.
Repeatedly pointing out that the current system manages to work is not a refutation of the claim that it can be improved. Why are you so unwilling to admit even the most superficial of flaws in your industry's standard practices?
Sometimes it's the companies that dictate standards. FD is notorious slow mover, yet it's 'an industry standard. I worked on one script with another writer. At that stage it was only two of us drafting the idea and I had a need to see what has changed in a script since I sent him my last version and vice versa. I mailed the dev of Fade In if he could just include a diff (yes, our beloved diff) somehow inside of Fade In. A day or two later it was in. I use it now constantly. Same ol' diff we're used to, kind of looks same as on github as well. Red and green lines and all.
This is an industry with lots of resources and a keen eye on new technologies. If you come up with something that optimises anything in workflow, it will come to open hands.
If you want to make it happen right away, make a royalty free stock footage archive in 4k of various snippets of code with metadata on what it's for. Less work=more adoption.
It would also be a recipe for misunderstandings and errors, while showing this information directly in the formatting of the output is instantly recognizable by all, even when printed out.
Clever-er isn't always smart, reminds me of the (not true, but used as a metaphor here) story of the US spending millions to make a pen that can write in space, upside down etc, whereas the russians just used a pencil.
I think this can't be overemphasized. I got into the movie business from software and was astounded by all of the quirks in the process that seemed to me to be fundamental inefficiencies. I was fairly certain that I could quite easily reinvent the process to make it smooth, efficient, and pleasurable.
Now, having been in the business for 14 years or so, almost all of it makes sense, and my arrogance in regard to their process is embarrassing at best. There are still a handful of quirks that don't make a ton of sense (even the property department will acknowledge that the only reason they handle cast chairs is because it's tradition), but generally speaking there is a good reason for everything being done the way it's done and it was a tough lesson that made it that way.
I also think it's important to remember that this is a 100 year old industry, and it's not an engineered process, it's an emergent process.
Just as an aside, the first time you have a week worth of wet work scheduled on a movie that's shooting so far out in the sticks that you have to have a satphone for set to communicate with basecamp, a lot of the reticence to adopt technologies that aren't completely bulletproof makes a lot more sense.
I just came off a shoot last week where we had an extremely tight schedule and I was directing and DP-ing (small shoot of 5 days, but with full everything - gear, people, sets..). I had a great plan to shoot some things on RED and most on Alexa cameras. I had both at hand. I did that, but during the one of the setups where I wanted to use the RED, camera was starting to weird out on me. I had to play tetris and get past level three on its small screen in order to get to the menu I wanted in order to set something up, and then it was showing one thing while idling and other while shooting. Not even my ACs knew what's up and I ditched the camera in favour of Alexa right then and there. People started to wait on me on set. That's a big no-no, since overtimes and everything stands still. Otherwise a fine camera, but set was being held up. I did a few setups with it that were great and all, but from now on I will avoid RED just because of that experience. Maybe on more lax productions I will consider it again. That's how these practices are forged. When 30+ expensive people are waiting for you, while you're tight on time. That's how everything in this business evolved.
I hate to say it, but it's not surprising given someone like Jim Jannard came along and said "This business is backwards! I can fix this!" without completely understanding the requirements. Move fast and break shit doesn't work for every industry.
The takeaway at the end of the day for me is that my idea of efficiency is built around conditional repeatability. However, the fundamental drive of art is to create a unique experience. For obvious reasons, these two goals run a bit counter to each other. There are absolutely repeatable elements, otherwise we wouldn't be able to relate to media (or each other, for that matter). But by the very nature of the requirement of uniqueness, there is no one size fits all. When the syntax is the program, it tends to exclude all but the most basic and robust standardization.
At least that's my drunk peanut gallery theory.
Because programmers, especially on the UNIX world, are known to be so much more forward looking and having it the "easy way"?
https://gist.github.com/cookrn/4015437
>Nothing in that article supports the assertion that screenplays have a different concept of versioning than software developers. Screenwriters just have a different preferred format for displaying diffs, a different convention for version numbering (but that's completely irrelevant to a tool like git), and a paper-saving way of handling insertions and removals, when printing out on paper (again, irrelevant to git).
So they have absolutely no use for Git, except as a lower-level versioning engine in the backend, with all the actual results and presentation layer reformatted into screenplay-specific views. So why should they care for Git specifically over any other engine that provides the latter out of the box?
Is that question actually meant for me? I made no suggestion that git should be used to replace any existing software in the screenwriting profession. I merely used it as a canonical example of a software development version control system, to refute the statement that "It's not versioning like you would find in software development." There's nothing fundamentally different about the revision tracking done for screenplays than for software; git's inadequacies for the task aren't in the core functionality but in the window dressings, which most software developers also complain about.
Do they also not run out of batteries when you're filming in the middle of a desert at night?
This is the first thread I've ever commented on for HN, mostly because I'm usually out of my depth. This one, I'm actually an expert on. This will also be my last comment here, mostly because I expected more than your bristling, reactionary, unobservant level of discourse.
All the best. I hope you find the peace you're looking for, and that it's not through arguments on the internet.
For example, camera cards and offloading them into a computer. There's a specific sequence which reveals where people got burned on each step.
This is what you do with Alexa for example. You get an unlocked card from assistant which you put into the camera. You go to either quick erase or full erase (format) on camera and then you have to press two buttons simultaneously on camera to OK the erase, like launching nukes. Then you get some red gaffer tape and write down the cards number (A016 for example) and tape it OVER the flapping gate cover of the card on the camera. Once you're done with the card, it's full or whatever, you take it out, lock the writing lock on it, put it in its plastic sleeve and tape shut that sleeve with that piece of red tape with number on it. If no sleeve, tape goes over the card. Then you give the card to your assistant or DIT or whoever is in charge of offloading on your set. He offloads the card, removes the tape, releases the lock and gives it back to you as ready to use. There are multiple cards like that in circulation at any given time.
Ease of use, for one.
But for the final presentation as these codified movie scripts, you better produce something very clear.
Story is very careful that it is just a very different (alien) perspective of mundane (no time travel) happenings. I haven't seen the film yet, but by all accounts it includes events requiring time travel (or retrocausation), which is definitely not the case in the story.
I would be much happier if she just got incredibly smarter after the interactions. I can see so many plot possibilities arising from that.
(People start distrusting her; she starts feeling more intimate with the aliens then humans; etc)
It resembles many programmer's belief that there just may be a 'holy grail programming language', where translating abstract thoughts into working code gets much easier, and your thoughts themselves are shaped by the structure of this language.
It might be arguable that Turing-complete languages in general are the holy grail, allowing expression of any algorithm into code. But indeed our productivity has risen so much with modern languages compared to the very first Turing-complete languages that one does wonder how far from optimal-human-productivity our languages really are.
The same musings naturally apply for human language of course :)
Also, the movie didn't bother to really explain this – though it made hints with the statement about how the heptapods see time, and Hannah's name being the same forwards as backwards, etc. – but there's no time-travel aspect specifically because there's no influencing going on at all. She doesn't cause anything to happen.
The first half of the movie, on the other hand, conveys the feeling of dread wonderfully.
I thought a lot of the tone, which made the short story so good, was lost, probably due to the medium and the audience. It became much more of a sci-fi alien thriller than a commentary on mortality.
I found that, having read the short story, I enjoyed the movie less than those who had not read it.
Heptapod communication, and later on Louise, is performative. Imagine that you're playing a part in a live theater. You've read the script. You know what's going to happen next. You have lines to say, and you say them. You choose to participate in the performance.
It's a really interesting resolution to the argument between free will and determinism.
In this case, we can infer that, even though Louise knew everything that would happen, including her daughter's disease, she chose to play her role anyway. She understood that it was just a smaller part of a much larger, pre-determined script, and she was content to be an actor in it at all.
For another example of this being hinted at in the movie, Abbott knew exactly the fate that awaited him. He could have avoided it, but didn't. That was his role.
It's a metaphysical argument that's deeply unsatisfying to a lot of people. For a lot of folks, free will means being able to force change as you see fit -- the Edge of Tomorrow position. But, for a minority of others, free will and determinism can coexist if you're willing to accept that just because things can be changed doesn't mean they should be. It's the ultimate form of stoicism.
(Also not addressed in the movie is the neat question of whether the heptapods' original decision to interact with humans for the sake of their own future is part of the script, or if it was an action they took outside of the script for their own sake.)
Like a lot of good sci fi I think this is one of those movies where I'd appreciate a Director's Cut or Extended Edition that includes obscure and technical and deep (read: nerdy) bits cut from wide release.
OFC my opinion.
I actually think the circle signature shouldn't have been cut because without it, I felt that the entire scene fell a bit flat.
I wonder: given that a Photon travels at the speed of light, then everything happens instantaneously, from its point of view. So, is choosing the path that takes the shortest time like a standing wave?
Note that photons don't take the shortest path; they take a variationally stationary path (not necessarily the shortest one). My suspicion is that this is because all the possible photon states are coherent along and immediately nearby that path, whereas if the path is not variationally stationary then it interferes with nearby paths. I've done very little variationally physics, however, so take this with a grain of salt.
Well, am new to the subject, so not sure what this means. Am curious if there is any relationship between Fermat's principle of least time (https://en.wikipedia.org/wiki/Fermat's_principle) and Quantum FFTs (https://www.quora.com/How-does-Shors-algorithm-work-in-layma...)
I don't know enough about quantum FFTs to know if there is an analogy. However, I'm taking Aaronson's class next semester so I'll be able to tell you in a few months ;)
Screenwriting software is ridiculously constrained. Margin sizes, line sizes, font sizes, font weights, everything, is ridiculously locked down. In the physical world, there were (are?) equally ridiculous constraints on things as minor as the binding of the scripts (which is to be done with brass "brads" [0]). It's absurd and absolutely intentional because it's what the real, ultimate customers of the screenwriting software want.
The real customers of the screenwriting software are the people who read and ultimately buy the scripts, and they want all these constraints in place. 99.99% of all screenplays never get bought, and 99.9999% of purchased screenplays never make it into theaters as a film [1]. In that sort of environment, the top goal of every screenplay buyer is to avoid wasting their time on screenplays that aren't worth reading.
If you send your genius script to people in Hollywood set in Helvetica or Gil Sans or whatever other font you favor, it will not get read. Period. That's an absolute deal killer for potential buyers. The reasons are a mix of process (constraining margins, font size, line spacing, font keening, etc. makes documents instantly comparable - look at the page count and you know the running time) and sociology (filmmaking is an incredibly complex and expensive multi-player art form with each project involving hundreds of people working together to build a product for millions or hundreds of millions of consumers, and the person considering buying your script wants to know do you understand how that works well enough for your idea to have a chance of surviving).
It's kind of like the way I put a footnote reference in my first paragraph starting with [0] instead of [1] or * the way the rest of the world would have if they wanted to put a footnote in plain text like this, it's an early visual cue that I might be a writer who gets how HN works. Ditto for the [1] footnote around the obviously made up, quantitatively inaccurate but qualitatively accurate stats on screenplays (the WGA does occasionally publish stats on the number of people who earn a "full time salary" writing screenplays and the numbers are amazingly depressingly small).
It doesn't matter how great your idea is, or how unique your personal creative vision is. If you put pictures in your screenplay, you are putting a giant HTML <blink> tag in your script that screams to every serious reader "My screenplay isn't worth your time to read, because I don't know what I'm doing." It's all well and good to be a special snowflake in your own mind, but if you want to be taken seriously in an incredibly sophisticated multi billion dollar industry where >99% of screenwriters never produce a script that gets produced, you distinguish yourself not by showing you know how to use blink tags but by showing you know how to grab people's hearts without them.
If you are a writer and you must have visuals, keep it to a single page, called a one sheet, with a single powerful evocative image on it to help the people you pitch to remember the concept for your story (and then try not to use or show it). Directors are buying your script, not your visuals.
When you're ready to produce and/or direct your own screenplay, then you can assemble your own visuals, but you still want to keep them out of the screenplay. Put them first in a mood board that conveys the feel of the story without the constraints of the script. Then put them in storyboards that convey the visual telling of the story at a high level (initially) to facilitate a deeper conversation with all the parties involved in the effort. Then produce a story reel, communicating the feel and pacing of the story. Or ditch them all and just make the film. The choice is yours if you are a producer/director but it's theirs if you are a writer. This is a big, complex, sophisticated industry you're working in. It's not always going to be optimized for you or your needs because it's optimized for the total needs of the set of all players in the industry.
All of which means please don't put images or blink tags in your screenplays and please don't put image embedding tools in your screenwriting software unless you want the people using your software to fail at their ultimate goal of having their screenplays made into actual films.
[0] https://www.writersstore.com/screenplay-fasteners/
[1] we'll come back to this in a moment
Unfortunately, you come across as highly antagonistic to a screenwriter (a) whose script was almost immediately made into a brilliant, popular film [73] and (b) whose story was foundationally about the way language affects our thinking.
Your main argument is that people's scripts will get rejected if they don't chain them into a box and use peculiar little brass fasteners to hold them together. This inclines me to believe that you've not yet actually seen "Arrival."
Sometimes I think a little change in the film industry would be productive, if it meant we'd get more movies like "Arrival," even if it means breaking rules, and fewer movies like "Captain America: Civil War," who follow the rules but composite dozens of mismatched absurdities and are basically mass-market children's cartoons.
[73] Common knowledge. By the way, championing your knowledge of "how HN works" WRT references just sounds self-congratulatory.
The bit about the footnote styles was more to help HN readers who are not screenplay experts validate an example from their world where formatting styles communicate in-group vs out-group status, because as you point out "everyone" knows footnote counters start at 0 on HN.
A very good intro is also http://blog.stephenwolfram.com/2016/11/quick-how-might-the-a... by Stephen Wolfram who helped with some science parts on the movie.
My short review: https://letterboxd.com/rurban/film/arrival-2016/1/
Also I'll chime in about the soundtrack; I loved it. Lucky me Google Play Music has it - it says I've listened to it about 9-10 times through so far.
I said it in another comment, but I seemed to enjoy the movie less because I had read the short story.
Forget aliens for a moment. Contact is a retelling of one of the oldest stories on record: the plight of the faithful. (Wo)man encounters awe-inspring supernatural force, tells the world but can't prove it, gets dismissed as a fraud/crazy, but goes on knowing in her heart that her experience was real. The rest is implementation details.
Putting a scientist in the role of prophet, and getting buy-in from science-y, atheist audiences by using aliens instead of God (you don't know that they're a God metaphor until you're on Jodie Foster's side), are precisely the kinds of intentional inversions that make adaptations of classic tales great. Present the message differently, using different tools, to appeal to different people in different ways.
In theater, Anais Mitchell's Hadestown takes Orpehus and Eurydice, sets it in 20s Appalachia, and expands a story about a hero into a story about the industrialized world: Orpheus and Eruydice's light, playful hearts are nature and beauty and song, while Hades is heavy industry, the coal mines, the Man, the capitalist machine. Also gives it some awesome music.
Sarah Ruhl's feminist Eurydice, from the same myth, subverts a story of a man's heroics in (trying to) rescue his damsel in distress, and transfers the agency to the damsel. These are both, in my mind, highly successful retellings.
Interstellar certainly has components that you'll find elsewhere - what story doesn't? - but none of them that I can place are a huge part of its identity. I wouldn't classify it as a retelling.
Arrival directly copies the setup of Contact: aliens make contact, we ingeniously decode their communication, the message is a blueprint/schematic for something that might be a weapon. Then it concatenates on the ending of Interstellar: time is non-linear, it can be transcended by the power of (love|language). It is neither faithful to nor a clever inversion of either story. It just takes half a plot from one movie, half a plot from another, and puts a different bridge in between them.
I had a good time in the theater, I'd probably see it again, I just don't have as much respect for the story as I did for either Contact or Interstellar individually. The linguistics were an interesting value add, just not that powerful.