W3C home > Mailing lists > Public > public-credentials@w3.org > May 2015

Re: Suggested renaming of RDF canonicalization spec

From: Gregg Kellogg <gregg@greggkellogg.net>
Date: Tue, 26 May 2015 11:54:39 -0700
Cc: public-credentials@w3.org
Message-Id: <8911B712-7CA9-4E09-A748-B346C9E3DF7D@greggkellogg.net>
To: David Booth <david@dbooth.org>
> On May 26, 2015, at 9:52 AM, David Booth <david@dbooth.org> wrote:
> 
> At present the draft at
> http://json-ld.github.io/normalization/spec/index.html
> usually uses the term "normalization" instead of "canonicalization", particularly in the title.  I think it would be clearer if we changed the spec throughout to use the term "canonicalization" instead, because "canonicalization" is more specific and more commonly used for this concept.
> 
> As a data point, the wikipedia entry for "normalization" lists *21* different senses of the term:
> http://en.wikipedia.org/wiki/Normalization
> In contrast, the wikipedia entry for "canonicalization" has only one -- the one we want:
> http://en.wikipedia.org/wiki/Canonicalization

There’s other discussion that is interesting: http://english.stackexchange.com/questions/35860/normalization-vs-canonicalization

I think that one argument for Normalization, vs Canonicalization, is that C14N is often used in different ways for RDF and XML (which will never be completely dis-entangled):

* N-Triples as a Canonical form (as does N-Quads, which we make use of). This is more to do with the syntax of a particular serialization than the notion that two N-Triples documents can be compared purely at a syntactic level.
* XMLLiterals are typically compared using Exclusive XML Canonicalization, which really has to do with the serialization form, rather than an abstract form such as a normalized dataset. 

OTOH, XMLDSIG [1] uses the term Canonicalization, and as digital signatures are a primary use case, using similar nomenclature makes sense.

Normalization may be a more “friendly” term for many consumers who probably don’t really care about the mathematical definition, but are more concerned with 

At this point, Normalization may be less overused than canonicalization.

Otherwise, I don’t have a strong opinion, and would go with the consensus.

To track this, I’ve created a GitHub issue [3], please redirect discussion to that thread.

Gregg

[1] http://www.hpl.hp.com/techreports/2001/HPL-2001-293.pdf
[2] http://www.w3.org/TR/xmldsig-core/
[3] https://github.com/json-ld/normalization/issues/2

> Thanks,
> David Booth
> 
Received on Tuesday, 26 May 2015 18:55:10 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 11 July 2018 21:19:23 UTC