7 ms·

The death of the protocol. People no longer release protocols, they release products. The open net, the open web has truly died. I wonder sometimes what we're

The death of the protocol. People no longer release protocols, they release products. The open net, the open web has truly died. I wonder sometimes what we're missing because the conditions that let the web grow out of the internet have disappeared. What next leap are we missing?

etaty 7y ago
Not products, but platforms.
- dexterdog 7y ago
  And in the past you could not be de-protocol-ed
  - bonesss 7y ago
    Not by mobs of angry students, but protocol managing entities like Microsoft, Novell, IBM, and others sure could make your interpretation of the protocol invalid. Particularly if you had some competing software out in the wild...
    - dexterdog 7y ago
      You can code around a protocol change. You can't do that with a de-platforming.
jasode 7y ago
>The death of the protocol. People no longer release protocols, they release products. The open net, the open web has truly died. People ascribe too much credit to a protocol. The real issue is the funding model to support spending money on that protocol. By focusing the lens on the funding, the web wasn't as "open" as people think. The so-called "open" web was a government project which wasn't fully open to the public until about ~1995. Before then, you had to work on government-funded projects (ARPANET & NSFNET). The "open" web was actually quite closed off to people that didn't belong to an institution to give them an ".edu" or ".mil" address. >I wonder sometimes what we're missing because the conditions that let the web grow out of the internet have disappeared. The "conditions" were government sponsorship to pay for protocols. - Vint Cerf developed the TCPIP protocol -- but he was funded by DARPA (USA government). Mr Cerf didn't have to go to venture capital firms on Sand Hill Rd and get funding to pay the salaries of his team to build a proof-of-concept or "MVP". He didn't have to worry about not making payroll if his TCPIP protocol project failed. - Tim Berners-Lee developed the http protocol and also HTML language -- but he was funded by CERN...which is one step removed from being funded by a consortium of governments[0]. Therefore, TBL didn't have to wear a "salesman" hat and pound the pavement begging enterprise customers to buy his "http protocol stack" so he could buy his expensive NeXT workstation[1] to help build and test the protocol. Those are prominent examples but it doesn't mean people can't create new protocols on their without government help. The SSL protocol was developed by Netscape -- a commercial business. I believe the the git protocol was developed by Linus without a government grant. (Also note that http v1 was funded by CERN (government sponsored) but http/2 & http/3 appears to be mostly funded by Google (non-government sponsored).) But notice what happens with "open protocols" like SSL and git: they eventually run into a wall where somebody has to spend money on it. The SSL protocol is free of royalties but the Certificate Authorities issuing certificates are not free of charge. The git protocol is free but enterprise accounts for Github are not. The ActivityPub protocol is also free but Mastodon servers are run by altruistic admins spending their own money or requesting Patreon donations. Somewhere, money has to be spent and that ultimately constrains what the "free web" will be. Releasing a "free to use protocol" is not enough. Therefore, it doesn't matter if the old IRC protocol is free of charge. People still have to spend money on auxiliary features outside the scope of the protocol. Slack spent money on those extra features (e.g. searchable history, etc) that nobody spent on IRC. [0] https://en.wikipedia.org/wiki/CERN#Member_states_and_budget https://en.wikipedia.org/wiki/CERN#Member_states_and_budget [1] https://en.wikipedia.org/wiki/Tim_Berners-Lee#/media/File:First_Web_Server.jpg https://en.wikipedia.org/wiki/Tim_Berners-Lee#/media/File:Fi...
- roncanepa 7y ago
  I think this is a very good explanation of things. Reminds me of the "free as in puppies" angle.
zimbatm 7y ago
Except people still release protocols. Just look at https://matrix.org/ https://matrix.org/ IRC is an early protocol and wasn't designed for extensibility purposes. A new chat protocol just needs to replace it.
- SamuelAdams 7y ago
  IRC was created in 1988. There have been several chat protocols created since then [1]. A few notable protocols that tick a lot of "modern" boxes: - Jami - Matrix - Signal Protocol - XMPP The better question: why aren't new and existing technology companies embracing these protocols and making them ubiquitous? Personally I think it's because companies have a vested interest in creating walled gardens. The more users that are forced onto a platform, the more money they make. It's against their business interest to create a product that works with their competitors. [1]: https://en.wikipedia.org/wiki/Comparison_of_instant_messaging_protocols https://en.wikipedia.org/wiki/Comparison_of_instant_messagin...
  - jasode 7y ago
    >It's against their business interest to create a product that works with their competitors. There's more to it than a simplistic "business interest" to explain it. For example, the Signal protocol is free to use. No payments for patent royalties or software licenses required. Moxie Marlinspike's company Signal LLC (Open WhisperSystems) is a non-profit entity. But notice that an organization that isn't pressured into making a profit and that's using an underlying free protocol -- does not want to federate with LibreSignal[0]. This an example of where the "openness of a protocol" runs into a hard reality of "closed wall of implementation". The protocols themselves can cost $0 but implementations (e.g. cpu+disk+bandwidth+staff) cost more than $0 which leads to complaints of it being "not really open". If people do not examine the economic forces surrounding any protocol, they will always be perplexed why the internet is not as free & open as they think it should be. Even non-profit entities are not immune from economic forces. [0] https://github.com/LibreSignal/LibreSignal/issues/37#issuecomment-217211165 https://github.com/LibreSignal/LibreSignal/issues/37#issueco...
  - achingtooth 7y ago
granshaw 7y ago
There’s no money to be made in creating and building on open protocols. Without money to be made there’s no funding. Without funding there’s no way to create great experiences. Without great experiences nobody will use the thing. Pretty straightforward, really
- pjmorris 7y ago
  And yet, as a counterexample, HTML and HTTP persist. Maybe things aren't as straightforward as they seem.
  - lima 7y ago
    With QUIC, even TCP is getting replaced because of the need to innovate despite protocol ossification. It's almost impossible to improve TCP at this point due to non-compliant endpoints and middleboxes. It's a similar "lowest common denominator" issue with IRC, XMPP, mail and many others. Yes, XMPP can - in theory - do almost everything Slack does, but there's dozens of clients supporting a different subset of the protocol. Of course nobody wants to use it.
  - vntok 7y ago
    I don't get how your point is a counter example to the parent's argument? I think there is a lot of money to be made in building on HTML and HTTP. For one, a faster HTTP protocol directly means more cost efficiency in hosting and streaming content, better customer conversions, etc.
    - pjmorris 7y ago
      IMO, HTTP and HTML are counter-examples to the statement 'There’s no money to be made in creating and building on open protocols.' because they're open protocols that have attracted a great deal of money, and plenty of people creating and building on them.
- hackerpacker 7y ago
  On the other hand, I don't think anyone accused slack, with multi-megabyte webpages and even worse mobile app, of being a great experience. I think I used it once, and said no more.
wildduck 7y ago
IMHO there need to be balance between open source and revenue, protocol and products. Just because everyone is going one way, doesn't mean it is correct. I have a friend from banking use to make fun of Richard Stallman 15 years ago. However, looking at it not what he said is quite prophetic. What I’m most concern about is cynicism.
eloisant 7y ago
We never got an instant messaging that works like email, for one.
- drzaiusx11 7y ago
  matrix is exactly this, at least from an client/server interop standpoint. email is the first massively successful federated application. matrix just needs an easy to use client/webclient. Riot.im is close, but needs some UX love.
  - wildduck 7y ago
    Agree. However the fact that E2E with perfect forwarding secrecy is not guaranteed is going to be an issue.
    - xf86alsa 7y ago
      Can you explain? My understanding is that megolm (what Matrix uses) ratchets the session forward every hour or every 100 msgs, whichever first.
CondensedBrain 7y ago
ActivityPub shows a lot of promise.
- wildduck 7y ago
  Also look at dat:// take a look. Beaker browser.
  - CondensedBrain 7y ago
    ActivityPub (AP) has the advantage of running on top of HTTP. You don't have to convince people to install a new browser. All the tools that communicate over AP run on familiar technologies. The hardest part is convincing people to make an account, so most growth is from people following friends--the key incentive--to Mastodon instances. It's not a big problem in practice. Monoculture concerns aside. AP doesn't have to run on HTTP, but any alternative has the same problem as anything else that needs its own client. Even browser support for FTP is on the way out, and it used to be standard.
wUabkSG6L5Bfa5 7y ago
Man, it's crazy how much was riding on Slack staying private.
Kinnard 7y ago
I don't know about that. There's a flurry of open protocols being built on Blockchain and maybe there's a golden age ahead . . . > "Historically many key protocols, such as TCP/IP and HTTP, have come from researchers. Subsequent iterations of these protocols were often handled by nonprofit organization that tried to wrangle with more or less success the various commercial interests that sprang up around this protocols (the companies that were making and selling software and hardware based on them). The more money was involved the harder this became. > "Now, however, we have a new way of providing incentives for the creation of protocols and for governing their evolution. I am talking about cryptographic tokens . . . > "I can’t emphasize enough how radical a change this is to the past. Historically the only way to make money from a protocol was to create software that implemented it and then try to sell this software (or more recently to host it). Since the creation of this software (e.g. web server/browser) is a separate act many of the researchers who have created some of the most successful protocols in use today have had little direct financial gain. With tokens, however, the creators of a protocol can “monetize” it directly and will in fact benefit more as others build businesses on top of that protocol. Given this new incentive I expect a lot of resources to be devoted to protocol innovation." ~ https://continuations.com/post/148098927445/crypto-tokens-and-the-coming-age-of-protocol https://continuations.com/post/148098927445/crypto-tokens-an...

etaty 7y ago

Not products, but platforms.

dexterdog 7y ago
And in the past you could not be de-protocol-ed
- bonesss 7y ago
  Not by mobs of angry students, but protocol managing entities like Microsoft, Novell, IBM, and others sure could make your interpretation of the protocol invalid. Particularly if you had some competing software out in the wild...
  - dexterdog 7y ago
    You can code around a protocol change. You can't do that with a de-platforming.

jasode 7y ago

>The death of the protocol. People no longer release protocols, they release products. The open net, the open web has truly died. People ascribe too much credit to a protocol. The real issue is the funding model to support spending money on that protocol. By focusing the lens on the funding, the web wasn't as "open" as people think. The so-called "open" web was a government project which wasn't fully open to the public until about ~1995. Before then, you had to work on government-funded projects (ARPANET & NSFNET). The "open" web was actually quite closed off to people that didn't belong to an institution to give them an ".edu" or ".mil" address. >I wonder sometimes what we're missing because the conditions that let the web grow out of the internet have disappeared. The "conditions" were government sponsorship to pay for protocols. - Vint Cerf developed the TCPIP protocol -- but he was funded by DARPA (USA government). Mr Cerf didn't have to go to venture capital firms on Sand Hill Rd and get funding to pay the salaries of his team to build a proof-of-concept or "MVP". He didn't have to worry about not making payroll if his TCPIP protocol project failed. - Tim Berners-Lee developed the http protocol and also HTML language -- but he was funded by CERN...which is one step removed from being funded by a consortium of governments[0]. Therefore, TBL didn't have to wear a "salesman" hat and pound the pavement begging enterprise customers to buy his "http protocol stack" so he could buy his expensive NeXT workstation[1] to help build and test the protocol. Those are prominent examples but it doesn't mean people can't create new protocols on their without government help. The SSL protocol was developed by Netscape -- a commercial business. I believe the the git protocol was developed by Linus without a government grant. (Also note that http v1 was funded by CERN (government sponsored) but http/2 & http/3 appears to be mostly funded by Google (non-government sponsored).) But notice what happens with "open protocols" like SSL and git: they eventually run into a wall where somebody has to spend money on it. The SSL protocol is free of royalties but the Certificate Authorities issuing certificates are not free of charge. The git protocol is free but enterprise accounts for Github are not. The ActivityPub protocol is also free but Mastodon servers are run by altruistic admins spending their own money or requesting Patreon donations. Somewhere, money has to be spent and that ultimately constrains what the "free web" will be. Releasing a "free to use protocol" is not enough. Therefore, it doesn't matter if the old IRC protocol is free of charge. People still have to spend money on auxiliary features outside the scope of the protocol. Slack spent money on those extra features (e.g. searchable history, etc) that nobody spent on IRC. [0] https://en.wikipedia.org/wiki/CERN#Member_states_and_budget https://en.wikipedia.org/wiki/CERN#Member_states_and_budget [1] https://en.wikipedia.org/wiki/Tim_Berners-Lee#/media/File:First_Web_Server.jpg https://en.wikipedia.org/wiki/Tim_Berners-Lee#/media/File:Fi...

roncanepa 7y ago
I think this is a very good explanation of things. Reminds me of the "free as in puppies" angle.

zimbatm 7y ago

Except people still release protocols. Just look at https://matrix.org/ https://matrix.org/ IRC is an early protocol and wasn't designed for extensibility purposes. A new chat protocol just needs to replace it.

SamuelAdams 7y ago
IRC was created in 1988. There have been several chat protocols created since then [1]. A few notable protocols that tick a lot of "modern" boxes: - Jami - Matrix - Signal Protocol - XMPP The better question: why aren't new and existing technology companies embracing these protocols and making them ubiquitous? Personally I think it's because companies have a vested interest in creating walled gardens. The more users that are forced onto a platform, the more money they make. It's against their business interest to create a product that works with their competitors. [1]: https://en.wikipedia.org/wiki/Comparison_of_instant_messaging_protocols https://en.wikipedia.org/wiki/Comparison_of_instant_messagin...
- jasode 7y ago
  >It's against their business interest to create a product that works with their competitors. There's more to it than a simplistic "business interest" to explain it. For example, the Signal protocol is free to use. No payments for patent royalties or software licenses required. Moxie Marlinspike's company Signal LLC (Open WhisperSystems) is a non-profit entity. But notice that an organization that isn't pressured into making a profit and that's using an underlying free protocol -- does not want to federate with LibreSignal[0]. This an example of where the "openness of a protocol" runs into a hard reality of "closed wall of implementation". The protocols themselves can cost $0 but implementations (e.g. cpu+disk+bandwidth+staff) cost more than $0 which leads to complaints of it being "not really open". If people do not examine the economic forces surrounding any protocol, they will always be perplexed why the internet is not as free & open as they think it should be. Even non-profit entities are not immune from economic forces. [0] https://github.com/LibreSignal/LibreSignal/issues/37#issuecomment-217211165 https://github.com/LibreSignal/LibreSignal/issues/37#issueco...
- achingtooth 7y ago

granshaw 7y ago

There’s no money to be made in creating and building on open protocols. Without money to be made there’s no funding. Without funding there’s no way to create great experiences. Without great experiences nobody will use the thing. Pretty straightforward, really

pjmorris 7y ago
And yet, as a counterexample, HTML and HTTP persist. Maybe things aren't as straightforward as they seem.
- lima 7y ago
  With QUIC, even TCP is getting replaced because of the need to innovate despite protocol ossification. It's almost impossible to improve TCP at this point due to non-compliant endpoints and middleboxes. It's a similar "lowest common denominator" issue with IRC, XMPP, mail and many others. Yes, XMPP can - in theory - do almost everything Slack does, but there's dozens of clients supporting a different subset of the protocol. Of course nobody wants to use it.
- vntok 7y ago
  I don't get how your point is a counter example to the parent's argument? I think there is a lot of money to be made in building on HTML and HTTP. For one, a faster HTTP protocol directly means more cost efficiency in hosting and streaming content, better customer conversions, etc.
  - pjmorris 7y ago
    IMO, HTTP and HTML are counter-examples to the statement 'There’s no money to be made in creating and building on open protocols.' because they're open protocols that have attracted a great deal of money, and plenty of people creating and building on them.
hackerpacker 7y ago
On the other hand, I don't think anyone accused slack, with multi-megabyte webpages and even worse mobile app, of being a great experience. I think I used it once, and said no more.

wildduck 7y ago

IMHO there need to be balance between open source and revenue, protocol and products. Just because everyone is going one way, doesn't mean it is correct. I have a friend from banking use to make fun of Richard Stallman 15 years ago. However, looking at it not what he said is quite prophetic. What I’m most concern about is cynicism.

eloisant 7y ago

We never got an instant messaging that works like email, for one.

drzaiusx11 7y ago
matrix is exactly this, at least from an client/server interop standpoint. email is the first massively successful federated application. matrix just needs an easy to use client/webclient. Riot.im is close, but needs some UX love.
- wildduck 7y ago
  Agree. However the fact that E2E with perfect forwarding secrecy is not guaranteed is going to be an issue.
  - xf86alsa 7y ago
    Can you explain? My understanding is that megolm (what Matrix uses) ratchets the session forward every hour or every 100 msgs, whichever first.

CondensedBrain 7y ago

ActivityPub shows a lot of promise.

wildduck 7y ago
Also look at dat:// take a look. Beaker browser.
- CondensedBrain 7y ago
  ActivityPub (AP) has the advantage of running on top of HTTP. You don't have to convince people to install a new browser. All the tools that communicate over AP run on familiar technologies. The hardest part is convincing people to make an account, so most growth is from people following friends--the key incentive--to Mastodon instances. It's not a big problem in practice. Monoculture concerns aside. AP doesn't have to run on HTTP, but any alternative has the same problem as anything else that needs its own client. Even browser support for FTP is on the way out, and it used to be standard.

wUabkSG6L5Bfa5 7y ago

Man, it's crazy how much was riding on Slack staying private.

Kinnard 7y ago

I don't know about that. There's a flurry of open protocols being built on Blockchain and maybe there's a golden age ahead . . . > "Historically many key protocols, such as TCP/IP and HTTP, have come from researchers. Subsequent iterations of these protocols were often handled by nonprofit organization that tried to wrangle with more or less success the various commercial interests that sprang up around this protocols (the companies that were making and selling software and hardware based on them). The more money was involved the harder this became. > "Now, however, we have a new way of providing incentives for the creation of protocols and for governing their evolution. I am talking about cryptographic tokens . . . > "I can’t emphasize enough how radical a change this is to the past. Historically the only way to make money from a protocol was to create software that implemented it and then try to sell this software (or more recently to host it). Since the creation of this software (e.g. web server/browser) is a separate act many of the researchers who have created some of the most successful protocols in use today have had little direct financial gain. With tokens, however, the creators of a protocol can “monetize” it directly and will in fact benefit more as others build businesses on top of that protocol. Given this new incentive I expect a lot of resources to be devoted to protocol innovation." ~ https://continuations.com/post/148098927445/crypto-tokens-and-the-coming-age-of-protocol https://continuations.com/post/148098927445/crypto-tokens-an...