> - sotd pagraph: (this document) is on prov-constraints, and self reference to [prov-sem] to be removed.
 > 

Done

 > - section 2.2: for convenience, we assume the order is total. What is the implication of this choice?

Raised ISSUE-635
 > 
 > - section 2.4: wasStartedBy/wasEndedBy are missing time

Fixed

 > 
 > - section 2.4: no bundle here? Shouldnt' we? or too slippery topic?

Added some explanation of this in the introduction

 > 
 > - section 2.5: could the final version have a go at an interpretation of mentionOf, even if marked as more speculative and in appendix?

I would prefer to avoid this.  If there's time maybe.

 > 
 > - section 3.2: term interaction not defined yet.

Fixed

 > 
 > - section 3.2.1: Can we provide some intution for value(obj,a) \subseteq value(thingOf(obj),a,t). What does it tell us about entities vs things.

Done

 > 
 > - section 3.2.4: (the first two sets may overlap) which? Why?

Fixed - this remark was out of date.

 >  Events is not defined yet.

Fixed - it's a forward reference to later in the section

 > 
 > - section 3.2.4.1: lifetime(evt)={time(t)} isn't it time(evt)?

Fixed

 > 
 > - section 3.3: EventActivity I think allows for an event not to be associated with any activity. Is this desirable?

It's tru that this is allowed.  We could make it a function, since every event is associated with exactly one activity.  At the time this was written this wasn't clear.

 > 
 > - 3.3.1: function derivedFrom linking each derivation to ITS path:
 >  I would suggest
 >        function derivedFrom linking each derivation to SOME path.
 >     in the case wasDerivedFrom(e1,e2,-,-,-) there may be multiple paths.
 > 

True, but since we are always considering formulas that specify an identity for the derivation relationship, I believe each distinct derivation has its own path.  I added a remark to this effect.

 > 
 > - section 4.3.1: we could add some intution: not all attributes of an
 >  entity object are required to be present in an entity formula.

Done

 > 
 > -  4.3.2, rule_13:
 >    "if st is specified", what do you mean? st is always present as a constant or variable.  Do you mean "if st is a variable"?
 >    How do we have to read min(lifetime(id))=st if st is variable?

Fixed.  A time can potentially be a variable if it is omitted and then expanded to an existential variable.  So we need to write rho(t) when we need its value.

 > 
 >   Elsewhere, when time occurs in a formula,  there is no clause "if t is specified".
 >   Why this differnce?

Fixed, this phrasing was outdated.

 > 
 > - formalism 16: who do we read time(evt)=t when t is variable?

Fixed, now reads rho(t)

 > 
 > - 4.4.7: nothing in the semantics seems to indicate that lifetime of
 >  agent precedes that of entity.

This constraint (and a number of other constraints) are not yet reflected in the semantics.  They will be, but doing this is somewhat repetitive, and I'm undecided as to the best way to do it.  This can be discussed under ISSUE-635.

 > 
 > - 4.4.12: there is some occurrence of obj_1/obj_2   which should be ent_1/ent_2
 > 

Fixed

 > - formalism 29 (precedes)
 > 
 > 
 > entity(e)
 > activity(a1)
 > activity(a2)
 > wasGeneratedBy(gen1; e, a1, 2011-11-16T16:05:00)
 > wasGeneratedBy(gen2; e, a2, 2012-11-16T16:05:00) //different date
 > 
 > 
 > gen1 <= gen2 and gen2 <= gen1
 > 
 > 
 > Formalism 29 implies: 2011-11-16T16:05:00 == 2012-11-16T16:05:00
 > (see separate email)

Raised ISSUE-635