W3C home > Mailing lists > Public > www-tag@w3.org > June 2012

Re: Registration of acct: as a URI scheme has been requested

From: Ivan Herman <ivan@w3.org>
Date: Mon, 25 Jun 2012 07:35:40 +0200
Cc: Kingsley Idehen <kidehen@openlinksw.com>, "www-tag@w3.org" <www-tag@w3.org>, "Henry S. Thompson" <ht@inf.ed.ac.uk>, Melvin Carvalho <melvincarvalho@gmail.com>
Message-Id: <37DAB127-B93D-4F0C-98B5-B08FAAB50B8C@w3.org>
To: nathan@webr3.org
(not specifically to you, Nathan, just one of your remark below triggered this...)

I just chime in on a non-technical issue. In my understanding the whole starting point of the acct: URI is the claim that most users, technical or non-technical, know and remember their email address. Ie, using something like acct:ivan@w3.org is cognitively easy for them. Easier than, let us say, a .well-known URI construction. And this claim is based that, these days, almost everybody uses emails in some way, let that be a technical geek, or a lambda user.

However... is this something that we can rely on in the future? Can we really take that as granted forever? I am not sure. In my experience there are more and more people around me (granted, youngsters, my child's age or even much younger) whose primary communication means is Facebook messaging, or personal messages via Twitter. They may have an email address, but they do not really use it; I bet in a few years some of these kids will grow up without a need for an email address, ie, without having one. And, for those young adults, acct:a@b.c will look just as cryptic as .well-known, or a longish foaf URI, or anything similar.

Just a thought...


On Jun 24, 2012, at 24:49 , Nathan wrote:

> Kingsley Idehen wrote:
>> On 6/23/12 3:12 PM, Nathan wrote:
>>> Kingsley Idehen wrote:
>>>> On 6/23/12 9:42 AM, Nathan wrote:
>>>>> So rather than creating an unstable pretty much useless URI for use internally within a specific protocol, why not take advantage of this provision and define the variable {acct} instead, such that you can do:
>>>>> https://gmail.com/.well-known/host-meta?acct=joe@gmail.com
>>>>> That way you tie in with web architecture, don't need a new URI-scheme, and still get to do what's required. 
>>>> In what context is any URI useless? Please remember URI abstraction re. context of my question.
>>>> Again: https://gmail.com/.well-known/host-meta?acct=joe@gmail.com , is a URL, a data access address. Webfinger folks don't want to present: <https://gmail.com/.well-known/host-meta?acct=joe@gmail.com> as a name to its end-users and developers when they use: <acct:joe@gmail.com> .
>>>> In a nutshell, you are implying that Linked Data is only achievable via http: scheme URIs. That simply isn't true. Even worse, you are making your case using host-meta which is all about delivering a generic resolver mechanism for URIs. Basically, decoupling the name/access functionality that's baked into http: URLs.
>>>> Being convenient and cost-effective doesn't make http: scheme URIs the sole option for Linked Data. It just doesn't.
>>> As you know I don't need convinced of the benefits of linked data, but I would like convinced that the acct: scheme is required; so far I've not seen any evidence of this, other existing techs can do the job, and RFC6415 appears to cover the cases where identifiers aren't URIs.
>>> I do accept though that saying a URI is useless is was far to strong, what I meant to say, or imply, was that creating a new scheme when not required may not be the best path to take - which I had thought was the point of this thread, and thus discussed then offered an alternative.
>> The net effect of creating an new URI scheme could be costly in some context (e.g. today's World Wide Web dominated by one type of user agent, the Web Browser). Not necessarily so re.,  other contexts.
>>> WebFinger is valuable, to the web, and the web of linked data, and I'd be keen to see it get to where it needs to be with as little red tape and limitations possible, if acct: can be swapped out for ?acct= without it impeding functionality, and speed up the process, then that's what I'd personal opt for.
>> Here's how I've used acct: (since I first encountered Webfinger) for both WebID and my Profile data in general:
>> 1. http://bit.ly/KtaGwI -- searching on my acct: scheme URI
>> 2. http://bit.ly/MEqals -- full profile
>> 3. http://bit.ly/KFeYG3 -- looking at inverse functional property effects on co-references
>> 4. http://bit.ly/KTWCCx -- ditto but via explicit co-reference via owl:sameAs relationships.
> I commend your use of URIs, however I still can't see what clear utility the proposed 'acct:' scheme brings to the broad internet community, beyond that which is already available.
> What requirement is fulfilled by the 'acct:' scheme that isn't already handled by other schemes?
> What does it enable, or what is gained by it's use?
> What other techs might it make possible?
> If there is a real gap for a scheme which can be used to mint identifiers for the class of things which are user accounts, then shouldn't that need be looked at closely and it be determined whether the scheme specification is open and robust enough to stand the test of time?
> acctURI      =  "acct:" userpart "@" domainpart
> This seems very limited, if acct: is to be rolled out at internet level, then why not open it up more, certainly acct:twitter.com/webr3 may be useful, are fragments useful, how many of these questions have been looked at, and how deeply has this new scheme been considered beyond the very specific use case it's been designed to work for/around?
> I'm not trying to knock or stall anything here, just to nudge towards the notion that it may be either (a) unneeded, or (b) needed and possible to improve/generalise to add even more utility.
> Best,
> Nathan

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Monday, 25 June 2012 05:36:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:45 UTC