[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [xri] Use of the word "synonym" in section 10
Les, Actually, the term “polyarchical
synonym” does not appear in the proposed text for section 11. I just used
that term in the example to distinguish between the different types of synonyms
in the two tables. I should probably stop using that term. Let me respond to the question as to
whether the two identifiers =steven.churchill and @ootao*steven should be
called synonyms. This really harkens to the original point that I tried to make
when I started this thread. It is that the term synonym should not be used outside
of a model/context where the concept of synonymity (the binding of identifiers
to identity) is defined. That’s why I feel it doesn’t belong in
section 10. I believe that the term can be used meaningfully and with value within
a constrained discussion regarding the Canonical ID Verification graph model,
because that model clearly lays out meaning of synonymity with respect to
identifier and identity. Yes that model does say @ootao*steven and
=steven.churchill are synonymous identifiers only if they return the same set of highest priority CIDs given the same set
of resolver input parameters (where one of these is cid=true.) From an “external”
perspective, I don’t consider the use of the term synonym misleading here.
A common use case for using CID verification is when a relying party uses a CID
as its primary key for an account record. If the CID =!C5FB.53B6.6E94.824 is
used as this PK, then its seems quite reasonable for my RP to consider
=steven.churchill and @ootao*steven to be synonyms because they both resolve
(for openid sep selection) to the same account. My RP considers them synonymous
identifiers in this respect. ~ Steve From: Chasen, Les
[mailto:les.chasen@neustar.biz] I must apologize here … I did not
look close enough at the XRD and XRDS produced as a result of your query.
You did not change the CID as I accused you of doing, sorry. Since you were doing service selection you
followed a ref in @ootao*steven that went to =!C5FB.53B6.6E94.824 as
demonstrated with this slightly changed query, http://beta.xri.net/@ootao*steven?_xrd_t=http://openid.net/signon/1.0&&_xrd_r=application/xrds+xml;sep=true;.
Since you only wanted the final XRD and SEP that is what you asked for giving
you the modified XRD for =!C5FB.53B6.6E94.824. In your write up you call
this a polyarchical synonym. I would not use that term because, IMHO, the
identity =!C5FB.53B6.6E94.824 is not the same as @ootao*steven. All we
have is a reference from @ootao*steven to =!C5FB.53B6.6E94.824. Also, if you had not told me, I could not
determine that =!C5FB.53B6.6E94.824 is =steven.churchill. Again, you as
the owner are the only one that knows @ootao@steve and =steve.churchill are the
same person. All we can tell is @ootao*steven has a reference and a
reference is not synonymous identities, atleast not to me. - les From: Chasen, Les You have taken @ootao*steven’s CID
and changed it from @!5BAD.2AA.3C72.AF46!0000.0000.3B9A.CA16 to
=!C5FB.53B6.6E94.824 because you did service selection? Where do we say
that the CID of @ootao@steven can be changed because service selection was
done? As I recall the rule for application/xrd processing says that the
only change to the XRD is to filter out all non selected services.
Everything else is preserved. What happened to persistence? From: Steven Churchill
[mailto:steven.churchill@xdi.org] Les, I will shorten the discussion. Hopefully
it will find its asymptote soon. You said: > I agree that is your definition of the word
synonym … actually you call it polyarchical > synonym. I don’t think =steven.churchill
and @ootao*steven are synonyms. > One refers to the other. And I don’t think
they should have the same CID. But they unequivocally do. I don’t
know about you, but when I click on the following two links, I get back the
same CID. http://beta.xri.net/=steven.churchill?_xrd_t=http://openid.net/signon/1.0&&_xrd_r=application/xrd+xml;sep=true; Again, I feel that http://dev.inames.net/wiki/XRI_CanonicalID_Verification
makes this point pretty clear. > =steven.churchill (=!C5FB.53B6.6E94.824) and
@oota*steven (@!5BAD.2AA.3C72.AF46!0000.0000.3B9A.CA16) > do not have the same CID in real life, and never will
unless you run your > spaces how can they have the same CID. We still
state that CID needs to be fully qualified based on the parent, don’t we? See the two links above. ~ Steve From: Steven Churchill
[mailto:steven.churchill@xdi.org] Les, At some point it will be time to give this
discussion up. I don’t think we are quite there yet. [Chasen, Les] L > I don’t know how else to say it
… in my mind a synonym is simply those XRI’s > that resolve to the same XRD or if
you want the same CID(s). I say CIDs > because it is possible to have more than
one persistent identifier that are synonymous. So, two XRIs are synonymous [with respect
to resolver input parameters?] if they resolve to the same set of CIDs. Again,
this is a fine model. [Chasen, Les] Yes. Input parameters should not play a
role in this definition. Would you agree that a second (and also
equally valid) model of synonymity is that two XRIs are synonymous (with
respect to resolver input parameters) if they resolve to the same set of
highest priority CIDs under Canonical ID verification? [Chasen, Les] I agree this is another model … I
don’t agree with it though. 1. I don’t think this definition
should be respective of input parameters and two I don’t agree with the
term highest priority CID. If there are two CIDs in an XRD and one iname
then all three are synonymous irrespective of the highest priority. Your statement, “In my mind a
synonym is …” is really another way of saying “under this
model, a synonym is …”. I am just reinforcing [well, repeating] my
point here: synonymity as a concept does not exist outside the model. It should
not be used within the XRI specification outside of the model’s
definition of identity. [Chasen, Les] OK > =les and =les.chasen and
=!3697.D141.3C5E.742F are synonyms because they all > resolve to the same XRD. (http://xri.net/=les?_xrd_r=application/xrds%2bxml).
> That is the model I work with.
Granted my model maybe slanted by the world I deal with every day. Again, unless we want to define the word
“synonym” in the spec to apply only under this model, then
let’s not use it. [Chasen, Les] I don’t know how this started but this
thread is about the definition of the word synonym. This is my view. > This XRD has a REF in it but I
don’t think that is a synonym. It is just a reference to another
authority. Agreed. [Chasen, Les] This is something J > I don’t believe input
parameters should play a role in this definition. Input parameters > should deal with the data …
with out processing it. Do you not agree that the XRIs
=steven.churchill and @ootao*steven are synonyms (under the Canonical ID
verification model) under one set of input parameters and not synonyms under
another? (I think that is made pretty clear at http://dev.inames.net/wiki/XRI_CanonicalID_Verification.) [Chasen, Les] I agree that is your definition of the word
synonym … actually you call it polyarchical synonym. I don’t
think =steven.churchill and @ootao*steven are synonyms. One refers to the
other. And I don’t think they should have the same CID.
=steven.churchill (=!C5FB.53B6.6E94.824) and @oota*steven
(@!5BAD.2AA.3C72.AF46!0000.0000.3B9A.CA16) do not have the same CID in
real life, and never will unless you run your own @ and = name spaces.
And even if you do run your own = and @ names spaces how can they
have the same CID. We still state that CID needs to be fully qualified
based on the parent, don’t we? I don’t have issues with your CID verification model
.. .i like it. I think our only difference is terminology. Sans the
‘highest priority’ stuff which makes me a little uncomfortable. > I would not call =steven.churchill
and @ootao*steven polyarchical synonyms … maybe > polyarchical references or just
references but not synonyms.
[Chasen, Les] No. see above. > Put simply the iname and cid are the
identifiers. If two or more of them exist in one authority they are
synonymous. I think you are now trying to create yet
another definition of synonymity. [Chasen, Les] no .. I am saying exactly the same thing I
said above. > Again .. this is just my
opinion. I do not care enough to debate it endlessly. Nor do I, but I’m not about to sit
back (especially at this point so late in the game) and ignore the statements I
see being made without rebuttal. [Chasen, Les] I am not seeing that we differ very
much. You seem to be saying that we need to have a definition that allows
for a polyarchical relationship. I am in full support of allowing a
polyarchical relationship through ref processing. I just rather not call
that relationship synonymous … but I can live with it. ~ Steve From: Steven Churchill
[mailto:steven.churchill@xdi.org] Les, In addition, your two definitions of
synonymity below have both left out completely the specification of whether
your identifiers are only synonyms with
respect to a given set of resolver input parameters.
~ Steve From: Steven Churchill
[mailto:steven.churchill@xdi.org] Les, I have to apologize for not letting go on
this. (“Letting go” on stuff like this is the primary reason this
TC can be far along in its process without having a clear agreement and
understanding of what constitutes the notions of identifiers, synonymity and
identity.) You said: > Synonyms are to mean things that are the same
identity. Today > http://xri.net/=les
and http://xri.net/=les.chasen and http://xri.net/=!3697.D141.3C5E.742F
> are synonyms because they are all the same identity and
produce the same XRD > except for those variable tags like query. I must restate my point. (At least I am
consistent): The statement that these identifiers are “synonyms
because they are all the same identity” is meaningless outside a model
defining of what constitutes identity. If you want to define a model that says
that “XRI identifiers are synonymous [perhaps with respect to resolver
input parameters?] because XRI resolution produces the same XRD except for
those variable tags like query”, then that is a perfectly fine model and
definition for synonymity. Note, however, that hat it is a different model altogether
from the three models presented in my email to Markus. That is, that the above
XRIs are not
synonyms under any of those models! To summarize and repeat myself:
identifiers cannot be called synonyms outside of a model that establishes the
meaning of synonymity. > I feel I am simplifying to a very straight
forward model. All XRIs that *are* the same identity are
synonymous. You know this because they resolve to the same CID(s).
But you said above that they are synonyms
because “produce the same XRD except for those variable tags like
query.” So now you are suggesting a second model for synonymity and
identity. You cannot have it both ways. Yet another model is that they are only
synonyms if the XRDs pass canonical ID verification. So again, we have three
different models. Certainly you are not proposing that the above XRIs are
synonyms under all three models. ~ Steve From: Chasen, Les
[mailto: From: Steven Churchill
[mailto:steven.churchill@xdi.org] Les, Yes we agree that the term in used too
often. In my view, I’m trying to eliminate it its use outside of model
that clearly establishes its definition. And thus I find myself needing to be
somewhat pedantic in my response. > from your examples, IMHO
=steve.churchill and =steven.churchill and the CID > that goes with it are synonymous
because they are the same identity. These identifiers cannot be synonymous
outside of a model constituting the definition of synonymity and identity. For
example, three separate models were presented in my previous email to Markus.
The identifiers above may be synonyms in all, some, or none of these models. So
I disagree with you saying that “they are synonymous”. This has no
meaning to me outside you first establishing the model. [Chasen, Les] [Chasen, Les] I am not disagreeing with you on the need for
a model. I laid out my view. > @ootao*steve is a reference that may
be the same person but no one can know that > for sure except you. It is just
a reference between two authorities. > I don’t really care if you want
to call a reference a synonym I just find that that term is used to often. In none of my examples did I say that
@ootao*steve was contained in the value of an xrd:Ref. If I we’re to
establish a reference to @ootao*steve, I would probably opt to use its CID in
the xrd:Ref element anyway. I think that what you are getting at here
is that the value contained in a Ref element could be thought of as a
“synonym” under some particular model of synonymity. Perhaps so,
but if that is how the Resolution Spec decides to employ the term, well, then
that is quite unfortunate, because it prevents or greatly discourages using the
term where it actually provides some real value (such as within the context of
describing the Canonical ID Verification model.) [Chasen, Les] [Chasen, Les] I am saying the value in a Ref is *not* a
synonym. That is however what I thought you said. I think it is a
reference nothing more and nothing less. Synonyms are to mean things that
are the same identity. Today http://xri.net/=les
and http://xri.net/=les.chasen and http://xri.net/=!3697.D141.3C5E.742F
are synonyms because they are all the same identity and produce the same XRD
except for those variable tags like query. In the context that you suggest (and which
the Resolution Spec currently employs), the term only serves to confuse. [Chasen, Les] [Chasen, Les] I feel I am simplifying to a very straight
forward model. All XRIs that *are* the same identity are
synonymous. You know this because they resolve to the same CID(s).
But you seem to think this view has no practical value. ~ Steve From: Steven Churchill
[mailto:steven.churchill@xdi.org] Les, > I agree. I think the word ‘synonym’
has been overused. IMHO, It should be reserved > for inames and inumbers that are in
the same XRD. I think that you are saying that a value of a CanonicalID is
synonymous with the value of LocalID is synonymous with the value of a Ref is
synonymous with the value of a Query given that the values show up in the same
XRD. This definition of “synonym” has no practical
value that I can see. On the other hand, the property of XRI identifier
synonymity is one of XRI Resolution’s most valuable features. I would
hate to “lose” the use of the term by associating it with your
meaning above. ~ Steve From:
markus.sabadello@gmail.com [mailto:markus.sabadello@gmail.com] On Behalf Of Markus Sabadello
On 6/8/07, Steven
Churchill <steven.churchill@xdi.org>
wrote: Drummond,
I
don't understand the use of the term "synonym" in section 10. In English,
if two words are synonyms, then they have the same (or roughly the same)
meaning. In the Identify field, if two identifiers are synonyms, then they
refer to the same identity. As
I try to express in my text for section 11, these concepts are dependent upon
the "model" (or "system") defining the notions of
synonymity and identity. In a Banking system, identifiers may include a
person's name(s) and account number. In that same model, identity may be defined as "the human
being person having a given SSN and DOB". (Don't get confused by the fact
that a SSN is also a type of identifier. That is incidental to the model. We
are talking here about establishing the definition of identity. Once that is done, then Steven
Churchill and Steve Churchill can be synonymous identifiers for the same
identity.) In the XRI Canonical ID Verification model, there is a clear notion
of synonymity and absolute identity. In any case, the model needs to be
formally defined before we can start referring to this thing or that thing as a
being a synonym. A synonym to what? Section
10 introduces the term synonym without establishing a model that defines either
synonymity or identity. As Gabe and others have correctly pointed out, XRI
supports many such models. What are synonyms within one model are not
necessarily synonyms within another model. It
is not helpful to use the term synonym outside a model defining how two
identifiers get mapped to the same identity. ~
Steve
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]