tcp: URI scheme

[Erik Wilde-3]

hello.

reading the discussion around the proposed ws: URI scheme made me  
realize that there is no tcp: URI scheme. is that something the world  
has missed so far? if yes, and if there should be one, would we  
consider tunneling it over http://tcp.example.com/ URIs to facilitate  
fallback behavior? i don't have the answers (but of course some  
opinions), but maybe if we consider how a TCP URI scheme would be  
approached (if at all) and why we might find having one useful, we  
might gain some insights into how to best deal with websockets. after  
all, the very name of it is borrowed from the world of TCP.

cheers,

dret.

[Bjoern Hoehrmann]

* Erik Wilde wrote:
>reading the discussion around the proposed ws: URI scheme made me  
>realize that there is no tcp: URI scheme. is that something the world  
>has missed so far?

There are quite a number of applications that support something that
looks like a 'tcp' scheme and at times they go on and call it that.
It's just not standardized or registered.
--
Björn Höhrmann · mailto:bjoern@... · http://bjoern.hoehrmann.de
Am Badedeich 7 · Telefon: [phone] · http://www.bjoernsworld.de
25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/ 

[Erik Wilde-3]

hello bjoern.

thanks for your email.

Bjoern Hoehrmann wrote:
> * Erik Wilde wrote:
>> reading the discussion around the proposed ws: URI scheme made me  
>> realize that there is no tcp: URI scheme. is that something the world  
>> has missed so far?
> There are quite a number of applications that support something that
> looks like a 'tcp' scheme and at times they go on and call it that.
> It's just not standardized or registered.

would you mind giving a couple of examples and/or pointers? if there are
applications using such a "scheme", there actually might be value in
having a common convention of how to identify a TCP "resource". since
this stretches the concept of a "resource" a bit, i am wondering how
people think about such a URI scheme, but since mailto: and tel:
basically do the same (not identifying a "resource" in the sense of a
document, but more an endpoint of a communications protocol that
provides connectivity), it might be useful to be able to talk about TCP
endpoints using a URI such as tcp://www.example.com:80, if you want to
talk about the TCP level of things.

cheers,

erik wilde   tel:[phone] - fax:[phone]
        dret@...  -  http://dret.net/netdret
        UC Berkeley - School of Information (ISchool)

[Bjoern Hoehrmann]

* Erik Wilde wrote:
>> There are quite a number of applications that support something that
>> looks like a 'tcp' scheme and at times they go on and call it that.
>> It's just not standardized or registered.
>
>would you mind giving a couple of examples and/or pointers? if there are
>applications using such a "scheme", there actually might be value in
>having a common convention of how to identify a TCP "resource".

http://www.google.com/codesearch?q=%22tcp%3A%2F%2F%22
--
Björn Höhrmann · mailto:bjoern@... · http://bjoern.hoehrmann.de
Am Badedeich 7 · Telefon: [phone] · http://www.bjoernsworld.de
25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/ 

[noah_mendelsohn]

Question:  if we do have such a scheme, what is the resource that a
tcp:xxxxx URI identifies?  I presume it's an endpoint on the network,
typically an IP address and port, that supports TCP connections?

Noah

--------------------------------------
Noah Mendelsohn
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








Bjoern Hoehrmann <derhoermi@...>
Sent by: uri-request@...
08/18/2009 02:37 PM
 
        To:     Erik Wilde <dret@...>
        cc:     URI <uri@...>, (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Re: tcp: URI scheme


* Erik Wilde wrote:
>> There are quite a number of applications that support something that
>> looks like a 'tcp' scheme and at times they go on and call it that.
>> It's just not standardized or registered.
>
>would you mind giving a couple of examples and/or pointers? if there are
>applications using such a "scheme", there actually might be value in
>having a common convention of how to identify a TCP "resource".

http://www.google.com/codesearch?q=%22tcp%3A%2F%2F%22
--
Björn Höhrmann · mailto:bjoern@... · http://bjoern.hoehrmann.de
Am Badedeich 7 · Telefon: [phone] · http://www.bjoernsworld.de
25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/ 

[Erik Wilde-3]

hello noah.

noah_mendelsohn@... wrote:
> Question:  if we do have such a scheme, what is the resource that a
> tcp:xxxxx URI identifies?  I presume it's an endpoint on the network,
> typically an IP address and port, that supports TCP connections?

that would be my theory and my approach for defining such a scheme,
since there is little else to TCP other than the definition of an
endpoint. to me, that looks very similar to the tel: scheme, which also
just identifies an endpoint of an end-to-end connection service, with no
specification of the applications talking over that connection (other
than the payload type supported by that connection type), or which
application-level data is exchanged.

cheers,

dret.