Re: [Uri-review] Is TCP a URI scheme?

Graham Klyne <> Fri, 22 February 2019 11:19 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id E09DD12EB11 for <>; Fri, 22 Feb 2019 03:19:32 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -4.2
X-Spam-Status: No, score=-4.2 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3] autolearn=unavailable autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id jEWIaMaOGNo3 for <>; Fri, 22 Feb 2019 03:19:30 -0800 (PST)
Received: from ( []) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 3BD101277CC for <>; Fri, 22 Feb 2019 03:19:30 -0800 (PST)
Received: from ([]) by with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.89) (envelope-from <>) id 1gx8rX-0004xp-cJ; Fri, 22 Feb 2019 11:19:28 +0000
Received: from ([] by with esmtpsa (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.89) (envelope-from <>) id 1gx8rW-0008yQ-F4; Fri, 22 Feb 2019 11:19:26 +0000
Message-ID: <>
Date: Fri, 22 Feb 2019 11:19:27 +0000
From: Graham Klyne <>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.11; rv:31.0) Gecko/20100101 Thunderbird/31.3.0
MIME-Version: 1.0
To: =?windows-1252?Q?=22Martin_J=2E_D=FCrst=22?= <>, Eric Johnson <>
CC: "" <>
References: <> <> <> <>
In-Reply-To: <>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
X-Oxford-Username: zool0635
X-Oxmail-Spam-Status: score=0.0 tests=none
X-Oxmail-Spam-Level: /
Archived-At: <>
Subject: Re: [Uri-review] Is TCP a URI scheme?
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Proposed URI Schemes <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Fri, 22 Feb 2019 11:19:33 -0000

I have a couple of comments prompted by this thread..

1. I propose that it is quite possible to use a string that conforms to URI 
syntax (explicitly or otherwise) without it being a URI.  So what makes it a 
URI?  Two touchstones that I apply are: (a) "what does it identify", and (b) 
"does it make sense to use it in a context where other URI schemes might equally 
be used" (e.g., as an href in an HTML document, or as a link relation type on an 
HTTP Link: header field).

2. Registration of *provisional* URI schemes is now first-come, first-served, 
without formal review, subject to satisfying some simple administrative 
requirements (cf. [1]).

    Permanent registration is a different matter, and does involve stricter 
requirements and formal review [2], which in turn involves some judgement by the 
reviewer about whether the requirements are satisfied.  As the current reviewer, 
I will typically look for indications of working group approval, similar from 
other recognized bodies like W3C, or "IETF review" (formerly "IETF consensus") 
[3] to inform that judgement.





PS: since [4], I've personally come to further downplay the URI/URL distinction. 
I regard it as more to do with context of use than something inherent to the 
actual string form used, in that *any* identifier can be a locator given the 
right infrastructure (e.g,. DOIs), and any locator can potentially be used as an 


On 22/02/2019 04:54, Martin J. Dürst wrote:
> Hello Eric, others,
> On 2019/02/22 09:49, Eric Johnson wrote:
>> Speaking somewhat on behalf of TIBCO, I'm not sure that the use of a string
>> that begins "tcp:" means that what is contained is necessarily a URL. I did
>> a small amount of very informal questioning of some of my fellow employees,
>> and didn't get the impression that the use of "tcp:" was anything other
>> than a signal for the type of thing that followed the prefix, and not
>> strictly speaking a "resource" locator.
> This overlaps quite a bit with what I'd have written as an answer to the
> original question if I had had some spare time when that comes up.
> One of the core points of URIs is that they unify syntax at a very high
> level (the level of foo:). That makes them very useful for pointing at
> various things, from various places. The original place was the href
> attribute of the <a> element in HTML. There, the main schemes were
> http:, https:, ftp:, mailto:, and so on.
> Other places where there was a need to point to something reused URIs,
> but once in a while, there was a need to point to something that didn't
> yet have an URI scheme. With the use of URIs expanding, such things got
> more and more diverse. That somebody somewhere wanted to point to a
> (potential) tcp connection, and would used tcp: for that, was just a
> matter of time.
> For quite a while, registration of URI schemes was handled rather
> strictly. It's supposed to be easier today, but it's still some work,
> some discussion, and some wait. Also, the documents describing URIs are
> not necessarily an easy read for everyone, because they use general
> terms that can easily be misunderstood.
> Also, please note that we are talking about URIs here, not URLs, so it's
> resource Resource Identifier, not Locator. And "resource" is really
> extremely general, even if for some schemes (e.g. mailto:), some care is
> needed when describing what exactly it stands for.
>   > On Fri, Feb 1, 2019 at 9:18 AM Ted Hardie <> wrote:
>   >
>   >> I don't believe it has ever been registered, even provisionally, but you
>   >> can find example syntax in both IBM and TIBCO documentation of their
>   >> usage.  See, for example:
>   >>
>   >>
>   >>
> Looking at that page, it says "TCP Discovery URL Format". I'm not
> familiar with this "discovery" part, but it may have been better to use
> something like "tcp-disc:" or so as a prefix.
> Regards,   Martin.
> _______________________________________________
> Uri-review mailing list