Re: Suggestions for Errata RDFa Core 1.1

The Wikipedia entry for "Entailment" redirects to "Logical Consequence" [1], which corresponds to my understanding of the term. The German entry for this is Implikation [2], or "Implication" in English. This might be a good term to use. I've CC'd Richard Cyganiak, who is the editor of RDF Concepts, which defines Entailment [3]. I also believe that Richard is a native German speaker, he might be able to help.

Gregg Kellogg

[1] http://en.wikipedia.org/wiki/Entailment
[2] http://de.wikipedia.org/wiki/Implikation
[3] http://www.w3.org/TR/rdf11-concepts/#entailment

Sent from my iPad

On Apr 15, 2013, at 1:52 AM, "Stefan Schumacher" <stefan@duckflight.de> wrote:

> Hello Shane,
> 
> thanks for your fast work and your comments.
> 
> I still have some comments on the earlier suggestions, I cut out the 
> fixed things, so the quoted list below is much shorter now.
> I kept the open issues, where you asked for input from the WG.
> 
> I have two more new things, that I discovered while proof reading my 
> translation. They are kind of connected. 
> 
> 7.2 Evaluation context, second list, item 5 
> ---
> now:
> ... new subject value, which once calculated will set the parent 
> subject 'property'...
> --- 
> suggestion:
> ... set the value for the parent subject
> --- 
> comment:
> The use of the word 'property' is my concern. Property in RDFa is 
> used as synonym to 'predicate', here it is used in the sense of the 
> value for a subject.
> 
> Second:
> 7.2 Evaluation context, second list, item 6:
> --- 
> now:
> A value for the current property value, the literal to use when 
> creating triples that have a literal object, or IRI-s in the absence 
> of @rel or @rev.
> 
> This sentence kills me.
> 
> 1) Is it a value for a predicate in general? 
> 2) Is it a value for the attribute @property?
> 3) Is it a value for an object, that is a literal?
> 4) Is it a value for an object, that can be literal or IRI?
> 
> I would call 1) 'current predicate value'.
> I would call 2) 'current property value'.
> I would call 3) 'current object literal'
> I would call 4) 'current object value' 
> 
> My ansers to the above: 
> It cannot be 1) or 2), because they would require 
> TERMorCURIEorAbsIRIs not a literal, like stated in the explanation.
> It cannot be 3), because it could be IRIs also.
> I could be 4), but the term 'current property value' doesn't really
> allow that.
> 
> So what now?
> 
> 
> Below are some more comments to the old stuff. Have to rush out now, 
> some things below I'll finish commenting later.
> 
> Stefan
> 
> 
> On 13 Apr 2013 at 12:34, Shane McCarron wrote:
>> I have gone through your comments.  Thanks so much for the feedback!   My
>> replies are in-line.  An updated draft is available at
>> http://www.w3.org/2010/02/rdfa/sources/rdfa-core/Overview-src.html
>> 
>> There are a couple of comments were I could use responses.
>> 
>> Suggested Errata
>>> for RDFa Core 1.1
>>> http://www.w3.org/TR/2012/REC-rdfa-core-20120607
> 
>>> Assumed Errors
>>> Location in the document
>>> Text in the current document.
>>> Suggested text for errata. 
>>> Comment.
>> 
>> 
>>> 5.0 Attributes and Syntax, definition vocab
>>> A IRI
>>> an IRI
>>> The 'a' or 'an' is in small letters elsewhere in the document.
>> Fixed
> The PER doesn't show it as yet.
> 
>>> 7.6.1, third paragraph
>>> ... allow the developer, if they would ...
> developer, they
>>> ... allow the developers, if they would ...
>>> Similar like above with author.
>> Fixed.
> The PER doesn't show it as yet.
> 
> 
>>> 7.6.1, last paragraph
>>> ... allow the caller to specify if they ...
>>> 1. ... allow the caller to specify if he/she ...
>>> 2. ... allow the callers to specify if they ...
>>> Similar like above.
>> Fixed.
> The PER doesn't show it as yet.
> 
> 
>>> 8.3.1.3 XML Literals
>>> Although the rendering of this page has highlighted the term the user
>>> searched for, setting @datatype to nothing ensures that the data is
>>> interpreted as a plain literal, giving the following triples:
>>> 1. 'Rendering of this ...searched for. Setting ... *Skip 'Although', make
>>> two sentences.*
>>> 2. setting @datatype to 'an empty string'
>>> 3. ... giving the following 'triple'. *Only one triple in the example.*
>> Fixed.
> 
> 
>>> 2. Syntax Overview, last NOTE
>>> In some of the examples below we have used IRIs with fragment identifiers
>>> that are local to the document containing the RDFa fragment identifiers
>>> shown (e.g., 'about="#me"').
>>> In some of the examples below we have used IRIs with 'local' fragment
>>> identifiers (that point to items/fragments inside of the same document) (e.g.,
>>> 'about="#me"').
>>> From my point of view, the expanation of a local fragment identifier in
>>> the current version of the sentence makes more confusion than just saying,
>>> 'local' fragment indentifiers are used. That's it. The rest of the sentence
>>> is saying the same, but makes it confusing.
>> This language came from the W3C TAG and I don't dare change it.
> Ok, I might go to another court and file my case.
> 
> 
>>> 4.2 RDFa Host Language Conformance
>>> All of the facilities required ...
>>> All requirements of this specification have to be included ...
>>> This is just a translator request, facilities can mean anything and is
>>> horrible to translate.
>> This language is historical - facilities is a term of art in the standards
>> industry.  I don't have an appropriate substitution.
> Fine, Terms Of Art shall be. Translators problem then. ;-)
> 
> 
>>> 4.2, note
>>> ... that are commonly used througout the Host Language.
>>> ... that are commonly used throughout documents written in the Host
>>> Language.
>> Changed to 'that are commonly used throughout the content model of the Host
>> Language'
> 
> 
> 
>>> 6., note, The production safe curie ...
>>> cannot
>>> must not
>>> Is this cannot used in the meaning of 'should not' or 'must not'?
>> Neither.  This is not a conformance requirement.  It literally means 'it is
>> impossible for this to happen'.  As in it is a logical impossibility.
> 
> 
>>> 7.2 and maybe other sections
>>> during the course of processing
>>> during/while processing
>>> during or while already say it is between the start and the end of
>>> processing. This 'course' is 'extra' and it complicates the translation
>>> (just a little bit).
>> This means 'things change over time'.  It is a subtle but important
>> concept.  I am reluctant to change it
> 
> 
>>> 7.2, first list, last list item
>>> a value to use as the prefix IRI when an undefined term
>>> unprefixed term
>> No.  What this means a term that is unknown to the processor.  I made a
>> change to clarify that.
> Yes, I like that one.
> 
>>> 7.5
>>> although the evaluation context used 'for each set of rules' will be based
>>> on previous rules that may have been applied.
>>> 'every cycle/run through the set of rules'
>>> (an evaluation context based on the resulting evaluation context from the
>>> last run will be used)
>>> I would prefer, if it is clearer, that the processing rules are applied
>>> again and again for each element, and that the evaluation context normally
>>> changes after each run. My suggestion surely needs some proper english, I
>>> put that on you. :-)
>> I appreciate that the language in this section is stilted.  But it is
>> actually the only important part of this document as far as I am concerned.
>> I don't dare change any of the text without a full review by the working
>> group.
> 
> 
>>> 7.6.1
>>> An web service RDFa Processor is defined as any RDFa Processor that is
>>> capable of processing a document 'by performing an HTTP GET, POST or
>>> similar action on an RDFa Processor IRI'.
>>> no suggestion yet
>>> What is an RDFa Processor IRI?
>>> 1. The IRI where the RDFa Processor is located? Or
>>> 2. the IRI that the Processor should look up and process?
>>> I wouldn't call the second IRI 'RDFa Processor IRI' it is more an 'RDFa
>>> document IRI'.
>>> I would say, an RDF processor is reachable under an 'RDFa Processor IRI'.
>>> For my translation, it would be nice if I get a response to this issue.
> 
>> It is the IRI for an RDFa Processor - in other words, the address someone
>> would use to query the processor via the web and extract triples from an
>> RDFa document.
> 
> 
>>> 8.3
>>> A literal object can be set by using @property to express a predicate, and
>>> then using either @content, or the inline text of the element that
>>> @property is on.
>>> A literal object can be set by @content or the inline text of the element,
>>> if @property is used to express a predicate.
>>> 'Using @content' is not precise, say what @content and the inline text do:
>>> they provide the literal.
>> Made a change that helps.   Some.
> 
> 
>>> 8.3, last paragraph
>>> Alternatively, the @property can also be used to define an IRI resource,
>>> in the presence of an @href, @resource, or @src and in the absence ... .
>>> Alternatively, @property can be used to define an IRI resource; this
>>> requires the presence of ... and requires the absence of ... .
>>> This would give the sentence a sharper edge.
>> Fixed.
> 
> 
>>> Skip 'the' before property, otherwise write it in full 'the attribute'.
>>> Skip 'also', alternatively says that already.
>>> Another idea would be to repeat, that @resource, @href, or @src are
>>> resource attributes (to burn it into the readers mind):
>>> ... in the presence of one the resource attributes @resource, @href, or
>>> @src ... ; and keep this order like it is in section 5.1 for
>>> into-brain-burning.
>> Fixed some.
> 
> 
>>> 8.4, before second example
>>> RDF has a set of predefined predicates that have an agreed-upon semantics
>>> of order.
>>> ... that follow (a) defined/given semantic(s?) of order.
>>> 1. an ... semantics, either a semantic or just semantics, where I think
>>> singular is fine.
>>> 2. agreed-upon, after you agreed upon that, you might have defined it, so
>>> it is given now? Make it easy to translate, please. :-)
>>> 3. Do you really say semantics of order?
> 
>> I don't care touch this - anyone else have an opinion?
> 
> 
>>> 9, last note
>>> entire note
>>> ul
>>> a nice list would make it better readable 
>> Fixed.
> Looks nice. There is the word 'Literal' in the last list item, it 
> might like a small letter? All the other literals have small letters.
> Except:
> Serching the doc for literal I found in the table of contents:
> 8.3.1.2 Typed literals. Here it should be capital, because in the TOC 
> all literals have capital letters.
> 
> 
>>> 10.1.1
>>> 1. RDFa-Vokabular-Entailment considers only the entailment on individuals
>>> 2. not on the relationships that can be deduced on the properties or the
>>> classes themselves
>>> none
>>> 1. Individuals, what is that? Is there any precise RDFish term, that could
>>> be used?
>>> 2. deduce on? Sorry, I don't understand, please help me translate. :-)
>> I can't really comment on this.  Anyone else?
> 
> 
>>> Location in the document
>>> For each IRI being the object of a triple in the output graph with the
>>> subject being the current document (base) IRI and the property being
>>> rdfa:usesVocabulary, that IRI is dereferenced.
>>> If there are one or more triples in the output graph, that contain a
>>> subject being the current document (base), the property (better:
>>> predicate?) rdfa:usesVocabulary, and an object being an IRI, the IRI of
>>> each triple is dereferenced.
>>> From my point of view it is written a bit confusing, from the context it
>>> is clear, but the structure of the sentence doesn't support, that the
>>> subject and property are 'in the triple'. In the suggestion, the flow of
>>> the sentence brings up a triple, inside a subject, property (what could be
>>> also referred to as predicate to keep the spec in one voice), and object.
>>> Plain structure.
>> I agree that this sentence is hard to parse.  I have taken a shot at
>> restructuring it.
> 
> 
>>> 10.1
>>> Note that if, in the second step, a particular vocabulary is serialized in
>>> RDFa, that particular graph is not expected to undergo any vocabulary
>>> expansion on its own.
>>> 1. graph replace with vocabulary graph
>>> 2. see comment
>>> How can a graph undergo something on its own, a graph is a file, not a
>>> processor, is there any detailed description with precise words for this
>>> whole sentence?
>> A graph is a concept - not a file.
> 
> 
>>> Throughout the doc
>>> production vs. definition
>>> definition
>>> In some parts of the document, production is used for definitions:
>>> 'term ::= NCNameStartChar termChar*'
>>> In 7.4.3 'definition' is used in a note for the same 'thing'. Is there
>>> anything, that would force the use of the term 'production' for this?
>>> Production is not nice to translate, and if it is in deed a definition, it
>>> would be nice to just call it definition. I'd like to see that in an
>>> editors guideline for the use of terms in all W3C specs. :-) Is there
>>> anything like that?
>> 'Production' is a term of art.  It is used in many W3C specs.  We really
>> should use it consistently.  I changed 'definition' to production in one
>> place.  There is no other suitable term that I know of.  Sorry.
> Well, well, I shouldn't have said anything, now I have to translate 
> one more production. Terms Of Art. Hm. Fine.
> 
> 
>>> Throughout the doc
>>> entailment
>>> no idea
>>> Entailment is a pretty tough word, it can be used in different ways, the
>>> translation into German is a mess. Is there any other nice english word,
>>> that would do, is more precise and is expected to have a nice translation?
>> Uggh.  This is something from the Semantic Web community, and I can't speak
>> to it. I don't think it has a translation that makes any sense  I would
>> honestly just use the word 'Entailment'.
> Yes, I did use Entailment in my translation, I will keep it then and 
> put a comment to explain it. 
> But, if anyone comes up with some nicer Term Of Art for this, I'll be 
> very happy.
> 
> 
> 
> -- 
> Stefan Schumacher
> Lonavala, Maharashtra, India
> +91 9923670737
> 
>