See Google Talk and Skype, both of which can do voice-only calls and video calls over 3G. It's not all perfect, though...
The problem with Skype on a mobile is that keeping a connection to their P2P network requires much more power than the usual push notification channel does. It's more expensive computationally and traffic-wise.
Google Talk, in comparison, needs less power, but the audio and video quality is just much worse (on my Nexus S with ICS). I also have fewer contacts in GTalk than in Skype. And for some reason, on many phones, there is no audio and video support at all.
Agreed. Over half of my calling is via skype or other voip service. The evolution is happening...but a clear winner has yet to be identified. Perhaps there will never be a clear winner...
This is called a phonebook, and people don't always want their numbers to be published in it. Also, how do we uniquely identify people? This post is so badly thought through.
IMHO, I haven't heard anyone telling/writing down their actual phone number for quite a long time, what I see happening a lot it's stuff like "you want my phone number? sure, let me give you a short call so that you can save it in your's phone's address book" (with the option "you want my Vodafone or Orange number"? I haven't actually never seen anyone changing their SIM cards, people just prefer to own 2 or more phones if it best suits their needs. Phones are getting damn cheep, at least in Eastern Europe).
The phone number exchange protocol you describe seems to demand that at least one of the persons is spelling out his/her phone number out loud ...
What I would expect is a protocol where both just should be able to connect their phones on short distance (bluetooth, whatever) and acknowledge an "exchange phone number with <person>?" dialog ...
Nokia phones have a method for such convenience. A contact stored in one phone can be sent to another phone as a special sms message called a 'Business Card'. When the other Nokia phone gets it, they can save the contact from the message view options. Of course, it's not free, or local (bluetooth).
Bluetooth supports vCards exchange, but it's much faster and arguably easier to do what GP said - one person spells their number, second person calls it.
>I haven't heard anyone telling/writing down their actual phone number for quite a long time, what I see happening a lot it's stuff like "you want my phone number? sure, let me give you a short call so that you can save it in your's phone's address book"
It's already a solved problem using using SIP. Some of my friends have SIP address. I can call them anywhere using the same address. Home phone, mobile, even using VoIP on a laptop.
Serious question: How easy is it for your friend to call you back immediately from the same device? An hour later from a different device? A week later from any device? If the convenience doesn't flow both ways, it's not a solved problem. [It might be, I really don't know.]
The return path works just like normal. Just hit the dial-back button on your phone or VoIP software and the phone that I used to phone you will start ringing. It gets interesting when we both have SIP/ENUM. I could call you from my mobile when I am at home and, when you ring me back, my home phone could ring because my SIP preference says to use that first.
The only tricky thing is your own SIP/ENUM preference. It needs to know where incoming calls need to go. But that can be automated (e.g. your phone sees your home wifi network and thus sets the preference to ring the home phone first).
Other neat stuff is possible too. One of my coworkers has linked his home phone to his cellphone. When you call his home number, both his home phone and cell phone will ring, even when he is at work. Whichever device answers first gets the call.
"There are only two hard things in Computer Science: cache invalidation and naming things." -- Phil Karlton
There is not an easy solution to the "problem" that this article describes.
The current situation allows us to look up numbers without letting some central authority know what numbers we are looking up and without letting them have our address book. Any solution to this "problem" is likely to reduce everyone's privacy by getting rid of one of these features.
EDIT: You can contact me using the same "username@domain" style address for email, XMPP and SIP. I guess that's something. People still need to store my "username@domain" address in an address book though.
Horse shit. There is nothing wrong with telephone system or the numbering scheme or the fact that we use numbers.
It's simple, doesn't require lots of technology (a analogue phone handset only has a few components in it) and it is ubiquitous across the entire globe.
The moment you add a DNS-type layer or indirection to it, it becomes more complex. Handsets become more complex, the network becomes more complex, usability declines.
The problems of remembering or writing down a name versus a number are far greater.
This is typical technophiles ignoring the minimal needs of about 90% of the populous of this rock.
Adding a DNS systems doesn't remove the phone numbers, not any more than DNS prevents you from typing IP addresses.
You could very well have it on smartphones and the likes without increasing complexity on either the actual phone network or the basic analog handsets.
EDIT: Personally, I think what Google Voice does - giving you a single number per person instead of per device/SIM card - fixes the major problem(s) without breaking backward compatibility. The only problem is the lack of international support.
Yes, but DNS also assumes there is exactly one match for a given text string.
In RL, there are a lot of "John Smiths", and it would be non-trivial for a system to figure out which one you wanted to call without some serious hinting (like access to your Facebook, LinkedIn, Twitter, G+, etc. accounts), and it might even have to access a private list of names/numbers to figure out that you probably want the John Smith who works with Alice Bell (an entry in your phonebook)...
Why not use the same system any web service in existence already uses - nicknames? Any text string is still easier to remember than an arbitrary number.
I've been using the nick 'brk' for maybe 20 years, but I still frequently come across random places where it has already been taken.
So, you'd need a nick registry, like domain names, and you'd run into issues with your friends that have "common" nicks having to register and use a different name. I think this merely trades one well-known set of limitations for a much different and less understood set of limitations without actually truly solving anything.
How do you spell that no wait did you say M is that upper or lower case how do I type an underscore into this thing oh the word underscore was that .net or .biz too late I typed .com automatically.
The places I've seen that give automatic email addresses have used a formula like first initial, last name, autoincrementing number. I think this is a pretty good system, but if you go worldwide you're going to get nice big numbers - entering name and number almost exactly like you do already. But if you misspell something based on a verbal communication it breaks harder.
How is that any simpler or faster than typing a number? One of the things I still miss to this day when switching to an iPhone was how quickly I could make a call on my old Motorola. I didn't even need to look at the phone. I've since switched to an Android device with a physical keyboard, and that's made typing a lot easier. But I really wouldn't want to have to type in some fully qualified name every time I wanted to make a call.
I guess you'd work around that by having a local address book on the phone, but now I've barely gained anything from that, given I already have a local address book on my phone.
As the parent commenter, I upvoted you. Yes, you can say that of most technological "advances". I agree entirely.
Very little genuine improvement happens in technology. Most of it is false starts. When it does happen, everyone notices it and everyone uses it.
Some major technological advances in the last 50 years: The Internet, Mobile Telephones, LCD screens good enough to replace CRTs, Vehicle safety improvements.
The telephone paradigm exists virtually unchanged in its mobile incarnation because it was that good to start with. Someone just got rid of the wires.
I can also agree that since phones and the global system is important infrastructure, it would probably be a disaster to disrupt it for the sake of a modest convenience boost that is only applied to the world's wealthy. So, the suggestions aren't there yet.
But, man, haven't you felt the phone number system is 10 times as archaic now that cell phones are around? I can never call my credit card company with questions, because they ask my phone number for verification, and I can't remember which is on file (/slight exaggeration)! That became such a problem that governments decreed people were entitled to keep their number when they switched cell providers. We want to be tied to a single, convenient contact identifier, I think.
You are just wrong. We might want to have our own contact identifiers because that is how the system works and it's convenient, but that's a symptom of the problem.
Let me phrase it like this: whose phone numbers do you actually know? If you're like most people, the answer is "almost nobody's." You identify friends based on name and context -- and sometimes you can pull an address out of your hat. The mapping to a phone number is accomplished by an address book. And that's true more broadly: the reason you want to keep your number is because you know there are a lot of pointers in a lot of address books containing this number, and you'd like them to keep pointing to you. You like your friends and don't want to hassle them with updating their phones.
The problem is foundational: given a description of a person and the contexts from which you know them, it is not easy for a computer to accurately reconstruct who that person was. We could solve the entire rest of the problem, actually, pretty easily. We could, if we wanted to, create this system: you register with a system and it creates a new business-card dropoff for you: "up8 m2x d4z". Feel free to register a couple of these if you want; there's something like 100 trillion of them and we're assigning them in pseudorandom order. Your online anonymous alias can be "h3c if3 fdc". That simple ID string points to a publicly-available business card: name, phone, email, current employer, photo, fax, RSA public key, dog's FTP server -- whatever you want or don't want to put there. Any address book can query it, either via the phone network or via the Internet or whatever else have you.
So the layer of abstraction needed is no more complicated than phone numbers are now, and if you worked out how to pay for it and scale it, then mobile phones could immediately query present or near-present data on the fly rather than demanding that it remains absolutely constant.
But the problem is really search. Looking up 't6o 19r 4q2' might be very easy for computers and might be very easy for humans to communicate, but when I want to say "crap, I lost my address book, what's my friend's mobile phone number?" I am still left checking Facebook and crap, they didn't put it up there, maybe their LinkedIn has it? No they don't have a LinkedIn. Hm. Maybe I should call someone else who knows them...
Names are far from unique identifiers, so people created standards for uniquely identifying people, you now want to create another one well guess what: http://xkcd.com/927/
The reality is it's not a technical problem in need of a better standard. If you meet someone at a party they can give you a phone number, an email address, a Facebook act, a linked in act, IM, skype, etc etc. But the universal system is simply Identifier:Service and we already have that and it works really well. Some of these come bundled with other Identifier:Service groupings, but just because I met you at a party does not mean I want to give out all of them. Backing up your address book is a side issue with simple direct solutions that are up to you to implement.
> Horse shit. There is nothing wrong with telephone system or the numbering scheme or the fact that we use numbers.
There is never anything wrong with the technology you know. I think the Unix shell is a thing of beauty. You have a user interface and programming language all rolled into one, available on even the simplest of hardware. Put it in front of the average computer user, however, and they'll look at you like you have two heads.
I do not know the telephone system. For me, the internet has pretty much always been there so I had little need to use the phone and I still only use it a handful of times per year. Making a phone call is as confusing to me as it is to someone who has to dive into the command prompt once per year. On top of that, it seems like every time I try to use it, it fails in some way (disconnects, misrouted calls, etc.), adding to my frustrations.
The telephone system is pretty ingenious from a technical perspective, but it is a user experience nightmare for anyone who did not grow up with it. I know exactly how older people feel about computers that did not grow up with them, because that is me with the telephone.
We are at least trying to make computers more accessible, why are we ignoring the telephone?
Honest question: how does the telephone system confuse you?
-Dial a number, receive a person (or voicemail).
-Most people don't even dial numbers any more. Select contact, press send, receive person.
Do you use a cellphone or landline? Are you in the US? I'm on (crappy) AT&T service with a cell phone and haven't had a dropped or misrouted call...ever.
Confuse may not be the best word, intimidated maybe, but I don't know. I realize it is just the execution of basic steps.
As per my previous example, how does if [ -r /etc/passwd ]; then cat /etc/passwd; fi confuse someone? It too is just simple, repeatable steps.
> Do you use a cellphone or landline? Are you in the US?
I have had problems with both, though cell does seem more prone to it, even in strong reception areas. This is in Canada, spanning multiple providers. I find it actually quite amazing how often it does fail for me, though I recall reading some stats not too long ago that showed the number of calling errors to be surprisingly high.
The tool free 800 and for-pay 900 systems are already one redirection layer in the telephone network. When you call an 800 number it's actually an SS7 signaling database lookup that then redirects to another number. The same is true for calling cards (remember those?).
In mobile phone systems the telephone number really is a simplier naming scheme. The call is routed based upon some id number for the SIM card/phone which is first looked up based on the phone number you dialed. For GSM this is a 64-bit International Mobile Subscriber Identifier (IMSI) and who wants to remember 64-bit values (IPv6 anyone?). AFAIK GSM says nothing about having a 1:1 mapping between phone number and IMSI, but it's been a long time since I did GSM work.
I am aware of phone number forwarders. I use them for company phone numbers.
As I said with Google Voice it solves the problem to the biggest extend.
But you as an expert might know that…
Are there any services that offer me a Global Number that won't require "international call" charges from international people who call me ?
Would this even be technically possible?
I'm sure there's someone out there that does it, but you'd have to pay the charges instead of your customer. Call routing is designed to make money first and route calls second.
Oh, if only phone numbers were ubiquitous across the entire globe.
Have you tried calling a phone in random foreign countries, going off a local phone number which hasn't been expressly written for foreigners? Figuring out which parts you have to dial and which parts are only for local people gets really annoying.
Additionally, there's no expiration mechanism. The volume has thankfully declined, but I still get calls on my home phone for the guy who had the number before me, even though he hasn't had that number for five years now. The only reason numbers are ever reused is because we have to keep them short enough to be dialed by a human on a regular basis, and the only reason we have that requirement is because we don't have something better.
Yes plenty of times. And I've had to muddle with foreign operators and they've had to deal with my really bad attempts at their languages. At the end of the day I'd got where I wanted to go though and had some interesting conversations on the way.
Where I want to call is somtimes a shack in the middle of some African state with a 60 year old phone dangling off a rusty, noisy aluminium pair with as many repairs as you can imagine on it (my cousin lives out there). It still works as well as it did 60 years ago for them.
One of my parents even had a UK BT area code in their company name. It got changed three times in the 90s. They are stil around. It still works for them.
What people are proposing regarding swapping out or augumenting numbers with something else is a wholesale reengineering of all voice communication networks worldwide. I.e breaking everything for convenience.
Numbers have been "good enough" for the last 100 years and probably will be "good enough" for the next 100 years.
Humans have this strange ability to contract out some of their brainpower so they don't have to think about something or can get around some minor inconvenience. It's almost like a numeric witch hunt in this case.
I'm going to call it "there's an app for that syndrome" roughly translated as "automate something simple so I don't have to use another 5 brain cells".
And yes I get people phone my line asking for the people who used to live here. But it happens less than I get junk mail through my door.
The telephone system does not work the way you think it works.
We could talk about POTS, but that's largely irrelevant. Worldwide, the vast majority of phone calls start and end with a GSM terminal. Each call is a digital, packet-switched connection from end to end and a supremely complex one at that. A call, even a solely fixed-line call, isn't connected through analog switching, but through a series of database queries that establish a route.
90% of telephone handsets in the world are complex digital devices. They each have multiple identities - in the case of GSM, the IMEI of the handset, the IMSI of the SIM, the MSISDN of the subscriber and one or more telephone numbers. This isn't some bolt-on to POTS anymore, it's the default mode. there are whole generations in the developing world who have grown up with telephony but have never even seen a fixed-line telephone.
Saying telephone switching is simple because it uses DTMF signalling is like saying sending an e-mail through Siri is simple because you just have to speak. The hidden complexity is immense. There are already multiple DNS-type layers involved in establishing the simplest of telephone connections.
You are solely considering the latest generation of equipment. There is still a lot of analogue kit (crossbar switches/relays) out there across the globe. There are even small pockets of it in the UK.
POTS is definitely not irrelevant if you throw away the notion that everyone is not an iPhone wielding hipster.
However you miss the point. The existing abstraction of "phone number" is so universally powerful that the underlying implementation, regardless of what encapsulation and routing algorithms used is effectively a logical replacement for the analogue switching.
It's like "virual analogue synthesizers" which are replacements for synthesizers. The abstraction was so powerful, the new technology emulates the old.
I've been using GV for a while now, which is certainly a next step in solving the problem (short of cell phone DNS).
The problem is that I get the impression that GV has stalled.
I still can't forward my GV number internationally, after years of this being one of their most-requested features. Google has GV in the US, and in NL, yet you can't forward between these countries or have a US & NL number associated with one GV account.
It's a bit silly, as many other services offer this feature. It's also easy money for Google.
I'm happy to see someone ranting about this, I couldn't agree more.
This was solved years ago in the VoIP world by ENUM, but it didn't get enough traction because nobody can earn any money with it.
Here is how it works: your phone "number" would be sip:saghul@domain.tld and you'd have a legacy alias which is a regular phone number, for example: +40317105163. Then someone which knows my legacy number may use DNS to get my SIP URI and call me for free:
You can put several things in DNS, the above example (real one) contains 2 websites, a SIP URI and a URL with location information (a link to Google Maps).
I find it rather odd that they didn't group the codes.
Wouldn't it be better if they made something like this: [phone number].[area code].[country code]?
I mean, I don't really have much information on the E.164 standard, but the information o Wikipedia seems to show that at least the country code is separated from the other numbers, allowing each country to have its own structure.
Well, using a reversed nibble format allows you to have a different nameserver at each level distributing the load and ownership. If as a telco, I have the block of phone numbers +14155551000-+141555519999, I can easily have my nameserver delegated for it as 1.5.5.5.5.1.4.1.e164.arpa.
IPv6 and IPv4 reverse lookups work the same way.
Besides, it isn't like users are typing this in directly. Might as well get in all the flexibility you can. You could do things like add extensions and have it route properly without any extra work or have a country with both 7 and 8 digit phone numbers.
On a related note, giving my contact information to you isn't a permission to forward it to someone else (granted in this case without your knowledge). I wish this aspect got more attention.
Firstly, the comparison with DNS doesn't really make sense.
DNS works because there is a one to one mapping between name and number. There is only one johnsmith.com anywhere on the web. OTOH, there are many thousands of John Smiths out in the real world, each with their own unique set of phone numbers (and other things that this guy seems to have overlooked from an "address book", like an actual address, or an email address, or relationships to other people).
DNS doesn't solve the fact that many people called John Smith may have their own websites. They won't all be found at johnsmith.com. This is solved through a combination of search engines (equivalent of a phone book) and bookmarks (which is pretty much the equivalent of a person's address book - so DNS alone clearly hasn't removed the need for address books in the online world).
Secondly, that's all totally irrelevant to the actual issue with Path - whether someone's got access to a complete list of your contacts with all of the details required to contact them, and got this without your permission.
I fail to understand your point. Yes, lots of people share names. But it should be enough to store _one_ handle to get to (=> resolve) your _current_ number.
That might be your domain or email address. Or something totally different.
The major part for me here is not 'do we need digits without meaning', it is 'bind some name to a way of contacting the person'.
DNS as a comparison makes sense for me. I store john@smith.com as a contact. When I want to call that guy my phone figures out how to reach him _and he can update that information_.
Just like you can move your domain name to point to a different number. If the name server is under your control (or your hoster loves you) and ignoring certain limitations of the protocol and buggy caching behavior, you could even change the result of a lookup every minute. That, for phone numbers, would be awesome.
a)that doesn't get rid of the need for address books - even with email I still need an address book to tell me which John Smith I want to email, and
b) that's got absolutely nothing to do with the Path issue - them uploading my address book with a single phone number/phone DNS whatever for my contacts without my permission is just as bad as them uploading my address book with 5 different phone numbers for each of my contacts
While I don't necessarily agree with everything in this post, I do see two problems with phone numbers:
1. Ask yourself, how many phone numbers do you actually know at this very moment? I'm willing to bet, on average, it's less than five. It might be slightly higher on this site, but for the average consumer, I'm guessing it's between 1 and 3. I ask people this all the time, and I know many who don't know a single number other than their own, or their parents.
2. You've all seen the Facebook groups that pop up when someone loses their phone. "I lost my phone, give me your number!".
Apple is doing a very good job of solving this for the masses right now with iCloud, but you have to be an iPhone user to take advantage.
That's a new phenomenon though, people had to remember more phone numbers before mobile phones or built-in address books.
From muscle memory alone, I can still finger dial about a dozen phone numbers belonging to friends when I was a kid, even though it's been 28 years since I last called them. Funny thing is that I can't recall the numbers without mock dialling them.
I have solved this problem a slightly different way, and I realize that it is not perfect.
90% of my business is conducted within the US, here I have a single number I give out. This number is managed by Ribbit (www.ribbit.com , no affiliation).
Ribbit knows a list of numbers for me:
Office desk phone
"Mobile" IP Phone (eg: sometimes it is at my house, or at my condo if I am spending more than a few days there, or possibly even travels with me if I am going someplace for an extended time).
iPhone 3GS which has numbers in the US, Dubai, Caribbean and Japan via multiple SIMS.
iPhone 4 which has a US number.
When you call my Ribbit number, it rings ALL these services simultaneously, whichever phone I pick up is the one that gets the call.
I can take numbers out of the routing loop if I want (for example when I travel, I don't need my home phone to ring and bug my wife).
So, to reach me, you dial a single number.
The downside to it is CallerID. While I can jump through some hoops to use apps and soft phones to get around this, it's a PITA (to me), so if I call you back from Japan for example, you're going to likely see my Japanese phone number as the caller ID, but I get more calls than I make, so it's not a huge problem.
The phonebook to DNS analogy is also off in that we all have "bookmarks" of sites we want to revisit and keep in a handy list. DNS has not solved or removed the issue of list keeping and organization/management.
LOL. Apparently the author is too young to remember phone books.
As an aside: Older folks will remember Sun receiving a lawsuit from the newly disintegrated phone system regarding their use of "yellow pages". They issued an update which took the "yp" prefix out of their commands.
So instead of an address book, we add telephone-dns and then we have a list of bookmarks? What's the upside again? Personally, I'm rethinking this whole DNS thing. I already have a huge list of bookmarks, I don't see why there couldn't just be IP addresses.
It's been a few years since I worked in softphones, but this "problem" is essentially solved with the SIP protocol. Everyone gets an address of the form user@domain. (The protocol scheme is sip: but the beauty of using user@domain is that one can quickly switch to mailto:). Then the immediacy of the communication dictates the medium. Don't need an answer right away? Email. Low priority? IM/Text. Must talk now? Voice.
I believe Office Communicator (or whatever it's called now) was designed with this in mind.
But that doesn't negate the need for an address book. One still needs to know the user@domain address, as others have pointed out.
What, as exactly as possible, is the desired user experience for contacting people for voice calls?
Serious question. All the discussion around whether phone numbers / something-like-DNS-for-phone-numbers-but-with-failover / GrandCentral-like things (GV, Ribbit also?) / Siri / FaceTime are okay (and for which user groups), is damn near impossible without identifying the actual goal state.
Subs eh - this is some one with a trivial knowledge trying to apply one model (dns) to another (POTS).
X.500 was the proposed solution (70's/80's) where you had interconnected directories that would allow a distributed worldwide directory to work. Only trouble is that a full working x.500 system is probably still at the bleeding edge of what is possible.
A Google or Microsoft could do it if they upped their game in terms of rigous adherence to standards by a few hundred % or so.
85 comments
[ 3.0 ms ] story [ 135 ms ] threadIf Facetime worked for just voice calls, and over 3G, everything would be perfect.
The problem with Skype on a mobile is that keeping a connection to their P2P network requires much more power than the usual push notification channel does. It's more expensive computationally and traffic-wise.
Google Talk, in comparison, needs less power, but the audio and video quality is just much worse (on my Nexus S with ICS). I also have fewer contacts in GTalk than in Skype. And for some reason, on many phones, there is no audio and video support at all.
IMHO, I haven't heard anyone telling/writing down their actual phone number for quite a long time, what I see happening a lot it's stuff like "you want my phone number? sure, let me give you a short call so that you can save it in your's phone's address book" (with the option "you want my Vodafone or Orange number"? I haven't actually never seen anyone changing their SIM cards, people just prefer to own 2 or more phones if it best suits their needs. Phones are getting damn cheep, at least in Eastern Europe).
What I would expect is a protocol where both just should be able to connect their phones on short distance (bluetooth, whatever) and acknowledge an "exchange phone number with <person>?" dialog ...
Handy, but the problem is that you still need a number for SMS, and Bluetooth is damn slow, particularly if you're not paired with the other device.
P H O N E C E P T I O N
The only tricky thing is your own SIP/ENUM preference. It needs to know where incoming calls need to go. But that can be automated (e.g. your phone sees your home wifi network and thus sets the preference to ring the home phone first).
Other neat stuff is possible too. One of my coworkers has linked his home phone to his cellphone. When you call his home number, both his home phone and cell phone will ring, even when he is at work. Whichever device answers first gets the call.
There are systems for mapping telephone numbers into DNS, like ENUM (http://en.wikipedia.org/wiki/ENUM), and e164.arpa. in DNS.
There is not an easy solution to the "problem" that this article describes.
The current situation allows us to look up numbers without letting some central authority know what numbers we are looking up and without letting them have our address book. Any solution to this "problem" is likely to reduce everyone's privacy by getting rid of one of these features.
EDIT: You can contact me using the same "username@domain" style address for email, XMPP and SIP. I guess that's something. People still need to store my "username@domain" address in an address book though.
Out of curiosity, where does this come from? And why, having never read this quote before, have I seen it three times in as many days?
It's simple, doesn't require lots of technology (a analogue phone handset only has a few components in it) and it is ubiquitous across the entire globe.
The moment you add a DNS-type layer or indirection to it, it becomes more complex. Handsets become more complex, the network becomes more complex, usability declines.
The problems of remembering or writing down a name versus a number are far greater.
This is typical technophiles ignoring the minimal needs of about 90% of the populous of this rock.
You could very well have it on smartphones and the likes without increasing complexity on either the actual phone network or the basic analog handsets.
EDIT: Personally, I think what Google Voice does - giving you a single number per person instead of per device/SIM card - fixes the major problem(s) without breaking backward compatibility. The only problem is the lack of international support.
It's all about the automation.
In RL, there are a lot of "John Smiths", and it would be non-trivial for a system to figure out which one you wanted to call without some serious hinting (like access to your Facebook, LinkedIn, Twitter, G+, etc. accounts), and it might even have to access a private list of names/numbers to figure out that you probably want the John Smith who works with Alice Bell (an entry in your phonebook)...
So, you'd need a nick registry, like domain names, and you'd run into issues with your friends that have "common" nicks having to register and use a different name. I think this merely trades one well-known set of limitations for a much different and less understood set of limitations without actually truly solving anything.
if its someone you have him in your friendcircle->app
I believe we need a proper open identity system in general but this could be a start.
The solution wont be facebook.com/nickname nor vox.io/nickname
It will be providerOrContextOrWorkplaces.com/name maybe
The places I've seen that give automatic email addresses have used a formula like first initial, last name, autoincrementing number. I think this is a pretty good system, but if you go worldwide you're going to get nice big numbers - entering name and number almost exactly like you do already. But if you misspell something based on a verbal communication it breaks harder.
How may John Smith's are there in your country ?
I guess you'd work around that by having a local address book on the phone, but now I've barely gained anything from that, given I already have a local address book on my phone.
I have a Nokia 6303. I press a couple of buttons and I'm calling the person I want. It's so unbelievably solved.
Thw fundamental problem is the same Numbers are a techincal layer i as a user dont zo know about
Buttons? Bah ;)
Edit: So your reply is essentially: "This is stupid! The way we currently do things is good enough!" and I get down voted?
Very little genuine improvement happens in technology. Most of it is false starts. When it does happen, everyone notices it and everyone uses it.
Some major technological advances in the last 50 years: The Internet, Mobile Telephones, LCD screens good enough to replace CRTs, Vehicle safety improvements.
The telephone paradigm exists virtually unchanged in its mobile incarnation because it was that good to start with. Someone just got rid of the wires.
I can also agree that since phones and the global system is important infrastructure, it would probably be a disaster to disrupt it for the sake of a modest convenience boost that is only applied to the world's wealthy. So, the suggestions aren't there yet.
But, man, haven't you felt the phone number system is 10 times as archaic now that cell phones are around? I can never call my credit card company with questions, because they ask my phone number for verification, and I can't remember which is on file (/slight exaggeration)! That became such a problem that governments decreed people were entitled to keep their number when they switched cell providers. We want to be tied to a single, convenient contact identifier, I think.
Let me phrase it like this: whose phone numbers do you actually know? If you're like most people, the answer is "almost nobody's." You identify friends based on name and context -- and sometimes you can pull an address out of your hat. The mapping to a phone number is accomplished by an address book. And that's true more broadly: the reason you want to keep your number is because you know there are a lot of pointers in a lot of address books containing this number, and you'd like them to keep pointing to you. You like your friends and don't want to hassle them with updating their phones.
The problem is foundational: given a description of a person and the contexts from which you know them, it is not easy for a computer to accurately reconstruct who that person was. We could solve the entire rest of the problem, actually, pretty easily. We could, if we wanted to, create this system: you register with a system and it creates a new business-card dropoff for you: "up8 m2x d4z". Feel free to register a couple of these if you want; there's something like 100 trillion of them and we're assigning them in pseudorandom order. Your online anonymous alias can be "h3c if3 fdc". That simple ID string points to a publicly-available business card: name, phone, email, current employer, photo, fax, RSA public key, dog's FTP server -- whatever you want or don't want to put there. Any address book can query it, either via the phone network or via the Internet or whatever else have you.
So the layer of abstraction needed is no more complicated than phone numbers are now, and if you worked out how to pay for it and scale it, then mobile phones could immediately query present or near-present data on the fly rather than demanding that it remains absolutely constant.
But the problem is really search. Looking up 't6o 19r 4q2' might be very easy for computers and might be very easy for humans to communicate, but when I want to say "crap, I lost my address book, what's my friend's mobile phone number?" I am still left checking Facebook and crap, they didn't put it up there, maybe their LinkedIn has it? No they don't have a LinkedIn. Hm. Maybe I should call someone else who knows them...
The reality is it's not a technical problem in need of a better standard. If you meet someone at a party they can give you a phone number, an email address, a Facebook act, a linked in act, IM, skype, etc etc. But the universal system is simply Identifier:Service and we already have that and it works really well. Some of these come bundled with other Identifier:Service groupings, but just because I met you at a party does not mean I want to give out all of them. Backing up your address book is a side issue with simple direct solutions that are up to you to implement.
There is never anything wrong with the technology you know. I think the Unix shell is a thing of beauty. You have a user interface and programming language all rolled into one, available on even the simplest of hardware. Put it in front of the average computer user, however, and they'll look at you like you have two heads.
I do not know the telephone system. For me, the internet has pretty much always been there so I had little need to use the phone and I still only use it a handful of times per year. Making a phone call is as confusing to me as it is to someone who has to dive into the command prompt once per year. On top of that, it seems like every time I try to use it, it fails in some way (disconnects, misrouted calls, etc.), adding to my frustrations.
The telephone system is pretty ingenious from a technical perspective, but it is a user experience nightmare for anyone who did not grow up with it. I know exactly how older people feel about computers that did not grow up with them, because that is me with the telephone.
We are at least trying to make computers more accessible, why are we ignoring the telephone?
-Dial a number, receive a person (or voicemail).
-Most people don't even dial numbers any more. Select contact, press send, receive person.
Do you use a cellphone or landline? Are you in the US? I'm on (crappy) AT&T service with a cell phone and haven't had a dropped or misrouted call...ever.
Confuse may not be the best word, intimidated maybe, but I don't know. I realize it is just the execution of basic steps.
As per my previous example, how does if [ -r /etc/passwd ]; then cat /etc/passwd; fi confuse someone? It too is just simple, repeatable steps.
> Do you use a cellphone or landline? Are you in the US?
I have had problems with both, though cell does seem more prone to it, even in strong reception areas. This is in Canada, spanning multiple providers. I find it actually quite amazing how often it does fail for me, though I recall reading some stats not too long ago that showed the number of calling errors to be surprisingly high.
In mobile phone systems the telephone number really is a simplier naming scheme. The call is routed based upon some id number for the SIM card/phone which is first looked up based on the phone number you dialed. For GSM this is a 64-bit International Mobile Subscriber Identifier (IMSI) and who wants to remember 64-bit values (IPv6 anyone?). AFAIK GSM says nothing about having a 1:1 mapping between phone number and IMSI, but it's been a long time since I did GSM work.
I am aware of phone number forwarders. I use them for company phone numbers.
As I said with Google Voice it solves the problem to the biggest extend.
But you as an expert might know that… Are there any services that offer me a Global Number that won't require "international call" charges from international people who call me ? Would this even be technically possible?
Have you tried calling a phone in random foreign countries, going off a local phone number which hasn't been expressly written for foreigners? Figuring out which parts you have to dial and which parts are only for local people gets really annoying.
Additionally, there's no expiration mechanism. The volume has thankfully declined, but I still get calls on my home phone for the guy who had the number before me, even though he hasn't had that number for five years now. The only reason numbers are ever reused is because we have to keep them short enough to be dialed by a human on a regular basis, and the only reason we have that requirement is because we don't have something better.
Where I want to call is somtimes a shack in the middle of some African state with a 60 year old phone dangling off a rusty, noisy aluminium pair with as many repairs as you can imagine on it (my cousin lives out there). It still works as well as it did 60 years ago for them.
One of my parents even had a UK BT area code in their company name. It got changed three times in the 90s. They are stil around. It still works for them.
What people are proposing regarding swapping out or augumenting numbers with something else is a wholesale reengineering of all voice communication networks worldwide. I.e breaking everything for convenience.
Numbers have been "good enough" for the last 100 years and probably will be "good enough" for the next 100 years.
Humans have this strange ability to contract out some of their brainpower so they don't have to think about something or can get around some minor inconvenience. It's almost like a numeric witch hunt in this case.
I'm going to call it "there's an app for that syndrome" roughly translated as "automate something simple so I don't have to use another 5 brain cells".
And yes I get people phone my line asking for the people who used to live here. But it happens less than I get junk mail through my door.
We could talk about POTS, but that's largely irrelevant. Worldwide, the vast majority of phone calls start and end with a GSM terminal. Each call is a digital, packet-switched connection from end to end and a supremely complex one at that. A call, even a solely fixed-line call, isn't connected through analog switching, but through a series of database queries that establish a route.
90% of telephone handsets in the world are complex digital devices. They each have multiple identities - in the case of GSM, the IMEI of the handset, the IMSI of the SIM, the MSISDN of the subscriber and one or more telephone numbers. This isn't some bolt-on to POTS anymore, it's the default mode. there are whole generations in the developing world who have grown up with telephony but have never even seen a fixed-line telephone.
Saying telephone switching is simple because it uses DTMF signalling is like saying sending an e-mail through Siri is simple because you just have to speak. The hidden complexity is immense. There are already multiple DNS-type layers involved in establishing the simplest of telephone connections.
You are solely considering the latest generation of equipment. There is still a lot of analogue kit (crossbar switches/relays) out there across the globe. There are even small pockets of it in the UK.
POTS is definitely not irrelevant if you throw away the notion that everyone is not an iPhone wielding hipster.
However you miss the point. The existing abstraction of "phone number" is so universally powerful that the underlying implementation, regardless of what encapsulation and routing algorithms used is effectively a logical replacement for the analogue switching.
It's like "virual analogue synthesizers" which are replacements for synthesizers. The abstraction was so powerful, the new technology emulates the old.
The problem is that I get the impression that GV has stalled.
I still can't forward my GV number internationally, after years of this being one of their most-requested features. Google has GV in the US, and in NL, yet you can't forward between these countries or have a US & NL number associated with one GV account.
It's a bit silly, as many other services offer this feature. It's also easy money for Google.
This was solved years ago in the VoIP world by ENUM, but it didn't get enough traction because nobody can earn any money with it.
Here is how it works: your phone "number" would be sip:saghul@domain.tld and you'd have a legacy alias which is a regular phone number, for example: +40317105163. Then someone which knows my legacy number may use DNS to get my SIP URI and call me for free:
You can put several things in DNS, the above example (real one) contains 2 websites, a SIP URI and a URL with location information (a link to Google Maps).There is an Android application called ENUMdroid which will do a ENUM query for each number you dial and it'll present a screen with it's findings: http://saghul.net/blog/wp-content/uploads/2010/02/CAP2010021...
Apple allows you to use FaceTime and iMessage with email-style addresses, lets see if they can push the model forward :-)
Wouldn't it be better if they made something like this: [phone number].[area code].[country code]?
I mean, I don't really have much information on the E.164 standard, but the information o Wikipedia seems to show that at least the country code is separated from the other numbers, allowing each country to have its own structure.
IPv6 and IPv4 reverse lookups work the same way.
Besides, it isn't like users are typing this in directly. Might as well get in all the flexibility you can. You could do things like add extensions and have it route properly without any extra work or have a country with both 7 and 8 digit phone numbers.
And then I become ChrisLomax103454751983, that's miles easier to remember.
DNS works because there is a one to one mapping between name and number. There is only one johnsmith.com anywhere on the web. OTOH, there are many thousands of John Smiths out in the real world, each with their own unique set of phone numbers (and other things that this guy seems to have overlooked from an "address book", like an actual address, or an email address, or relationships to other people).
DNS doesn't solve the fact that many people called John Smith may have their own websites. They won't all be found at johnsmith.com. This is solved through a combination of search engines (equivalent of a phone book) and bookmarks (which is pretty much the equivalent of a person's address book - so DNS alone clearly hasn't removed the need for address books in the online world).
Secondly, that's all totally irrelevant to the actual issue with Path - whether someone's got access to a complete list of your contacts with all of the details required to contact them, and got this without your permission.
That might be your domain or email address. Or something totally different.
The major part for me here is not 'do we need digits without meaning', it is 'bind some name to a way of contacting the person'.
DNS as a comparison makes sense for me. I store john@smith.com as a contact. When I want to call that guy my phone figures out how to reach him _and he can update that information_.
Just like you can move your domain name to point to a different number. If the name server is under your control (or your hoster loves you) and ignoring certain limitations of the protocol and buggy caching behavior, you could even change the result of a lookup every minute. That, for phone numbers, would be awesome.
a)that doesn't get rid of the need for address books - even with email I still need an address book to tell me which John Smith I want to email, and
b) that's got absolutely nothing to do with the Path issue - them uploading my address book with a single phone number/phone DNS whatever for my contacts without my permission is just as bad as them uploading my address book with 5 different phone numbers for each of my contacts
1. Ask yourself, how many phone numbers do you actually know at this very moment? I'm willing to bet, on average, it's less than five. It might be slightly higher on this site, but for the average consumer, I'm guessing it's between 1 and 3. I ask people this all the time, and I know many who don't know a single number other than their own, or their parents.
2. You've all seen the Facebook groups that pop up when someone loses their phone. "I lost my phone, give me your number!".
Apple is doing a very good job of solving this for the masses right now with iCloud, but you have to be an iPhone user to take advantage.
From muscle memory alone, I can still finger dial about a dozen phone numbers belonging to friends when I was a kid, even though it's been 28 years since I last called them. Funny thing is that I can't recall the numbers without mock dialling them.
The points you describe could also be said about e-mail addresses and website addresses.
This is about people putting too much trust in (electronic) technology.
90% of my business is conducted within the US, here I have a single number I give out. This number is managed by Ribbit (www.ribbit.com , no affiliation).
Ribbit knows a list of numbers for me: Office desk phone "Mobile" IP Phone (eg: sometimes it is at my house, or at my condo if I am spending more than a few days there, or possibly even travels with me if I am going someplace for an extended time). iPhone 3GS which has numbers in the US, Dubai, Caribbean and Japan via multiple SIMS. iPhone 4 which has a US number.
When you call my Ribbit number, it rings ALL these services simultaneously, whichever phone I pick up is the one that gets the call.
I can take numbers out of the routing loop if I want (for example when I travel, I don't need my home phone to ring and bug my wife).
So, to reach me, you dial a single number.
The downside to it is CallerID. While I can jump through some hoops to use apps and soft phones to get around this, it's a PITA (to me), so if I call you back from Japan for example, you're going to likely see my Japanese phone number as the caller ID, but I get more calls than I make, so it's not a huge problem.
The phonebook to DNS analogy is also off in that we all have "bookmarks" of sites we want to revisit and keep in a handy list. DNS has not solved or removed the issue of list keeping and organization/management.
reg bookmarks: bookmarks is nothing you type in and make work. Its shortcuts to your favourites. And you dont bother about technical layers below.
LOL. Apparently the author is too young to remember phone books.
As an aside: Older folks will remember Sun receiving a lawsuit from the newly disintegrated phone system regarding their use of "yellow pages". They issued an update which took the "yp" prefix out of their commands.
And imho: Telephone books are the like blackboards/dial up sites with lists of more sites.
http://en.wikipedia.org/wiki/.tel
I believe Office Communicator (or whatever it's called now) was designed with this in mind.
But that doesn't negate the need for an address book. One still needs to know the user@domain address, as others have pointed out.
Serious question. All the discussion around whether phone numbers / something-like-DNS-for-phone-numbers-but-with-failover / GrandCentral-like things (GV, Ribbit also?) / Siri / FaceTime are okay (and for which user groups), is damn near impossible without identifying the actual goal state.
X.500 was the proposed solution (70's/80's) where you had interconnected directories that would allow a distributed worldwide directory to work. Only trouble is that a full working x.500 system is probably still at the bleeding edge of what is possible.
A Google or Microsoft could do it if they upped their game in terms of rigous adherence to standards by a few hundred % or so.
Stick to the frocks hunni.