back to article Meet the open sorcerers who have vowed to make Facebook history

Once upon a time the internet ran on open protocols, and anyone could host servers that ran these protocols. Your first dial-up internet connection probably came with a bundle of tools for groups and chat. If you weren't happy with the service from your ISP you'd point the client at another. The internet was open and federated, …

Page:

  1. disgustedoftunbridgewells Silver badge

    Why IMAP and not XNMP?

    1. Anonymous Coward
      Anonymous Coward

      Re: Why IMAP and not XNMP?

      Do you mean XMPP?

      Either way, my guess is that as there is already a massive number of IMAP servers already deployed anyway, that this extension can piggy-back off, you could be nearer to adoption than by getting more XMPP servers deployed.

      1. disgustedoftunbridgewells Silver badge

        Re: Why IMAP and not XNMP?

        XNMP - a mix of XMPP and SNMP ( yes, I meant XMPP )

        1. bobajob12 Bronze badge
          Coat

          Re: Why IMAP and not XNMP?

          You can chat to your buddies, but only in ASN.1

          1. . 3

            Re: Why IMAP and not XNMP?

            There's already enough XMPP capacity in existence for billions of users (cf. ejabberd). However, there is no centrally stored message history so it's no Facebook replacement.

            Why stop at Faceache? An open protocol Ebay / Amazon must be a good decade overdue.

            1. Charles 9 Silver badge

              Re: Why IMAP and not XNMP?

              What eBay and Amazon provide cannot be decentralized. They succeed BY being middlemen: a moderator/facilitator in this case. They reduce the "caveat emptor" risks. Plus Amazon is itself a retailer: no middlemen in those cases.

              1. Anonymous Coward
                Anonymous Coward

                Re: Why IMAP and not XNMP?

                "What eBay and Amazon provide cannot be decentralized. They succeed BY being middlemen: a moderator/facilitator in this case. They reduce the "caveat emptor" risks. Plus Amazon is itself a retailer: no middlemen in those cases."

                I used Google to find stuff on Amazon, even though Amazon has a search function.

                Break down the functions. Product catalog management, reputation management, sales process mediation, escrow, insurance, payment handling, and other stuff. These functions can be provided a la carte. What Amazon and eBay do is bundle it for our convenience, which admittedly is a value provided as well. We like convenience, sometimes to our own detriment. It needn't be centralized, or i.e. monopolized. As protocols mature and interoperability advances, we can have a different world. I'm not betting on it, but I don't give up hope.

                1. JMcL

                  Re: Why IMAP and not XNMP?

                  "Break down the functions. Product catalog management, reputation management, sales process mediation, escrow, insurance, payment handling, and other stuff. These functions can be provided a la carte. What Amazon and eBay do is bundle it for our convenience, which admittedly is a value provided as well."

                  Amazon in fairness do add logistics into the mix, marketplace excluded, and we do all like the large discounts which that sort of scale affords.

          2. Anonymous Coward
            Anonymous Coward

            Re: Why IMAP and not XNMP?

            You can chat to your buddies, but only in ASN.1

            Given that ASN.1 can mean XML or JSON these days (the ITU standard is still updated). ASN.1 is quite cool and Web friendly.

            As it happens I'm using this feature of ASN.1 right now. I get JSON objects in through a Web API, and those get deserialised, validated, reserialised as XML and then uPER as they move throughout the system where bandwidth becomes scarce (uPER is handy for that). All the code to do this serialisation to multiple wire formats is generated automatically from a single ASN.1 schema, which is kinda cool.

            For my next trick I intend running the C code generated by one of the open source ASN.1 tools as a Web assembly. If that's not enough to cause widespread disgust throughout the developer community, I don't know what is...

            1. bobajob12 Bronze badge

              Re: Why IMAP and not XNMP?

              That is very cool...didn't know that. I'm still recovering from programming ASN.1 for OSI protocol stacks in the early 90s. I had no idea they had kept up with the cool kids.

              1. Michael Wojcik Silver badge

                Re: Why IMAP and not XNMP?

                I had no idea they had kept up with the cool kids.

                While DER, and even more so BER, are abominations and largely responsible for the myriad problems of ASN.1 implementations, ASN.1 itself is still overengineered and excessive. It's the very antithesis of "cool", regardless of encoding.

                OIDs in ASN.1 structures are useful, inasmuch as they add a typing mechanism, but even they are poorly designed. Hierarchical namespace: great. Represented with integers assigned by a numbering authority rather than human-readable text: dumb.

                Aside from OIDs I'm having a hard time thinking of anything valuable ASN.1 brought to the table. It's not like it invented the idea of describing structured types using a CFG. We've had Backus-Naur Form since 1960, and as Wikipedia points out, the general idea is around 2500 years old.

      2. Anonymous Coward
        Anonymous Coward

        Re: Why IMAP and not XNMP?

        > as there is already a massive number of IMAP servers already deployed anyway, that this extension can piggy-back off, you could be nearer to adoption than by getting more XMPP servers deployed.

        Group messaging over IMAP is almost a completely new protocol (even *sending* of E-mail isn't normally done over IMAP). So why build completely new protocol, and servers, and clients, when there are existing XMPP servers and clients?

  2. Anne-Lise Pasch

    another bloody silo that nobody needs

    We're good at adding silos. I'm lumbered with Slack, Skype (and For Business) and MS Teams. Id really like it if they'd amalgamate underneath.

    1. bombastic bob Silver badge
      Unhappy

      Re: another bloody silo that nobody needs

      extending IMAP probably isn't the right way to do it, especially if existing private e-mail systems NOT running Dovecot suddenly become "incompatible" because of it.

      Any kind of central "real world identity" system is RIPE for abuse. Period.

      Spammers and scammers WANT to be able to connect a real-world identity with an online one, especially if you use multiple e-mail addresses, IRC handles, gamer IDs, or "any other presence". If they can't track you CENTRALLY, in order to monitor you for advertising (or other) purposes, they'll do it via the "real world identity" thing.

      Don't forget, Micro-shaft NOW has that "Micro-shaft Logon" which was first excreted in the early part of the 21st century as part of their "Dot Not Initiative" - aka "Micro-shaft Passport", which was a TOTAL flop even when compared to the whole "dot bomb" thing.

      maybe its time to divorce from the Faecebook way of doing things, and go back to something that doesn't track you or spam you with ads, like USENET and IRC.

      1. John G Imrie Silver badge

        Any kind of central "real world identity" system is RIPE for abuse. Period.

        I may have misunderstood, but wasn't the ID bit to piggyback of DNS, which is defiantly not a centralised system.

        1. Anonymous Coward
          Anonymous Coward

          Re: Any kind of central "real world identity" system is RIPE for abuse. Period.

          DNS is a hierarchical decentralized system. The worst of both worlds.

          Federated social networks are yet another compromise between central and decentralized. It's not a new idea, it's been tried and failed before (eg. XMPP). What's the incentive for big ISPs to maintain a complex decentralized infrastructure when their users are happy enough with their simple central silos? There is none.

          1. Anonymous Coward
            Anonymous Coward

            Re: Any kind of central "real world identity" system is RIPE for abuse. Period.

            "What's the incentive for big ISPs to maintain a complex decentralized infrastructure when their users are happy enough with their simple central silos? There is none."

            The answer is dead easy. Existence. As in existential class actions against such ISPs (in a broad sense), which are not only foreseeable but foreseen by certain devs (that also happen to be trial lawyers) quietly building the technical infrastructure to create and support properly pleaded "lawyer-less decentralized class actions".

            If you want to predict the future, it's hard. Until and unless you actively create that future. You don't even need funding or community "support" until you choose to deploy, that just slows everything down. It's so easy to do that Zuck did it with his dystopian present, so why not anyone?

          2. Door Handle
            Linux

            Re: Any kind of central "real world identity" system is RIPE for abuse. Period.

            You don't necessarily need any new infrstructure or any other form of co-operation from ISPs. Check out Delta Chat: https://delta.chat/en/

            This is an open source Android chat app built on top of the existing SMTP and IMAP protocols. Pretty compatible with E-mail clients and has end-to-end encryption, it can use your existing mail accounts / servers. Of course gmail works as well. You don't need anything apart from the Android app to start using it.

            1. Michael Wojcik Silver badge

              Re: Any kind of central "real world identity" system is RIPE for abuse. Period.

              You don't need anything apart from the Android app to start using it.

              Presumably you need at least one interlocutor.

              And therein, I think, lies the rub. I haven't produced any content on Facebook since shortly after I first signed up several years ago: not a post, not a status update, not a "like". I do, however, read it for at least a few minutes most days, because my extended family and friends post heavily, and I respect them enough to try to pay a bit of attention.

              None of them are using Delta Chat. None of them are likely to ever start using some new decentralized IMAP-based social media system. True, several use Twitter or Instagram or god knows what else; I ignore all of that because the Facebook skim is all I'm willing to invest in this, respect or no. But they started using those other non-Facebook social media services because there was pressure to, and that pressure was largely generated by marketing campaigns.1

              I've yet to see an open protocol with a real marketing campaign.

              To the vast majority of ordinary users, a new social media service is just what they see in the client app. And if that doesn't do something that they see as novel and valuable, they won't adopt it.

              1Typically "stealth" campaigns of the sort described by Holiday.

      2. Anonymous Coward
        Anonymous Coward

        Re: another bloody silo that nobody needs

        "maybe its time to divorce from the Faecebook way of doing things, and go back to something that doesn't track you or spam you with ads, like USENET and IRC."

        What's keeping you here? Usenet and IRC aren't dead yet.

      3. daealc
        Thumb Up

        Re: another bloody silo that nobody needs

        IRC will never die :)

      4. Anonymous Coward
        Anonymous Coward

        Re: another bloody silo that nobody needs

        extending IMAP probably isn't the right way to do it, especially if existing private e-mail systems NOT running Dovecot suddenly become "incompatible" because of it.

        Agreed

        Any kind of central "real world identity" system is RIPE for abuse. Period.

        What's being discussed is a non-centralised, federated identity system.

        Spammers and scammers WANT to be able to connect a real-world identity with an online one

        Why not actually read the id4me overview? You'll see that identity providers can provide a different unique ID to each online system ("relying party") which authenticates against it. Two cooperating services don't need to know that you are the same person.

        "Identity authorities can decide whether to provide the same identity handle to all the relying parties having access to a specific identity, or whether to provide a different identity handle to each relying party. "

        especially if you use multiple e-mail addresses, IRC handles, gamer IDs, or "any other presence". If they can't track you CENTRALLY, in order to monitor you for advertising (or other) purposes, they'll do it via the "real world identity" thing.

        It's not necessary to link an online identity to a real-world identity. In fact, id4me does not have any existing way to do so. Any personally identifying information you care to share via id4me is unverified and can be as fake as you like. Or you can create an identity with no information at all.

        Don't forget, Micro-shaft NOW has that "Micro-shaft Logon" which was first excreted in the early part of the 21st century as part of their "Dot Not Initiative" - aka "Micro-shaft Passport", which was a TOTAL flop even when compared to the whole "dot bomb" thing.

        Clearly you don't use Office365 or OneDrive; these authenticate off a Microsoft account. Wouldn't it be a bit mad to have a separate username and password for each Microsoft service?

        1. Charles 9 Silver badge

          Re: another bloody silo that nobody needs

          The point is that there is a financial incentive to de-anonymize the Internet's netizens. Meaning there will ALWAYS be someone out there trying to connect the dots, no matter how much effort you put into scattering chaff, much like a letter in the mail, the required information (addresses) alone are enough to start painting a picture. Frankly, if you want to stay anonymous, don't communicate. EVER. In fact, don't even step out into public streets (as you never know where a camera is hidden).

    2. small and stupid

      Re: another bloody silo that nobody needs

      No silos are great.

      You can have a separate identity in each and if it goes titsup at least the consequences are limited.

  3. MacroRodent Silver badge

    The hardware underneath

    The big thing Facebook, Google etc bring to the table is a huge network of maintained servers. We can dream up alternative social networks and software all we want, but it has to run somewhere, And that costs real money, and cannot be done as a hobby (at least not for long).

    1. DougS Silver badge

      Re: The hardware underneath

      A lot of people have unused bandwidth, CPU cycles and gigabytes of storage. Theoretically you could build some sort of massively redundant distributed system, though it would probably be easier to rent cloud resources from Microsoft or Amazon and ask for donations to fund a spying-free social networking alternative.

      1. Charles 9 Silver badge

        Re: The hardware underneath

        I don't think you CAN make it spying-free unless you roll it completely on your own. Otherwise, SOMEONE in the loop's gonna peek.

        1. veti Silver badge

          Re: The hardware underneath

          And by "roll it completely on your own", you mean "running your own software, on your own operating system, your own hardware, with chips designed, fabricated and soldered into place by you personally."

          Anything less is just posing.

          Ideally you should also be laying the fibre and installing connectors too, but that would be unreasonable. So long as there's no silicon within your server centre that's provided or manufactured by anyone else, that's probably good enough.

          Assuming you're perfect, of course.

          1. Jack of Shadows Silver badge

            Re: The hardware underneath

            The roll-your-own movement is still alive and kicking as a direct result of governmental overreach. They are a solid part of the people I hang out with on the weekends. There already exist CPU's which have been verified as being per specifications, without additions. Operating system kernels formally verified. Each program as well. And so forth as we work up the stack. Note: Everything outside the device is considered insecure. There's some special theories about securing communications beyond the end point, considered an essential feature in the face of a determined attacker.

            There's also a solid group trying to figure out ways to operate in an environment where you can't trust any device or its components. That group is operating at a level that loses me. I really need to upgrade my maths to correspond to what they are practicing.

            Interesting problems and it's not like I have much else to do these days.

            1. Charles 9 Silver badge

              Re: The hardware underneath

              I frankly don't see how that problem can be solved. It's like the First Contact problem (Alice and Bob trying to prove themselves to each other without anything, even a Trent, in common between them). Basically, if you can't trust your environment, you can't trust your work in it, period. Meaning all the bad guys need to win is to install a perpetual paranoia.

      2. Jack of Shadows Silver badge

        Re: The hardware underneath

        The major problem with trying to utilize unused bandwidth, CPU cycles, storage, ad nauseum are the cable, telecomms, and other ISP's themselves. Try running a server and see how far you get. As a matter of fact, I would really like a business internet connection via Comcast Business. They aren't even willing to try even though we have their all up service here. That's completely ignoring the asymmetric nature of normal services. That's the other reason I wanted a business connection here, it's symmetric. 1 TB on each of storage on several services doesn't do much good if you can't upload fast enough to do any good. Oh, and that's ignoring all sorts of free consumer accounts.

        If Dovecoat goes anywhere, I'm willing to find ways with or around our ISP. Hell, just host it on somebody else's server (Amazon, ....).

        1. DougS Silver badge

          Re: The hardware underneath

          The asymmetry is going away thanks to some breakthroughs in self interference cancellation earlier this decade. DOCSIS 3.1 has a full duplex mode that is standardized and will be deployed later this year at up to 5 Gbps symmetric by Comcast or Spectrum (can't remember) Terabit DSL is also symmetric - though it may never see the light of day in the US, since telcos have been unloading their copper to companies who are letting it rot while milking grannies who cling to their landlines.

      3. Michael Wojcik Silver badge

        Re: The hardware underneath

        Theoretically you could build some sort of massively redundant distributed system, though it would probably be easier to

        use Usenet.

        Hey, look, it's NNTP. An open protocol for a redundant distributed social-media system.

        I know. It'll never catch on again.

    2. Aitor 1 Silver badge

      Re: The hardware underneath

      I have had in the past forums, etc.

      You mostly have some extra power there.. and you would only need more power if you have more users. Then you should get more money in (donations, subscriptions, selling their personal data, ads, whatever).

    3. Michael Wojcik Silver badge

      Re: The hardware underneath

      The big thing Facebook, Google etc bring to the table is a huge network of maintained servers

      That's only part of it. Orlowski's "trivially simple" comment is, in fact, quite wrong. More is different.

      When you scale up a system to the kind of transaction rates that Facebook handles, you need quite a lot of non-trivial software. It's not just a question of throwing hardware at it. Even a very low failure rate1 turns into quite a lot of failures when multiplied by that load.

      Take a look at Realtime Data Processing at Facebook, say, or the SVE paper, or TAO.

      The big social-media players do quite a lot of software R&D. It is not trivial, nor simple.

      And yes, Dovecot may handle, in aggregate, the email of a couple of billion people. Email workloads are orders of magnitude smaller than social-media workloads.

      I'm perfectly happy to see people extending IMAP (though I've never been a fan of IMAP, particularly) or other open protocols. I've spent much of my professional career working with both open and proprietary data-comm protocols, and even the gnarliest open ones (IIOP, say, or if you want an IETF-blessed example, Telnet) are generally much nicer than the proprietary alternatives (ah, SNA, so many years you have claimed). But minimizing the technical challenges helps no one.

      1And the high tolerance for failure in social-media applications, which really don't care about consistency and reliability.

  4. Anonymous Coward
    Anonymous Coward

    It's a good idea but how do you monetise it? Do email providers start charging money or serving ads? How long till we're back at the same problem?

    1. Julz

      I already pay for my email service.

      1. Doctor Syntax Silver badge

        "I already pay for my email service."

        So do I, partially so I can hand out and delete unique email addresses but there are millions who won't.

        1. David 164 Bronze badge

          More like billions and billions of users is what the likes of facebook and google and yahoo operate at.

          The number of individual who actually pay for their own email service is in the low millions.

  5. highdiver_2000

    Why not XMPP? Unless you want to de throne Whatsapp, then good luck.

    Why Facebook messaging?

    The key is the content, good, bad, inane, horrible content. These drive the user interactions and messages. Be it individual posts or Group pages. But the article says it is not about content, so WTF?

  6. Doctor Syntax Silver badge

    The problem with this scheme is that so many users will go for the cheapest ISP and then "free" services such as FB which cash in by selling the users. If a bunch of ISPs stepped up and offered services with similar levels of traffic to FB etc their costs would go up and hence their prices unless they also sold their users. They wouldn't attract custom away from cheap ISP and "free" social network combination.

    1. teknopaul Bronze badge

      Hotmail, Gmail and Arsebook can all play ball and give away free to users that want their chats scanned, but those of us that don't and have a different email address can still chat.

      1. Doctor Syntax Silver badge

        "but those of us that don't and have a different email address can still chat"

        And continue to have no measurable difference on FB's viability.

  7. Anonymous Coward
    Anonymous Coward

    I'm looking for a Scottish croft that has no internet or mobile phone reception to save me from this anti-social madness.

    1. d3vy Silver badge
      Joke

      From experience you dont need to be too remote... Pretty much anywhere north of Glasgow should meet your requirements :)

    2. Stuart Halliday

      Try Oban. Pretty much zero GSM.

      1. Anonymous Coward
        Anonymous Coward

        >Try Oban. Pretty much zero GSM.

        Ah wonderful Oban, you go there for the scuba diving and the great fish and chips by the harbour.

      2. d3vy Silver badge

        "Try Oban. Pretty much zero GSM."

        Pah, you dont know you're born... I spent the late 90s on North Uist!

      3. JMcL

        "Try Oban. Pretty much zero GSM."

        Who needs GSM when you've got whisky

    3. Snar

      From our very own El Reg!

      This place was up for sale a few years ago - I wish I'd had the cash to buy it.....

      https://www.theregister.co.uk/2010/07/09/scottish_bunker/

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019