W3C home > Mailing lists > Public > www-tag@w3.org > October 2011

Re: ISSUE-50 Bilder, Geoff. Identify This! Identifiers and Trust.

From: Martin J. Dürst <duerst@it.aoyama.ac.jp>
Date: Mon, 03 Oct 2011 17:08:40 +0900
Message-ID: <4E896D88.1090506@it.aoyama.ac.jp>
To: Jonathan Rees <jar@creativecommons.org>
CC: www-tag@w3.org
Hello Jonathan,

I don't think you're taking too many liberties below, but stuff like 
this has been too obvious for some of us for a long time, so that's 
probably why I had to ask you about what you found noteworthy (and yes, 
I meant the syntax).

In my view, people who use http://dx.doi.org/... are just (at least 
implicitly) evaluating the total of persistence and resolvability of 
various approaches.

My calculations would look about as follows:

http://dx.doi.org/...:

    resolvability: 100pt

    (it's not exactly perfect, but it can't get better, at least as far
     as the DOI people are concerned)

    persistence:    95pt

    (I take one point for the risk at ICANN, although I think it's
     extremely low because while they have no way and no interest to
     declare anything permanent, they have no interest at all to take
     something existing and working away (as long as the relevant fees
     are paid;
     then I take one point for the risk at .org, just because there is
     some risk there, even if small;
     then I take three points for the risk at doi.org, because they could
     indeed miss out on paying their registration renewal fees (this
     is something that actually has happened in other cases))

urn:doi:

    resolvability: 10pt

    (I'm being generous, the race is over here anyway)

    persistence:   99pt

    (IANA registers these, and IANA is (at least currently) part of
     ICANN, so if you don't trust ICANN, you may not be able to trust
     this one, either. There are some RFCs that very clearly tell IANA
     what to do and what not to do, but that doesn't guarantee anything.)

So it should be clear why http://dx.doi.org/ is used all over the place. 
And the order stays pretty much the same even if you change the weights 
or the scores a bit (e.g. you weight persistence more,...).

Regards,    Martin.

On 2011/09/29 21:58, Jonathan Rees wrote:
> On Wed, Sep 28, 2011 at 8:48 PM, "Martin J. Dürst"
> <duerst@it.aoyama.ac.jp>  wrote:
>> On 2011/09/29 1:29, Jonathan Rees wrote:
>>>
>>> FYI. Note the interesting syntax used for the DOI.  -Jonathan
>>
>> Hello Jonathan - Can you say what you found noteworthy?  Thanks,  Martin.
>
> I hope the noteworthiness of the article relative to the ISSUE-50
> discussion is clear, so I assume you mean the noteworthiness of the
> syntax. The article's landing page has a line that says "doi:
> http://..." which, read naively, says that the DOI of the article is
> http://... (of course we don't really know what the publisher intends
> it to mean). The DOI specification says that technically the DOI is
> only the part starting with "10."  Remember that the old URL/URN
> debate revolves around whether http: URIs can serve as "persistent
> identifiers". My naive reading says that in practice, among journal
> publishers, they do - that some community (other than W3C) is actually
> treating some http: URIs (those in the dx.doi.org domain) as
> persistent identifiers (DOIs are supposed to be persistent). If this
> is true, the practice requires some explanation (would ICANN agree
> with the statement that at least one URI beginning http://dx.doi.org/
> can be considered "persistent"?), but it suggests that the URL/URN
> debate is in some sense over, in some quarters, with URLs having won.
>
> This idea is reinforced by the interesting practice of putting
> http://dx.doi.org/ hyperlinks in PDF files that are intended to
> persist indefinitely (e.g. through the LOCKSS system). This applies to
> nearly every bibliographic citation in nearly every research article
> published by a major publisher. There are so many of these hyperlinks
> that, I speculate, if dx.doi.org URIs were to stop resolving
> correctly, you would suddenly see PDF viewers and/or browser plugins
> and/or network stacks and/or DNS resolvers and/or DNS roots sprout up
> that would "fix" the problem. That is, I speculate, ICANN and PIR have
> in fact already lost "authority" over this part of the URI namespace
> (although the question is, at present, totally academic since the URIs
> *do* resolve correctly, aften enough).
>
> I'm certainly taking liberties here; I intended my brief comment to be
> impressionistic and provocative, not rigorous.
>
> Jonathan
>
>
Received on Monday, 3 October 2011 08:09:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:40 UTC