Re: RDF Canonicalization is a Rec

> On 27 May 2024, at 23:00, Gregg Kellogg <gregg@greggkellogg.net> wrote:
> 
>> On May 26, 2024, at 12:27 AM, Ivan Herman <ivan@w3.org> wrote:
>> 
>> 
>> 
>>> On 25 May 2024, at 16:07, Melvin Carvalho <melvincarvalho@gmail.com> wrote:
>>> 
>>> 
>>> 
>>> út 21. 5. 2024 v 12:48 odesílatel Phil Archer <phil.archer@gs1.org <mailto:phil.archer@gs1.org>> napsal:
>>>> Creating standards is never quick. Nor should it be. It takes time to reach consensus that something needs to be standardized, to ensure that there is a community that wants it and will implement it. Then you actually create the standard itself and prove it works. In this case, that initial phase took a long time. The need for an RDF Canonicalization algorithm was discussed back at the start of the Sem Web but there wasn't the consensus that it was needed or that there was a best way to do it. Once that need was agreed the work kicked off in September 2022 and now, less than 2 years later, RDFC is a W3C Recommendation. See https://www.w3.org/news/2024/rdf-dataset-canonicalization-is-a-w3c-recommendation/
>>>> 
>>>> If you have an RDF graph and you want a canonical representation of it, for whatever reason, there's now a W3C Rec for that, along with a test suite and implementation reports.
>>>> 
>>>> Thanks are due to many people who have thought about this over the years including Jeremy Carrol and Aidan Hogan.  For the production of the Rec itself, thanks go especially to Dave Longley, David Lehn and Manu Sporny at Digital Bazaar; Gregg Kellogg as lead editor supported by Dave Longley and Dan Yamamoto. Active WG members include Sebastian Crane, Ted Thibodeau and Kazue Sako. Markus Sabadello and I had the privilege of working together as co-chairs with W3C staff Pierre-Antoine Champin and Ivan Herman.
>>>> 
>>>> Thanks everyone!
>>> 
>>> Congrats!
>>> 
>>> Does it support RDF 1.2 out of the box?
>> 
>> Unfortunately, no. 
> 
> RDF 1.2 is really not far enough along to target.
> 
>> The algorithm works with the current official version of n-quads, meaning that it is not prepared to handle Triple Terms[1].
>> 
>> I see two ways to solve this (but, honestly, I have not given too much thoughts):
>> 
>> 1. the algorithm has to be modified to handle triple terms.
>> 2. we define a "canonical" mapping of RDF1.2 to N-Quads without triple terms.
>> 
>> (1) might be a delicate process: the algorithm is fairly complex, so are the implementations, and it is always difficult (and error-prone) to engage in this type of revision. (The fact that RDFC is closely related to security-related applications makes it even more delicate.) I would think that the need for (2) may arise independently of RDFC-1.0, in order to use deployed implementations of RDF for the purpose of RDF 1.2. (I do not follow the development of RDF 1.2 very closely, but I seem to remember that some work was done in 2 when we still called these things RDF-Star. But I might be mistaken.)
> 
> I believe that RDF 1.2 will define a mapping from RDF 1.2 Full (including triple terms) to a “classic” profile, which does not contain triple terms (perhaps uses classic reified triples). A hypothetical update fo RDF Dataset Canonicalization would be to transform the incoming dataset from Full to Classic prior to running the RDC algorithms. Other changes in RDF 1.2 (e.g., literals with text-direction) should not require any changes to the RDC algorithms.

That would be my favorite approach indeed. This would mean that the current algorithm (and the core of the implementations) would not change, and RDF 1.2 would only mean a pre-processor of some sort. I personally would be scared touching the edifice of the RDFC algorithm…

Ivan


> 
> Gregg
> 
>> In any case, this must be reopened as soon as RDF-1.2 is closed to its own finish line.
>> 
>> Cheers
>> 
>> Ivan
>> 
>> [1] https://www.w3.org/TR/rdf12-n-quads/#triple-terms
>> 
>> 
>>>  
>>>> 
>>>> Phil
>>>> 
>>>> 
>>>> ---
>>>> 
>>>> Phil Archer
>>>> Web Solutions Director, GS1
>>>> https://www.gs1.org <https://www.gs1.org/>
>>>> 
>>>> https://philarcher.org <https://philarcher.org/>
>>>> +44 (0)7887 767755
>>>> @philarcher1
>>>> 
>>>> CONFIDENTIALITY / DISCLAIMER: The contents of this e-mail are  confidential and are not to be regarded as a contractual offer or acceptance from GS1 (registered in Belgium). 
>>>> If you are not the addressee, or if this has been copied or sent to you in error, you must not use data herein for any purpose, you must delete it, and should inform the sender. 
>>>> GS1 disclaims liability for accuracy or completeness, and opinions expressed are those of the author alone. 
>>>> GS1 may monitor communications. 
>>>> Third party rights acknowledged. 
>>>> (c) 2020.
>>>> 
>>>> 
>> 
>> 
>> ----
>> Ivan Herman, W3C 
>> Home: http://www.w3.org/People/Ivan/
>> mobile: +33 6 52 46 00 43


----
Ivan Herman, W3C 
Home: http://www.w3.org/People/Ivan/
mobile: +33 6 52 46 00 43