W3C home > Mailing lists > Public > public-prov-wg@w3.org > June 2011

Re: PROV-ISSUE-8: defining generation in terms of `IVPT of'

From: Graham Klyne <GK@ninebynine.org>
Date: Sun, 12 Jun 2011 08:26:35 +0100
Message-ID: <4DF46A2B.3050800@ninebynine.org>
To: Luc Moreau <L.Moreau@ecs.soton.ac.uk>
CC: "Myers, Jim" <MYERSJ4@rpi.edu>, "public-prov-wg@w3c.org" <public-prov-wg@w3c.org>
Luc Moreau wrote:
> Jim
> A key difference between your definition and mine is that for me the generation relationship is between an ipvt and a process, for you it's between a thing and a process.
> Khalid's new derivation is also between ipvts.
> I thought last week's agreement Was that we were focusing on ipvts, which give us the kind of invariance we need for provenance.
> How do we reconcile this apparent disagreement?

Adapting my earlier response:

By understanding (implicitly?) that a thing might be an IVPT of some other 
thing.  Thus, we use the notion of IPVT in the explanations, but the definitions 
still apply to things.


To the extent that I properly understand them, I find Jim's explanations are 
capturing the same issues that I'm trying to express.


> Professor Luc Moreau
> Electronics and Computer Science
> University of Southampton 
> Southampton SO17 1BJ
> United Kingdom
> On 10 Jun 2011, at 21:02, "Myers, Jim" <MYERSJ4@rpi.edu> wrote:
>>> In this definition, you say when X didn't exist(before P) but you
>> don't say
>>> when it starts to exist. Is it intended?
>> I updated the wiki as follows:
>> * X did not exist before P began,
>> * X began to exist sometime before P ended, and
>> * X would not have begun existing if P had not occurred/P was necessary
>> for X's existence to happen.
>>> Also, did you really mean before P began? Can you explain why ?
>> If X is generated by P, it can't have existed before P did. I think this
>> is just saying P must start before the creation of X if X is generated
>> by P.
>> Is it a matter of wording or is something mixed up logically?
>>> Can X change, i don't understand what happens then? Does the last
>> condition
>>> necessarily apply?
>> The generated thing/resource X can change after being generated, but
>> only in ways consistent with its definition/type. An egg (defined as an
>> ovoid object produced by a chicken) generated by a chicken can
>> participate in boiling and painting (but I can't say what color such an
>> egg is since color is not part of the definition of what makes one
>> instance of the egg class unique/identifiable.) Other definitions of egg
>> (IVPTs of each other) differ in how mutable they are. If you want a
>> fairly immutable concept of egg, define it as a ovoid product of a
>> chicken that has a particular color, is cooked/uncooked,
>> cracked/uncracked, etc. Then we can talk about the Easter Egg that was
>> derived from the uncooked/uncolored egg from a particular chicken
>> through a series of processes. (Or, we could just say that after being
>> generated, the egg sat and at some point we recognized that it existed
>> in a state that corresponds to our notion of Easter egg and we assert
>> that EasterEgg Y is an IVPT of egg X.)
>> Does the last condition apply? - Is that "must there always be a Y that
>> X is an IVPT of"? I don't know, but I suspect the answer is yes in any
>> practical sense. By the IVPT definition I posted, all I need is some
>> process P that would destroy/consume X and define Y as "X considered
>> mutable with respect to P" - "the idea of this egg that would be
>> unaffected even by the end of the universe" or some such might be the
>> end of the line though.
>>> Are you also saying there needs to be another notion? Modification?
>> I think the other notion is some form of participation (also uploaded
>> definitions there). I was saying participates in/is modified by for the
>> thing/process relationship. (Agency/control would be one form of
>> participation). The egg layed by the chicken participates in boiling and
>> coloring during its lifetime. It is not modified in the sense that any
>> part of what identifies it changes, but attributes of it we consider
>> transient (can't be used to discover this egg reliably because they
>> change) do change.
>> -- Jim
>>> Professor Luc Moreau
>>> Electronics and Computer Science
>>> University of Southampton
>>> Southampton SO17 1BJ
>>> United Kingdom
>>> On 10 Jun 2011, at 18:14, "Myers, Jim" <MYERSJ4@rpi.edu> wrote:
>>>> Posted on the wiki - basically the same definition as from Jun and
>>>> Daniel with more discussion of how the IVPT concept and the idea of
>>>> modifiable objects relate.
>>>> -- Jim
>>>>> -----Original Message-----
>>>>> From: Luc Moreau [mailto:L.Moreau@ecs.soton.ac.uk]
>>>>> Sent: Friday, June 10, 2011 12:57 PM
>>>>> To: Myers, Jim
>>>>> Cc: public-prov-wg@w3c.org
>>>>> Subject: Re: PROV-ISSUE-8: defining generation in terms of `IVPT
>> of'
>>>>> Jim and all,
>>>>> Could you put forward a revised definition that addresses better
>> your
>>>>> concerns?
>>>>> Professor Luc Moreau
>>>>> Electronics and Computer Science
>>>>> University of Southampton
>>>>> Southampton SO17 1BJ
>>>>> United Kingdom
>>>>> On 10 Jun 2011, at 17:51, "Myers, Jim" <MYERSJ4@rpi.edu> wrote:
>>>>>>> Hi Jim
>>>>>>> I think we are discussing two issues here.
>>>>>>> 1. You suggest that the egg is itself an IVPT.
>>>>>>>  There are different ways of looking at this:
>>>>>>>  a. I was saying that an egg was a thing (identified, and  typed
>>>>>> according to
>>>>>>> an ontology)
>>>>>>>      Then, it's a question of choice of a same ontology, or
>>>>>> ontology
>>>>>>> refinement, a classical problem,
>>>>>>>      which we will not solve here.
>>>>>> I would claim the things in IVPT relationships with egg are also
>>>>>> things that can be identified, typed, etc., not necessarily in the
>>>>>> same class or ontology as 'egg'. (A logical picture is not the
>> same
>>>>>> class of thing as a jpg file corresponding to a particular
>>>>>> manifestation of that image).
>>>>>>>  b. We could say that there is in an alternate account, which
>>>>>> describes the
>>>>>>> egg in terms of molecules.
>>>>>> Yes - my object in that account is 'set of molecules' and 'egg' is
>> a
>>>>>> convenient label for when those molecules are in a particular
>> subset
>>>>>> of all the configurations they can be in. If you say egg is an
>>>> object
>>>>>> and require a different type of thing to be used to describe
>> things
>>>>>> that invariant views of my set of molecules, I can't use it in my
>>>>>> account, and we don't yet have any mechanism to make it clear that
>>>>>> somehow my 'state of set of molecules' corresponds to your notion
>> of
>>>> 'egg'.
>>>>>>>  c.  Alternatively, we have IVPTs of IVPTs of IVPTs ...
>>>>>>>       is there a base case? I fear we are going to reach quantum
>>>>>> mechanics ...
>>>>>> I don't think we have to be afraid of this - and I would rephrase
>>>> and
>>>>>> say we have things of different types that can be in IVPT
>>>>>> relationships with each other and your concern is then whether the
>>>>>> fact that we can make deep hierarchies is an issue. I'd answer
>> that
>>>> by
>>>>>> saying that the base case is in the middle - things like eggs are
>>>>>> useful not because they are somehow true objects where other
>> things
>>>>>> are just views, they are useful views because of the
>> natural/common
>>>>>> processes they participate in. The fact that the model allows one
>> to
>>>>>> describe a set of quantum wave functions and claim one view of
>> them
>>>> is
>>>>>> an egg doesn't mean that this will happen in practice (though
>> there
>>>>>> are scientists who do essentially this on a daily basis at the
>>>> nano-scale).
>>>>>> Whether it is truly 'turtles all the way down' is a philosophical
>>>>>> question I'm not sure we have to answer - modeling it that way
>>>> covers
>>>>>> the middle ground without requiring any connection to real base
>>>>>> objects (or more neutrally, without identifying a particular set
>> of
>>>>>> objects as real with all others some form of constructed view) -
>> in
>>>>>> this sense, I would ask you a) whether you see a
>>>>>> consequence/limitation of a model that does not define which
>> objects
>>>>>> are 'real'? and b) given the debates about ontologies in the
>> world,
>>>> do
>>>>>> you think we can reach a consensus on what the base reality is?
>>>>>>> 2. You are commenting on the word modified.
>>>>>>>    If I crack the egg,
>>>>>>>      Y-> crack -> X
>>>>>>>      Y and X are IVPTs of egg
>>>>>>>      Y->X (we have a derivation)
>>>>>>>    So looking at generation only, I feel it's OK to say the egg
>>>> is
>>>>>> modified, since
>>>>>>> we
>>>>>>>   have now a new IVPT Y about the same egg.
>>>>>>> So, could you maybe make some suggestions on how you would revise
>>>> the
>>>>>>> definition?
>>>>>> I'm not sure what point you're making about the cracking example,
>>>> but
>>>>>> I'd say generation is just a case where we are more
>>>>>> familiar/comfortable with the thing produced by a process
>> execution
>>>> as
>>>>>> a useful thing to discuss/track the provenance of, and potentially
>>>>>> where the inputs of the process execution are uninteresting. A
>>>> chicken
>>>>>> lays an egg not because we can't talk about a set of atoms that
>> the
>>>>>> chicken rearranges into a state we want to identify as an egg but
>>>>>> because that view is not very useful, so identifying the 'set of
>>>> atoms
>>>>>> in the chicken' that is used to produce the egg or the 'set of
>>>> atoms'
>>>>>> that exists before and after egg laying that comprise the egg
>> after
>>>>>> laying isn't useful and we record chicken controls eggLaying which
>>>>>> generates egg. Both/all variant accounts are valid and consistent
>> in
>>>>>> the model I'm advocating, specifically egg does not change type
>> from
>>>>>> being a real object to a state of a set of atoms - 'set of atoms'
>> is
>>>> a
>>>>>> real thing that has a stateful view corresponding to a real thing
>>>>>> called an egg (an IVPT relation with egg). The fact that 99+% of
>> us
>>>>>> would just report generation of an egg and stop is OK (good in
>> fact
>>>> -
>>>>>> we don't want to needlessly talk about alternate views any more
>> than
>>>>>> we should feel pressure to expand all processes into fine grained
>>>>>> steps or include info about the movement of electrons in
>> describing
>>>>> computations).
>>>>>> Thanks,
>>>>>> Jim
>>>>>>> Thanks,
>>>>>>> Luc
>>>>>>> On 06/10/2011 02:28 PM, Myers, Jim wrote:
>>>>>>>>> -----Original Message-----
>>>>>>>>> From: Luc Moreau [mailto:L.Moreau@ecs.soton.ac.uk]
>>>>>>>>> Sent: Friday, June 10, 2011 3:26 AM
>>>>>>>>> To: Myers, Jim
>>>>>>>>> Cc: public-prov-wg@w3c.org
>>>>>>>>> Subject: Re: PROV-ISSUE-8: defining generation in terms of
>> `IVPT
>>>>>> of'
>>>>>>>>> Hi Jim,
>>>>>>>>> I had not seen your comment in line, my responses are also
>>>> inline.
>>>>>>>>> On 10/06/11 02:28, Myers, Jim wrote:
>>>>>>>>>> This would mean that a heating process modifies an egg to
>> create
>>>> a
>>>>>>>> warm egg,
>>>>>>>>> it does not transform a cold egg into a warm egg?
>>>>>>>>>> Or do you mean both - a process execution can turn one thing
>>>> into
>>>>>>>> another,
>>>>>>>>> these things can be considered IVPTs of a thing that
>> participates
>>>>>> in
>>>>>>>> the process
>>>>>>>>> execution/ is modified by the process execution? And in an open
>>>>>> world
>>>>>>>>> assumption, a witness doesn't have to report the modified thing
>>>> or
>>>>>>>>> can
>>>>>>>> decline
>>>>>>>>> to identify/report either of things in IVPT roles depending on
>>>>>> their
>>>>>>>> ability to
>>>>>>>>> observe and the use case they wish to enable?
>>>>>>>>>> ________________________________
>>>>>>>>>> From: public-prov-wg-request@w3.org on behalf of Luc Moreau
>>>>>>>>>> Sent: Thu 6/9/2011 6:44 PM
>>>>>>>>>> To: Provenance Working Group WG
>>>>>>>>>> Subject: PROV-ISSUE-8: defining generation in terms of `IVPT
>> of'
>>>>>>>>>> Hi all,
>>>>>>>>>> - if a new thing is created, it is clear that we have a new
>>>>>> of
>>>>>>>>>> that thing
>>>>>>>>>> if a chicken creates an egg is it just an IVPT of an egg?
>>>>>>>>> I would think the physical object is the egg.
>>>>>>>>> I thought we had agreed that for a provenance purpose, we had
>> to
>>>>>> talk
>>>>>>>> about
>>>>>>>>> an IVPT of that egg.
>>>>>>>> But 'the egg' is also an IVPT of the stuff that goes into the
>> cake
>>>> -
>>>>>> a
>>>>>>>> temporary 'state' in which yolk and white are together and not
>>>>>>>> mixed/chemically altered, etc.
>>>>>>>>>> - if the thing is modified, then it is a requirement that a
>> new
>>>>>> view
>>>>>>>> (IVPT) is
>>>>>>>>> generated ...
>>>>>>>>>>       otherwise, it would still be a view that existed before
>>>>>>>>>> can't I say the egg was heated without reporting its cold and
>>>> warm
>>>>>>>> states? I.e.
>>>>>>>>> don't we want to be able to report that something was modified
>>>>>>>>> without
>>>>>>>> having
>>>>>>>>> to report the IVPTs? A document was edited four times by
>>>> different
>>>>>>>> people but I
>>>>>>>>> don't wan't to/can't tell you what each wrote at each stage?
>>>>>>>>> These comments were made in the context of defining Generation
>> of
>>>>>> an
>>>>>>>> IVPT.
>>>>>>>> OK - but you said "if the thing is modified"... For generation,
>> I
>>>>>>>> would say the chicken participates in an egg laying process
>>>>>> execution
>>>>>>>> that generates an 'egg'. That 'egg' is an IVPT of the chemicals
>> in
>>>>>> it
>>>>>>>> (which existed before). The 'egg' can also have further/more
>>>>>> stateful
>>>>>>>> IVPTs that are more useful for discussing heating, cracking,
>> etc.
>>>>>> The
>>>>>>>> sense in which generation is special is that it is a derivation
>>>> from
>>>>>>>> things we don't consider logically an aspect/IVPT of something
>>>>>>>> greater. I.e. the 'mass of egg-bound chemicals' in the chicken
>>>> just
>>>>>>>> changes its state to become the 'egg', nothing really appears or
>>>>>>>> disappears (conservation of mass and energy). Unless/until a
>>>>>> scientist
>>>>>>>> wants to look at the potential for different processing of
>>>> chemicals
>>>>>>>> going into the egg versus those used to build the chicken's own
>>>>>> body,
>>>>>>>> 'mass of egg-bound chemicals' isn't something we'd usually think
>>>>>>>> about, but it's a valid perspective and consistent with the view
>>>> of
>>>>>> an
>>>>>>>> egg being generated (both views can be drawn on the same graph
>> in
>>>>>> the
>>>>>>> way I've been describing).
>>>>>>>>> The document was edited four times could be expressed by 4
>>>> process
>>>>>>>> execution
>>>>>>>>> and something like opm:wasTriggeredBy in between.
>>>>>>>>>> - if the process execution was taking a long time to
>>>> modify/create
>>>>>>>> the thing,
>>>>>>>>> there is only one
>>>>>>>>>>   instant at which the (invariant!) IVPT appears
>>>>>>>>>> I thnk we could define it that way, but if a cracking process
>>>>>> takes
>>>>>>>> time, saying
>>>>>>>>> the cracked egg appears instantaneously basically means you
>> want
>>>>>>>> 'cracked egg'
>>>>>>>>> to be defined by some threshold - the cracked egg might become
>>>> more
>>>>>>>> cracked
>>>>>>>>> over time ) invariant only in that it is always above the
>>>> threshold
>>>>>>>> and the
>>>>>>>>> instance of the creation of the IVPT relationship occurs ata
>>>>>>>> aspecific instant.
>>>>>>>>> Yes, agreed.
>>>>>>>>>> - I think this captures well a stateful objects, where
>> processes
>>>>>> can
>>>>>>>> modify the
>>>>>>>>> object, resulting in
>>>>>>>>>>   different IVPTs corresponding to the various states
>>>>>>>>>> IVPTs are not a separate kind of thing and their invariance is
>>>>>>>> relative. If they
>>>>>>>>> are truly immutable sates/snapshopts, they can only exist for
>> an
>>>>>>>> instant because
>>>>>>>>> some part of the state of the thing (a part we may not care
>> about
>>>>>>>>> such
>>>>>>>> as age)
>>>>>>>>> will change immediately.
>>>>>>>>> I am not sure I agree, here. IVPTs are a view/perspective on a
>>>>>> thing.
>>>>>>>> Perhaps I should argue from the other direction - the notion of
>> a
>>>>>>>> thing is also a perspective/view. "eggs" don't exist - only
>>>>>> temporary
>>>>>>>> co-locations of particular molecules. Galton and Mizoguchi's
>> paper
>>>>>>>> argues that objects are defined by what processes you consider
>> to
>>>> be
>>>>>>>> internal to and external to the object - if you change the set
>> of
>>>>>>>> processes you are concerned about, you describe the world using
>>>>>>>> different objects. In this sense, eggs feel more 'real' because
>>>> the
>>>>>>>> set of processes we see happening frequently to them preserve
>>>>>> aspects
>>>>>>>> of their state, so egg as an IVPT is useful/predictive/etc.
>>>>>>>>> Alternative views asserted by other asserters may co-exit.
>>>>>>>>> - it's a decaying egg
>>>>>>>>> - it's a duck egg, not a hen egg
>>>>>>>>> - it's a chocolate egg
>>>>>>>> I agree, so again from the other direction - I don't see why
>> 'egg'
>>>>>> is
>>>>>>>> more real/more correct/more special than any of these, they're
>> all
>>>>>>>> IVPTs.
>>>>>>>>> It is a requirement of any specific perspective to be
>> invariant.
>>>>>>>> Invariant relative to a view in which only certain processes are
>>>> of
>>>>>>>> interest (are observable/reported?).
>>>>>>>>> So, it's not a snapshot of the global egg state, but it's a
>>>>>> snapshot
>>>>>>>> according to a
>>>>>>>>> view.
>>>>>>>>> An analogy would be several cameras pointing to a same egg.
>>>>>>>>> From one camera, the egg is still, no change occurring.
>>>>>>>>> From the other, we see a crack appearing.
>>>>>>>>> So one asserter can describe change in a physical object, while
>>>>>>>> another does not
>>>>>>>>> describe any change.
>>>>>>>>> But it's the same egg.
>>>>>>>> Right - those are two views of something. But that something is
>>>> just
>>>>>>>> another view - one camera sees a soup of chemicals that are
>>>> swirling
>>>>>>>> and mixing (slowly for an egg) while another sees one thing (the
>>>>>> egg).
>>>>>>>> Both of these are just IVPTs too. The chemical view is mutable
>> by
>>>>>> more
>>>>>>>> processes than the egg view, but it is more persistent (it lasts
>>>>>>>> longer because we've defined it in a way that the processes that
>>>> can
>>>>>>>> create/destroy it are less frequent).
>>>>>>>> I know that thinking of everything as an IVPT is not necessarily
>>>>>>>> intuitive, and that one can argue that it is just one way to
>> model
>>>>>> the
>>>>>>>> world/a philosophy, but I think it is a model that has the right
>>>>>>>> conceptual power to deal with the use cases we have (and the
>>>> general
>>>>>>>> set we can envision) while also being one that, in practice,
>> will
>>>>>> fade
>>>>>>>> away
>>>>>>>> - most people will agree that 'egg' and not 'mass of chemicals'
>> is
>>>> a
>>>>>>>> more useful IVPT to talk about and we'll see 'eggs' used in cake
>>>>>>>> baking and the world will mostly look like OPM (straight
>>>>>>>> thing-execution-thing chains), but we'll still have the power to
>>>>>> drop
>>>>>>>> down and talk about cracking or go up and talk about
>> conservation
>>>> of
>>>>>> mass
>>>>>>> when needed.
>>>>>>>> I also don't know what a coherent alternative is that, once we
>> add
>>>>>> in
>>>>>>>> all the features necessary to cover the use cases, we'll like
>>>>>> better.
>>>>>>>> There are certainly other ways to model - my question really is
>>>>>>>> whether there are others that will end up being more intuitive
>>>> once
>>>>>>>> all the needed features are dropped in.
>>>>>>>> -- Jim
>>>>>>>> Jim
>>>>>>>>> Luc
>>>>>>>>>> What do you think?
>>> http://www.w3.org/2011/prov/wiki/ConceptGeneration#Definition_of_Gene
>>>>>>>>> r
>>>>>>>>>> ation_by_Luc
>>>>>>>>>> Cheers,
>>>>>>>>>> Luc
>>>>>>> --
>>>>>>> Professor Luc Moreau
>>>>>>> Electronics and Computer Science   tel:   +44 23 8059 4487
>>>>>>> University of Southampton          fax:   +44 23 8059 2865
>>>>>>> Southampton SO17 1BJ               email:
>> l.moreau@ecs.soton.ac.uk
>>>>>>> United Kingdom
>> http://www.ecs.soton.ac.uk/~lavm
Received on Sunday, 12 June 2011 07:32:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:31 GMT