The Confusing State of RCS | Hacker News Reader

The Confusing State of RCS

substack.com

74 pointshocuspocus85 comments

This post left out the Android situation for custom roms, rooted devices, and non-Google blessed devices. On these devices Google willfully and maliciously blocks RCS for these subsets of would-be users: https://www.theverge.com/2024/3/1/24087418/google-messages-b...

I say willful because they admitted to it, malicious in that at no point will the device inform the user that RCS is blocked. It shows connected and will not relay messages. Spoofing fingerprint and other props it will immediately begin to work again. For users affected, they will have inbound messages silently blocked (because other users will see the device registered on RCS, but Google will not relay it) but not know about the disruption unless they were aware of this news.

EDIT: In case anyone thinks this is me being alarmist, RCS is beginning to be used for 911 here's Google's own press release on it: https://blog.google/products/messages/google-messages-rcs-91...

Ha, I admit last time I worried about custom ROMs, I was using a Nexus 5.

I think it's a lost cause and not only because of RCS, see also banking applications. On the other hand, even a $200 Galaxy A gets 4/5 years of updates today, which benefits the vast majority of customers a lot more than an unlockable bootloader.

I think you’re getting downvoted so heavily because most of HN dislikes anticompetitive behavior from big players. Having no FOSS solution for RCS is a problem. So is silently breaking the OSS user’s text messages.

I agree that Google's behavior is at best questionable, and I tried to convey that in the article. Especially on Android, RCS support could be split into a subsystem exposing a public API and a front-end client, exactly like third party SMS clients work.

But I also live in the real world and things like Android's SafetyNet make custom ROMs almost unusable as a daily driver. Frankly, nothing forces Google to keep selling unlockable Pixels.

I strongly disagree that SafetyNet makes custom ROMs unusable as daily drivers. I don't need a banking app on my phone every day - in fact I don't use one at all. I also don't use Google Pay; these are not essentials to a phone.

But reliable cross-platform messaging? That is essential to a phone. MMS sucks, and I can't convince everybody in my life to switch to Signal.

Google Pay and SCA with banking apps are pretty essential to a lot of people.

I'm using "essential" here in its narrow (and proper, as I would argue) sense. I can use a physical credit card instead of GPay. I can use my bank's website instead of the app. But if an MMS fails to download from a group message I'm in (which happens to me fairly frequently), I have no alternate recourse to see what the message said (except possibly asking a group member to send me a screenshot by email).

The client situation for RCS is really quite bad. With SMS and MMS I can fairly easily run my own client on a linux system that has a GSM modem. This is not the case at all with RCS. Even worse, Google Messages is not open source and so far there have not been public APIs available on android to interact with RCS messages.

I can't upvote this enough. As an Android developer of an app that uses text messaging, RCS is completely locked out.

>the relationship between carriers, Google and Apple.

I am begging you to accept the Oxford comma into your heart.

I think anyone who couldn’t infer the meaning of that sentence deserves whatever dire consequence you imagine they might experience.

Plenty of people don't know the definition of "carrier," and it's especially confusing because Google is a reseller (via Google Fi).

A single comma would clear up any possible confusion and require nearly zero time to add.

To be fair, this was an egregious sentence to not add a comma to.

It needs to be reworded. "Between" implies two parties: so a relationship between (one carrier) and (one mobile OS/app dev/vendor).

Since there are many carriers and at least two OS vendors, now we have "...relationships among..."

Well technically...

https://fi.google.com/

The Oxford comma isn't a panacea. Sometime it makes a list read like a parenthetical phrase

In every example I recall seeing where using an Oxford comma causes a problem it is because some sort of appositive or parenthetical phrase has been set off with commas.

Commas are the most common way to set off such phrases, but they are not the only way. Most grammarians seem to think that em dashes or parenthesis are acceptable, and I've seen styles guides that recommend doing that if there are commas in the sentence.

As far as I can tell if we just stopped using commas to set off such phrases when other commas are in the sentence (or just stopped using commas to set off such phrases all the time) that would get rid of all the cases where including the Oxford comma in a list makes the list ambiguous, without changing the cases where not having an Oxford comma is ambiguous.

Cue the em-dash, semicolon, colon, and parenthetical as secondary clause separators. If you still can't use the (in my opinion mandatory) serial comma without ambiguity then you need to rephrase.

I don't agree. Can you share an example?

They went to Oregon with Betty, a maid, and a cook.

Is Betty a maid or did they go with 3 people?

This could be two people, but would normally be written with a different separator: "Betty, a maid; and a cook" (just removing the comma doesn't help because then Betty could be a maid and a cook). As-is, the implication is that this is three people. If you would like to make that more explicit, you would instead re-structure the sentence†, so it's not highly relevant to the serial-comma-vs-no issue.

†For example:

Betty, one maid, and one cook.

Betty, and a maid and a cook (a little awkward)

A maid, a cook, and Betty (depends on how you want Betty's inclusion to land for the reader)

Right, you can change punctuation to clarify it. However, it doesn't change the fact that the Oxford comma could make the list readable as a parenthetical phrase.

I'm not saying the Oxford comma is bad. I'm just saying that it isn't 100% perfect as many people imply.

Three people; otherwise it should be, "They went to Oregon with a cook, and Betty, a maid."

Or better yet: "They went to Oregon with a cook, and maid named Betty."

Thanks, fixed :)

The misery of RCS ultimately was rooted in money, not technology. Carriers wanted a way to monetize messaging services on LTE like they did in the 1990s with SMS. They thought that inventing a new standard, ostensibly built around SIP (ie more integrated into the telco session layer than SMS, which was a bit kludgier), would be the path.

What they failed to realize what that LTE brought easy access to data, and in particular OTT messaging like WhatsApp. Consumers, remembering only too well the horror stories of being gouged by telcos for heavy SMS messaging, were not inclined to trust the carriers. So WhatsApp etc. took off while RCS stagnated.

On the technical side, vendors soon discovered that every telco wanted to do RCS slightly differently, and that implementing RCS in an IMS network was not as easy as slapping in an app server. So when Jibe showed up, a lot of the air left the room. Problem was, the telco people saw precisely what Google were doing: taking away a potential revenue stream from their control and monetization. So again, things languished.

The technical people I worked with at mobile operators across this time were very sharp people who could make most anything work. But the leadership was so terrified of losing money, and so bad at making money, that many good plans never saw the light of day.

Exactly this! Why would anyone use RCS when there are still carriers out there, that take ancient SMS (usually 0.19€) and, should you send photos, even MMS prices (0.69€) for every single RCS message.

Again I was thinking this was the "real" RCS:

https://www.gnu.org/software/rcs/

seems we are running out of acronyms.

I don't know, my mind went to the messaging system that's being deployed now, instead of an ancient VCS. There are only 14k possible three-letter acronyms anyway.

There are only 14k, but they are very unevenly distributed, so you might be surprised how many are left for the taking: https://gwern.net/tla

You only have to go as far as 'CQK' before there is, AFAICT and based on Wikipedia entries, essentially no even slightly important use of a TLA. There are then another >2k more or less unused, and if you go to four-letters, you've got ~390k to play with.

‘CQK’ sounds like piece of trivia I would’ve had to memorize for an amateur radio license exam.

Close-Quarters Kinematics, used in robots in tightly constrained environments.

RCS brings to my mind Reaction Control System, as in spacecraft. Been reading a lot about space recently…

Although I suppose the headline doesn’t make much sense in that context - spacecraft systems are rarely in a “confusing state” (except maybe Boeing Starliner)

We ran out a long, long time ago. 26^3 is only 17,576 and 26^4 is less than half a million (even if we intentionally set out to increment them rather than use them as meaningful acronyms) yet the birthday paradox gives grave odds for even longer acronyms via random assignment.

And that's including awkward ones nobody wants like QWFW

Not only that, but it appears that there are at least five expansions to "R___ Control System": https://en.wikipedia.org/wiki/RCS

I was working at a large telco back in 2010 when they started introducing RCS. Yes it’s that old…

Point is, we never got very far because every carrier and every phone manufacturer wanted to have their say. It was (and probably still is) "standard definition by committee". And that just takes forever.

I really, utterly, and completely hate RCS. It's incredible how Google and Apple managed to screw up the messaging market.

I want to have very straightforward functionality: send rich notifications to my clients via RCS.

It's simply impossible. Carriers control RCS, and you need to make commercial agreements with each one to make sure your messages are delivered. Wonderful.

But it gets worse. There's RCS Business, so once you do get agreements with carriers, you can spam users as much as carriers approve. There is no way to opt out of automated messages. India got a dose of it and RCS got disabled pretty quickly.

I wish carriers just looked at Telegram or Signal for examples of _good_ messaging infrastructure.

Article author here, I didn't look into this as I naively assumed that the consolidation around Jibe would make this use case better than SMS, not worse.

I see a lot of providers on this page: https://jibe.google.com/partners/messaging-partners/

Including Twilio, Sinch, ...

Yep, to make RCS business messaging useful, you'll have to use a middle-man like Twilio. They take care to sign contracts with individual RCS carriers. And some carriers might refuse interconnections entirely (as a result, Twilio supports only a subset of countries).

And of course, this is not cheap. Carriers don't have any market pressure to compete on the RCS interconnection price. So you'll be paying not-insignificant per-message fees over the IP infrastructure.

On Android iOS interop via RCS

The user experience for Android users is very poor.

I'm on a 4 people group chat with 3 iOS users.

Sometimes when I message the group, I see 5-6 threads in my Android messages app. Groups with every combination of 3 people, 2 people and 1 person. That's right. There is a group where I'm messaging myself.

So 2024 isn't the year of interop. Lot more needs to be fixed.

Shortly after RCS became supported, I disabled it. After going years without receiving a single unexpected SMS, I was suddenly flooded with fake DHL schemes via RCS from +63 numbers (to my German number). I don't see a reason to ever re-enable RCS

Interesting.

I had never gotten any spam via RCS until this year, and it's been in relatively small numbers compared to WhatsApp, Telegram, and mostly SMS which stays the top spam vector for me. In any case, Google Messages spam filtering works incredibly well.

The article does not clearly state whether Android RCS and iOS RCS will work together.

Interoperability depends on how RCS is implemented behind the scenes. Just because iOS RCS is here doesn't mean you can assume your Android friends will receive your messages or vice versa.

iOS RCS relies on mobile operators's infrastructure, while Android RCS uses Google services.

You can see where this is going, right?

Unless your telco invests in infrastructure to support RCS in-network, those two ecosystems will not talk to each other.

>iOS RCS relies on mobile operators's infrastructure, while Android RCS uses Google services.

1. I thought Android RCS (ie. google messages) only uses Google's infrastructure if the operator doesn't provide it?

2. Does Google's infrastructure interoperate with other mobile networks? In other words, if I'm on a network that provides its own RCS infrastructure, and I send a message to my friend who's on another network that doesn't provide RCS and therefore falls back on Google, what happens to the message? Does it get routed through my network, which routes it to google, or through google exclusively? If it's the former, it's plausible that RCS would work for iOS users, at least for users on networks that provide their own RCS infrastructure.

Correct. Where operators don't provide, Google will use their servers. So the fact one uses RCS on Android does not mean an iOS user will have access. It may be using Google's servers...

Today with Google Messages, the backend is always Jibe, which is interconnected worldwide. That's kinda the point.

The only question is whether Google managed to make the operator pay for it.

Hi, maybe this part is not entirely clear, but everywhere RCS was enabled by Apple, it works.

The thing is, outside China and a few other exceptions, there's no MNO infrastructure. It's Jibe or nothing.

Google seems to run several "catch all" domains to provision Google Messages with the IMS configuration in all countries where Jibe is running, for instance:

config.rcs.mnc${MNC}.mcc${MCC}.jibecloud.net

While the iOS client should default to 3GPP well-known configuration domains:

config.rcs.mnc${MNC}.mcc${MCC}.pub.3gppnetwork.org

But in almost all cases this should be Jibe endpoints too.

It doesn't work in New Zealand. Operators have not deployed the infrastructure needed for it.

Again, there's no operator deployed infrastructure, it's Jibe, which works in NZ since 2020 like in most of the world:

https://www.geekzone.co.nz/forums.asp?forumid=97&topicid=275...

I like how you are replying to me with my own website.

Here is some interesting comment from a telco staff https://www.geekzone.co.nz/forums.asp?forumid=76&topicid=315...

What I meant is, interoperability between iOS RCS and Android RCS won't work until telcos deploy their own servers and Google fallback to those.

And again, telcos don't deploy servers. The backend is provided by Jibe, it's already running.

The iOS 18 RCS client provisioned in the US, UK, Spain, Germany, ... connects to Jibe and can send messages to Android devices running Google Messages.

Anecdote: iPhones on Google Fi still can’t use RCS, with no timeline for getting the carrier bundle bit flipped.

I mentioned this in the article, as it's pretty ironic.

Apparently Fi doesn't warrant a dedicated carrier bundle, meaning Fi subscribers will need to wait until the generic T-Mobile MVNO setting gets flipped on. I've read (on Reddit) that Apple does this because Fi doesn't sell iPhones, but I obviously don't know if that's true. Afaik it took Apple 2 years and a half to enable 5G on Fi, which makes me believe it's indeed Apple being petty, and not some conspiracy theory.

Google Voice doesn't have RCS either.

This is generally a good chronological recap, however, the confusing state seems to arise from an unnatural focus on finding drama / attributing motives / a lens of corporations personified playing chess with full information at all times, and then getting confused about what these inferred chess moves mean.

Ex. Some MVNOs haven't updated their iOS cellular with apple yet because it's an obscure technical thing for a feature most of them don't know exists, yet, for an OS that came out...3 weeks ago? It's not confusing or indicative of some grand plan on anyone's behalf

My flag went up for "need to read this closely, they're substituting facts for lengthy contextual verbiage that feels true" as soon as it's talking about gTalk doing a video call a year before FaceTime. The implication that this went away, and more importantly, the implication Google "lost" to FaceTime, and even more importantly, framing not buying WhatsApp as a "miss", are unambiguous signs of immature commentary.

Smaller aside, not intended as a singular reason to find it lacking basis: I'm no Google fan, but over the years, I've learned to ignore handwaving about how the Nth rebrand of hangouts was somehow leagues ahead and then disappeared and there was a brand new one that lost meaningful core features.

I get the Google service stuff is confusing and not good stuff, but truth is you could start a gChat Gmail Chat gTalk etc. in 2009 and have it in your Google Chat in 2024.

Nobody cares because everyone in the US weirdly uses iMessage while the rest of the world uses either WhatsApp, WeChat or Line.

Honestly, at this point, would probably be easier for Google to just push WhatsApp in the US despite it being a Meta app if they really want to break iMessage stronghold.

Google sells RCS services, and they botched their way out of the social market. For a company which depends on ad targeting that’s pretty huge and Meta is hardly going to help their biggest competitor in advertising but the phone companies are far more negotiable.

Google don't want to "break the iMessage stronghold". They want to crack open Apple customers' messaging so they can feed it into their advertising systems. WhatsApp is useless to Google for that.

Maybe Google wants their users to have a decent, interoperable phone-base messaging solution?

They don't make any money or sell any ads via text messages with Android users, why would they suddenly with iPhone users?

Clearly Apple uses iMessage to create a significant lock in effect in the USA. So, yes, as their main competitor, Google should want to attack the market share of iMessage. It’s probably the biggest thing standing in the way of Android. Pushing a competitor with great interoperability between Android and iOS would make sense.

I mean they clearly believe that they can do that by themselves in the US market but WhatsApp has the advantage of being able to lean of expatriates and people with contacts abroad as a steppingstone.

If you really think that the only reason people use iPhones is iMessage then you are completely detached from reality.

I didn't see OP write that at all. You should also re-read the HN guidelines.

That wouldn't send a ton to data into the hands of Google.

Doesn't matter what Apple does, they blinked, now they have to catch up.

"Over the top?" Come on, use descriptive terms.

That's the term: https://en.wikipedia.org/wiki/Over-the-top_media_service

It doesn't make sense, because "the top" is not defined.

Regurgitating dumb terms merely perpetuates their use; it doesn't make them informative.

Another great (non-tech) example: "red flag warning."

He is using estiablished terms.

https://en.m.wikipedia.org/wiki/Over-the-top_media_service

Puts on tin foil hat Established by a cable company lobbyist to disparage their internet-based competitors. Makes it sound like netflix is hopping over the top of the gate, to the senators they were begging to ban it. Source: made it up.

You needn't make comments like this when a quick google would tell you that the CRTC - Canada's FCC counterpart - popularized this term 13 years ago.

https://crtc.gc.ca/eng/publications/reports/rp1110.htm

Ah yes, a government agency dedicated to new business regulations, no lobbyists involved writing those. That's why "over the top" is still a common term for "internet based", we didn't all have to google the arcane term, and it didn't fall into disuse soon after cable/telco companies lost their legislative battle to monopolize media distribution. Sorry, I'm leaving my tin foil hat on for this one. The term has a smell.

I think apples adoption of RCS is good, but it lacks IMHO two huge things. 1. A better warning to users (other than being green) that RCS messages are NOT end to end encrypted and that your conversation is an RCS conversation. 2. Force read responses from the other party should be enabled.

Small other complaint, anyone using SMS should get auto ignored on Apple now, or at least a setting for auto-ignore messages that aren’t iMessage or RCS

1. Being green is fine. Why do you need a second differentiator? Green has always been without E2EE and continues to be without. The last thing I want is some extra icon adding visual noise.

2. Forcing read receipts, you mean? Absolutely not. The ability to turn off read receipts is vital for privacy. Nobody should be forcing anything there.

> Small other complaint, anyone using SMS should get auto ignored on Apple now

Please don't. The messages I send fall back to SMS when I don't have good data connectivity, which happens all the time on the go. Why should my messages be auto-ignored because of my connectivity issues?

>Small other complaint, anyone using SMS should get auto ignored on Apple now, or at least a setting for auto-ignore messages that aren’t iMessage or RCS

This would exclude any people who aren't on mobile data all the time. If iMessage becomes an Internet only messaging platform, it would have no compelling edge over Signal or WhatsApp.

It already is an Internet-only messaging platform, and its compelling edge and sole reason why people care about the blue/green bubble thing is being installed by default on iOS.

People care because it’s enormously better than SMS and also better than RCS. Nobody would care if green didn’t mean things like no guarantee of better than same-day message delivery.

People outside of the US have solved this problem by adopting an OS-agnostic Internet-based messaging platform like WhatsApp, Signal, Facebook Messenger, or Telegram. There are one or two people I exchange SMS messages with, but other than that, my SMS inbox is full of automated notifications, and the actual chats happen elsewhere.

So where does that leave the theory that iMessage’s “compelling edge and sole reason why people care about the blue/green bubble thing is being installed by default on iOS”? WhatsApp, et al. are not the default on any platform but appear to be doing fine.

WhatsApp et al. are doing fine and are the primary choice in Europe. But in the US, for some reason, people prefer not installing messaging apps, and instead keep the default messaging app, which means iMessage on iOS.

Most Americans have those apps installed, too, so it’s probably not the app being the default - as it is everywhere - but something else about social graphs. Looking at the data vs. SMS pricing 1-2 decades ago is one explanation I’ve heard but there are likely others.

>Small other complaint, anyone using SMS should get auto ignored on Apple now, or at least a setting for auto-ignore messages that aren’t iMessage or RCS

I dunno man, RCS is pretty unreliable, even in Android-to-Android communications. SMS is your standard fallback.

Honestly I try avoid the whole thing and prefer to use Telegram or FB Messenger, which also have the benefits of reasonably well on computers.

So you're saying I should never be able to text iPhone users? Great.

Doesn't iOS have RCS support now?

I don't have access to RCS from my phone number, but I do have access to SMS.

I would hazard a guess that most iMessage users don’t associate blue == E2EE. Even if it were made explicit, how many would care?

Reply on news.ycombinator.com