Why text messages are limited to 160 characters | Hacker News Reader

Why text messages are limited to 160 characters

latimesblogs.latimes.com

40 pointsbootload51 comments

  Text messaging has become the prevalent form of mobile communication worldwide.

  an average of 357 texts per month in [Q2 2008 vs] an average of 204 calls

I would think this suggests that calls are still the bigger method of communication. The average call will last much longer than what could be written in 160 characters. Also, a ninety second call discussing something could easily span a number of text messages back and fourth.

Edit: New point after reading further:

  Just look at your average e-mail today, he noted. Many can be summed up in the subject line, and the rest often contains just a line or two of text asking for a favor or updating about a particular project.

First, I don't accept the premise. Trying to suggest that 160 characters is more or less enough for the average email is laughable. Hell, a lot of the time it's not even enough for an SMS message or tweet.

Second, why are they even being compared? Is SMS meant to be just the same as email but on your phone? Not in my book...

> Second, why are they even being compared? Is SMS meant to be just the same as email but on your phone? Not in my book…

They are both ways to interact with others. The argument here is that most communication do fit in 160 characters when pared down to its essential components, and thus for most communications SMS are sufficient to replace emails.

Now say that in 160 characters. :)

Both are communication channels. 160 characters is enough for most thought to come across. Therefore SMS can replace email in most cases (though not all).

calls are still the bigger method of communication

I coined an expression: "one lunch is worth a thousand tweets." It got a good laugh from the audience at the last founder showcase. What I really mean of course, is that human contact is making a come back. Yes, you can tweet and text. But at some point, it feels all so detached. You'll crave actually meeting people in person. I think that will be a key trend of 2011, "back to reality".

Running some code over War and Peace from Project Gutenberg gives these numbers for sentence length. I cleaned the text of non-alpha | {sentence terminators}.

Then I discarded any sentences with less than 4 characters under the assumption that they were spurious (ie, artifacts from the boilerplate PG text).

Text with artifacts assumption:

STDEV: 95.0115812751

MEAN: 74.4095505188

N: 26681

Text without artifacts assumption:

STDEV: 82.3010550235

MEAN: 76.425092356

N: 30805

Looking at the high point of the first standard deviation of the artifacts-removed text, we see that it is at 169 characters.

Same point for the text with artifacts, we have 158 characters.

So assuming that the translated War and Peace represents a reasonable sample of English, 160 characters is fairly close to the length of a sentence.

edit: formatting.

> So assuming that the translated War and Peace represents a reasonable sample of English, 160 characters is fairly close to the length of a sentence.

I would say the mean is close to the length of a sentence, and the first high mark via standard deviation is a long sentence.

So a standard SMS should be able to fit 2 or so normally worded sentences or 3 concisely worded sentences.

IMO post-Hemingway English has shorter sentences on average than it used to; I wouldn't mind targeting, say, Twilight with my code and getting the stats on it, but I am loath to steal Twilight just for personal amusement.

"First, I don't accept the premise. Trying to suggest that 160 characters is more or less enough for the average email is laughable. Hell, a lot of the time it's not even enough for an SMS message or tweet. Second, why are they even being compared? Is SMS meant to be just the same as email but on your phone? Not in my book..."

Have you considered the idea that maybe your writing is just too wordy? You could easily reduce your 327-char comment to 110. Look at what I did below - much better!

SMS is not long enough to replace emails. Often, it's not even long enough for me to write a thought or tweet.

This is kind of a weird article, it almost contradicts itself a bit.

As the article mentions, SMS'es are kind of piggy-backed with control signals sent on the network. At first they found they could tack on a 128 character message, but then reduced the character set further to get to 160 characters.

What isn't clear in the article is when this 'proof' that 160 was enough occurred. Did it occur after they figured out they could jam 160 in, in order to validate that 160 would be useful enough? Or did it really occur in the order that the article talks about, that they first picked 160 characters out of the sky, then managed to jam that into the signal.

I find the latter to much less likely. 160 just happens to be the most you can fit in to the packet, which he then validated was probably enough to be useful.

Anybody know definitely otherwise?

Here is part of what the Wikipedia article has to say, which appears to show that the size limits were indeed the result of the existing design, and so the only 'proof' here was that the existing amount of data would be sufficient for the intended purpose. Which makes the linked article another example of the mainstream press getting themselves all confused when it comes to cause and effect in technology. A key sentence in the Wikipedia article was that SMS, so limited, could be implemented with software changes only, making its rollout by far easier and more likely to occur. http://en.wikipedia.org/wiki/SMS:

"The innovation in SMS is Short. The GSM is optimized for telephony, since this was identified as its main application. The key idea for SMS was to use this telephony-optimized system, and to transport messages on the signaling paths needed to control the telephony traffic during time periods when no signaling traffic existed. In this way, unused resources in the system could be used to transport messages at minimal cost. However, it was necessary to limit the length of the messages to 128 bytes (later improved to 140 bytes, or 160 seven-bit characters) so that the messages could fit into the existing signaling formats.

This concept allowed SMS to be implemented in every mobile station by updating its software. This concept was instrumental for the implementation of SMS in every mobile station ever produced and in every network from early days. Hence, a large base of SMS capable terminals and networks existed when the users began to utilize the SMS."

Having spent two and a half years of my life getting intimate with GSM-based (ie, not CDMA, though a lot of the protocol stack is the same) SMS messages (I worked for a company that did security, anti spam, anti fraud and other VAS systems (parental control, blacklist, SMS backup etc) for mobile operators)[1], I know a little bit about it. Admitedly I could tell you a lot more about the protocols used by SMS than its history

GSM networks consist of two signalling links: voice and control. The contol is used to, for example, set up or terminate voice calls. This meant that, in practice, the control signalling links were generally idle and so they piggy-backed SMS on it. I don't know anything about the 128 bytes, but the 160 characters comes from the 140 byte payload, when using the 7bit GSM default packed alphabet. I don't know the numbers offhand (been a year since I've worked there now), but quite possibly this is, indeed, the maximum packet size available, though there certainly is a large range of differently sized packets. In fact, a lot of things are piggy backed on the SMS payload. For example, when using concatenated messages (or binary data - there was a range of predefined things that are biggy-backed on SMS as "user data"), the header telling the phone which message it is is placed at the start of the message payload (reducing the maximum characters per text by (I don't remember offhand anymore) ~6 octets. Of course, 160 characters only applies to the 7bit GSM default alphabet. SMS messages also support 8bit and 16bit payloads.

Mobile operators absolutely love SMS as it is probably the most expensive form of communication they offer.

Somebody else below mentioned ubiquitous internet as an SMS replacement. While I think its true for a lot of people in well built-up parts of the world (and who have a smart phone, a lot of people still don't), its certainly not global. Hell, even here in Ireland theres areas where cell phone reception is pretty bad, so while you might manage to send or receive text messages, if you want to go online you're out of luck. Other, poorer, parts of the world are even worse off. Actually, I remember one company at the 2009 GSMA Mobile World Congress in Barcelona was demonstrating a solar panel powered mobile cell tower, for use in places like Africa, which was designed to provide better coverage of mainly SMS to large open spaces.

[1] http://dublindan.posterous.com/things-ive-worked-on-4 We did a lot using transparent Home Routing http://en.wikipedia.org/wiki/SMS_home_routing The wikipedia article says that transparent home oruting is limited. I disagree. At least in the lab, we implemented everything from the technical realization (and then some) on a transparent home routing system. I believe when I left the company, others were working on a non-transparent version too (because it is easier/cheaper to get and install hardware for non-transparent operation than for transparent). But then, we did a lot of non-standard things, we just did them in ways that don't interfere with the rest of the network.

EDIT: comment didn't end up as interesting as it sounded in my head. Hopefully somebody found it useful nontheless.

I thought that SMS was initially just a hidden feature that engineers could use to test the phones - validating that control signals were coming through. Some teenagers found it, and discovered that it wasn't billed, and started using it. Then the telcos caught on, and realized that a new product was born.

Mr. Hillebrand is not a rich man, although given that the text message is a number 1 overpriced product (http://www.walletpop.com/2010/09/27/10-most-overpriced-produ...) it seems that his work has made huge amounts of money for the others.

I'm surprised that Printer Ink Cartridges don't appear on that list.

I can confirm that mobile operators absolutely LOVE SMS because it is so highly overpriced for the amount of data you get to send. They love it so much, in fact,t hat they are willing on spending large sums of money on services to improve the SMS network in some way (eg, we did anti-fraud and anti-spam stuff, amongst other more niche things).

Slightly off-topic, but it's a little mind-boggling that they charge 10 cents per message for something that rides on control signals and gets the lowest priority on the network.

Here in Kenya we pay between Kshs 1-2 per message to send.

Receiving is free.

Considering that the exchange rate is Kshs 80 to 1$ a message is about 1.25 to 2.50 cents.

In the UK you can combine several texts together seamlessly so this limit is meaningful only with respect to billing, and most plans provide so many free texts that it simply isn't something you think about - is that similar in the US?

Yes for modern phones, although I can't say how many free texts is considered enough. If you're not on an unlimited plan and you start texting with someone who is, you can quickly stack up hundreds of texts without realizing it...

True, and that is of course applicable here in the UK too, however it is rare, in my experience at least, to see people being too concerned about text length. In the past people would be very anal about using 'text speak' i.e. hi m8, how r u? partly to fit into the char limit.

The feature itself is a feature of GSM and part of the GSM specifications (see lower in the comments thread, I am not sure this also works on CDMA and remember having read it didn't, but can't find the source again. If anybody knows, I'm interested), so it is available at least in GSM networks (AT&T and T-Mobile).

On the other hand, I do not believe US operators to be quite as generous as UK (or more generally european) ones when it comes to SMS bundles, though I can be wrong: I at least know that many (most?) US cellphone customers pay not just for outbound but also for inbound text messages. And from checking the Verizon and Orange UK websites (I'd have checked O2 but i can't access their shop from outside the UK):

* At Verizon, texts are either 20c each or bundles of $5/250 texts (2c/text) or $20/5000 (0.4c/text), there are apparently unlimited options with the $10 and $20 bundles but they're only between VZ phones within precise coverage areas ("National Enhanced Services and Extended National Enhanced Services Rate and Coverage Areas")

* At Orange UK, two out of three plan lines (Dolphin and Canary) have unlimited text messaging from the start, Racoon (apparently the one used to talk to your grandparents) and value have limited amounts of texts (an unlimited texts option is available for £5, prices for out-of-plan texts are similar to VZ's). And I believe incoming texts are not double-dipped (only the sender pays, unless it's an international text maybe)

Be careful to be aware of the limit to that though. IIRC the standard allows up to 255 messages to be concatenated, but devices impose much smaller limits.

While this usually manifests as a limit that you can't pass, or if your phone sends a message longer than the receiving phone like it will manifest as a split message, my Android phone switches to sending a multimedia message once what I'm wittering about doesn't fit in three SMS messages.

This means it might cost you depending on how your provider charges for MMS messages (in some cases one MMS might be chargeable where several SMSs would come out of a pre-paid/free allowance), and it will cause hassle if the receiving phone (or network) does not handle MMS well (or at all).

Somehow this reminds me of the story that the length of audio CDs was chosen so that Beethoven's Symphony No. 9 would fit onto one CD (see for example http://www.snopes.com/music/media/cdlength.asp )

Seems it may actually be more than just a myth.

See the penultimate paragraph on http://www.dutchaudioclassics.nl/?strPage=Info&strBrand=...

Wikipedia also gives http://www.aes.org/historical/store/oralhistory/?code=OHP-01... as a source for the Beethoven story being true, although you have to buy that interview to actually hear it.

U.S. mobile users sent an average of 357 texts per month in the second quarter of 2008 versus an average of 204 calls, the report said.

The average person sends ~ 10 text messages and makes 7 calls a day?

Wow, I'm behind the curve ...

the mean isn't a robust statistic. i don't think your inference is correct in this case.

a single user can theoretically throw the mean off from the central tendency completely. i would imagine there are a few users that send a lot more text messages and calling more than most people, biasing the result. it would be more useful to talk about the median in this case, i think.

http://en.wikipedia.org/wiki/Robust_statistics#Examples_of_r...

Agreed. I know many ppl who send 1000 or more texts per month. But I don't think there exist many in the world who make that many calls for personal use.

What an anachronistic technology in a world of ubiquitous internet connections. I, for one, hate SMS.

a world of ubiquitous internet connections

Haha, I dunno what world that is, I regularly travel to places where there's no GPRS/EDGE/whatever but there is good ol' GSM. Or sometimes there is data but it's prohibitively expensive.

Incidentally if you hate SMS you probably just have a crap phone. BlackBerry for one presents the same interface for SMS, IM, whatever, the underlying transport is irrelevant to the experience. I guess iPhone and Android are the same.

> I guess iPhone and Android are the same.

Nope, the interface is not unified, they have no equivalent to BBM.

Though in my experience, their SMS experience is quite acceptable.

I can tell you what world that is. It is the 'old world', Europe. I live in Germany. I have an iPhone. I have 'net access virtually everywhere including all of western Europe. I pay 30$ per month for it.

I have lived in the US for a while, and I can tell you: US mobile carriers offer incredibly bad service for incredibly high prices. All of western Europe is way ahead of AT&T or Comcast in these regards. I have not had a single 'dropped call', ever. This problem just does not exist outside the US (except when you enter a tunnel, of course).

Hence, I can always send emails, or even Facebook or Twitter updates from anywhere. Why would I want to send an SMS instead?

Your "old world", western Europe, is an outlier. Head out into the middle of nowhere in northern Norway and see how your Facebook works.

You could argue that Norway would be 'Northern Europe', but I guess that is hypocritical.

Germany, France, Switzerland, Austria, Netherlands, Belgium and England have very good coverage, though.

Those countries combine for approximately 242.63 million people and 1,236,644 km^2, or 3.5% of the world's population and about 0.8% of the world's total land area.

Perhaps England has gotten better of the past few years, but I found that 3G coverage in northern England (outside of larger towns and cities) was very spotty about 3 years ago. Hell even GSM was spotty in quite a few place.

In Ireland the service is much worse than mainland Europe. Also, "anywhere" in your last sentence does not include large parts of: Africa, The middle-east, Asia, Central and South America, large parts of Australia. So, the problem does exist outside of the US.

Ok, Ok, I get that you mean in equivelent non-US places the problem doesn't exist.

I can fully confirm this (I've too lived in Europe and the US).

Europe (Austria): I currently get data traffic at 4 Euros per GB. No monthly costs. Alternatively I can also get 9 GB for 8.80 per month. Also possible in combination with 4 Eurocents per minute for voice calls - of course also without monthly costs (or alternatively: 8.80 Euros per month for 1000 minutes AND 1000 SMS). Usually you've got cellphone coverage _everywhere_ (this also includes many mountain tops in the Alps).

US: I wasn't able to buy just a GSM SIM but had also to buy some cheap phone to get the SIM. Data rates are just crazy - something like $20 for 100 MB IIRC. Coverage is quite poor outside of metro areas: In Austria I would get upset if I don't have 3G coverage, in the US I was happy when I had GSM coverage. And of course, T-Mobile and AT&T use incompatible 3G frequencies - meaning if you buy a phone with one of them you cannot use it with the other one.

Extrapolation from a small and rich country like Austria to the entirety of Europe is very dangerous and quite likely to give inaccurate results.

So you are arguing that the US should actually be rated a 'poor' country by central/western European standards.

What? No not at all. I've said nothing one way or the other about the US.

Ubiquitous internet is decades away. You don't have to wander far off the beaten track for SMS to be the only viable form of mobile communication. SMS is great in that it works in places where the signal is too weak for phone calls. Also sending an SMS is a good order of magnitute cheaper than sending an email when roaming.

They're not limited to 160 characters, any GSM phone is able to create and receive segmented SMS (a single message over multiple segments), up to 255 segments of 134 bytes.

I have read CDMA phones are not able to perform such an amazing feat though.

Is that true? I’m sure it is nowadays, but I remember getting a phone upgrade (from a previous GSM phone) that had this ability.

User Data Header => UDH Length + Information Element Identifier + Information Element Length + Concatenated Message reference number + Sequence Number of Part + Number of Parts. Each is one octet, for a total of 6. Hence 134 octets, Note that the user data header can, conceptually (I have not seen it in practice, but the spec allows it) contain multiple information elements of varying sizes.

I don't know much about CDMA, but from the little I did interact with it, the upper levels of the protocol stack (TCAP upwards) looked almost identical to me. Does this mean that CDMA "phones" never implemented this or is it a protocol issue?

> Does this mean that CDMA "phones" never implemented this or is it a protocol issue?

I do not know, I just read recently that CDMA had no segmented SMS (in relation to the CDMA iPhone I believe, but for the life of me I can't remember the source). I just threw it out there in case somebody had the knowledge to infirm or confirm it.

If it was in relation to the CDMA iphone indeed, I'd expect it's either in the baseband or a protocol issue, since GSM iPhone has no issue with segmented SMS (in either production or consumption)

It would be very strange if the phones just didn't implement it, so at a guess I'd say its a protocol issue too, but as far as I remember, the upper levels of the protocol stack were the same between GSM and CDMA so I don't see why the protocol wouldn't support it (since the user data is inside the message payload itself and doesn't interact with the rest of the payload. I guess CDMA must not support a user data header for some reason?

I'd be very interested in hearing if this is true or not, if anybody knows.

Googling a bit, I found the following on an SMSC's website:

> In CDMA environments, no standard for segmentation exists. Therefore, the WAP Forum defined its own standard for the sending of segmented WDP messages in a CDMA environment.

> A key requirement of this standard is that each segment of the message must use the same CDMA SMS message id. However, the CDMA SMS message id field is a value that is generated by the SMSC, and cannot be set by an application that is submitting messages to the SMSC, such as an MMSC or WAP Push Proxy Gateway.

A number of email threads seem to indicate concatenated SMS are available on some CDMA phones but not others (and most notably appear missing on a number of CDMA Blackberry devices, to the point that there are software built for auto-splitting: http://www.mobihand.com/product.asp?id=28752&n=Beyond160), I have found no indication on whether the process was as seamless as on GSM, but the quote above seems to indicate the limitations are quite heavy and segmented SMS are not supported by CDMA itself.

Oh, interesting. Thanks for that.

I only worked with GSM messages myself, so I don't know how it compares to CDMA. From what little I did see of CDMA I thought they would be fairly similar, but I guess theres more involved. GSM is a pretty versatile system, though.

Any modern GSM phone, certainly. I had a few phones in the late 90s/early 2000s which could not do so (an Alcatel, an Ericsson, and a Sony), although it was pretty trivial to notice the garbled characters at the start of an SMS and wait for a second/third incoming SMS :)

I wonder whether the GSM spec got updated, or an RFC-style process took place, or simply that one phone manufacturer implemented the functionality in a sensible manner and the rest followed?

Because cellular providers love massive piles of cash?

Reply on news.ycombinator.com