RE: [iri] #73: Decide on organization-specific schemes

After the discussions at the IETF and also in Happiana, I'm leaning toward removing the distinction between "provisional" and "permanent", and moving toward FCFS for initial registration, with "expert review" to weed out spam and camping.  

We would then recommend that people avoid duplicates even with unregistered schemes by using a naming convention.

Is there any advantage in recommending com.eventbrite.attendee and just eventbrite.attendee ...?

I'm a little more concerned about whether we should allow registration of scheme prefixes ( "web+" from HTML). 
I know I used a prefix myself "xmp." ...

I think using DNS names as URI prefixes might actually be harmful rather than helpful.
I'd rather see people be able to register a unique prefix (e.g., "eventbrite" in this case) since URI schemes need to be longer-lived than DNS names.

If bell labs had had some URI schemes, what would have come of them in the various splits and mergers?

One example doesn't mean it was a good idea.

Larry



-----Original Message-----
From: "Mykyta Yevstifeyev (М. Євстіфеєв)" [mailto:evnikita2@gmail.com] 
Sent: Friday, December 09, 2011 9:43 PM
To: Peter Saint-Andre
Cc: iri issue tracker; chris@lookout.net; duerst@it.aoyama.ac.jp; Larry Masinter; public-iri@w3.org
Subject: Re: [iri] #73: Decide on organization-specific schemes

09.12.2011 20:37, Peter Saint-Andre wrote:
> <hat type='individual'/>
>
> On 11/16/11 7:00 PM, iri issue tracker wrote:
>> #73: Decide on organization-specific schemes
>>
>>
>> Comment (by duerst@…):
>>
>>   Moved from #106 (duplicate):
>>
>>   The reason is that there were no actual usages. But now, there is a use
>>   (see  http://www.iana.org/assignments/uri-schemes/prov/com-eventbrite-

>>   attendee), as reported at  http://www.ietf.org/mail-archive/web/uri-

>>   review/current/msg01548.html.
>>
>>    I guess we should restore this text.
> Seems fine to me.

Well, I really think there is no point in distinguishing between some vendor schemes and other schemes, as it's done with media types.  I think developing and registering the scheme according to usual naming rules should be fine in this case.  So my proposal remains the same.

Mykyta Yevstifeyev

>
> Peter
>

Received on Saturday, 10 December 2011 17:18:18 UTC