|From lachance@chass.utoronto.ca Fri Sep 21 15:54:36 2001
|Date: Thu, 20 Sep 2001 19:39:51 -0400 (EDT)
|From: Francois Lachance <lachance@chass.utoronto.ca>
|To: lou.burnard@computing-services.oxford.ac.uk
|Subject: P4 Chapter 16
|
|Review of Chapter 16 Feature Structures
|
|16.1
|space missing paragraph one reads "chapter26"

Fixed

|
|16.2
|I would suggest harmonizing the descriptions for the "rel" attribute to disambiguate "given content", "actual content" and "actual values". Might be accomplished with a reference to 16.9 Indirect Specification of Values Using the rel Attribute. 

I don't understand your comment. The descriptions are pretty similar,
I think. There is a cross reference to 16.9 a little further in the
text, where it says that the attributes are not being discussed here.

|
|<fs>
|rel 
|indicates the relation of the given content to the actual content or value of the feature structure. [Reference to 16.9]
|
|<f>
|rel
|indicates the relation between the values that are given as the content of the feature or pointed at by the fVal attribute and the actual values of the feature.  [Reference to 16.9]

In general, we dont put cross references  inside attribute
descriptions. 

|
|16.2 Formal Declaration
|
|<!ELEMENT %n.fs; %om.RR; ((f | fAlt | alt)*)> 
|
|Are the parentheses surrounding the inner group (f | fAlt | alt)* necessary?


Yes, though the outer ones are not!


|
|16.3 
|Case sensitivity
|id="TXT1"  but IDREF=txt1  from <linkGrp> to <text>
|id="FSL1"  but IDREF= fsl1 from <linkGrp> to <fsLib>

This is a recurrent problem throughout the chapter: I am making
them all uppercase

|

|Also the <link> elements target <c> elements with id values with
| upper case letters but the idref values use lower case.

As above.

|
|The <link> id values are not unique. "LS" and "LZ" are repeated.

Most bizarre! I've removed the duplicates

|
|Order of discussion
|I suggest moving the discussion of the binary feature values to before the discussion of the use of the "feats" attribute of <fs>. That is, present fLib, fvLib and fsLib before the linking mechanism example. 
|


|16.4
|Suggest striking the first sentence. 
|"By separating out feature values as content of <f> elements, we are able to classify those values into types."
|
|Suggest adding explicit mention of <plus> and <minus> in the next sentence.
|"16.2 Elementary Feature Structures: Features with Binary Values , the two empty elements which represent binary values are defined."
|
|Suggest removing comments <!-- ... --> used to indicate ellipsis. Or having them appear on a separate line. 
|
|Missing quotation marks
|<f name=case>
|should read
|<f name="case">
|
|The example <f name="number of bedrooms" rel="eq"> needs a wrapper to make it a conforming XML snippet.
|
|Missing quotation marks
| <f name=monthly.rent>
|should read
| <f name="monthly.rent">
| 
|To help emphasize XML requires attribute values to enclosed by delimiters and for smoothness suggest rewording
|"with a per=month attribute"
|to 
|"with a per attribute with a value of "month",
|
|16.5
|
|Missing quotation marks 
|id=Nkab027
|id=txaustin
|
|Case mismatch
|The paragraph asks readers to suppose a value of "Nkab0927" as the value of an id attribute. The the value of the idref in the example reads 
|fVal=nkab027
|(quotation marks missing)
|suggest changing "Nkab097" to "nkab0927" since lower case is used throughout all the other examples.
|
|Suggest adding a cross-reference to 16.3 for the example introducing the use of feats attribute because the fLib is not explicitly demonstrated the ID/IDREF mechanism is not clearly visible. The cross-reference would serve as a useful reminder where to get the info on how feats works.
|
|Case matching
|in the prose:
|Pmbf009   Parn002
|in the example:
|pmfb009  parn002 
|
|In the pdf format a line gets cut
|Mr.</seg> and <seg
|in the html, it fails to wrap
|
|16.6
|The discussion of single, set, bag and list might benefit from the addition of a table presentation :
|
|								order		repetition
|single 			 n/a			n/a
|set						n/a			n/a
|bag						n/a			yes
|lists					yes			yes
|
|Paragraph two:
|Missing the word "chapter" before 26, Feature System Declaration
|
|Paragraph three:
|attribute values change to lowercase to match examples
|PKAB027
|PANB005
|PMFB010
|Pzrb001
|enclose in quotation marks
|
|Next paragraph:
|Missing quotation marks:
|org=list
|name=promotion history
|
|As well, 
|org="list" 
|needs to be added to the example for the element <fs type="employment record" rel="sb"> since the concluding sentence of the paragraph introducing the example mentions it.
|
|
|Missing quotation marks
|name=status.code
|<f name=agreement> (2 places)
|
|"chapter" dropped before 26 Feature System Declaration
|
|Missing quotation marks in the description of <null>
|org=set
|org=bag
|org=list
|org=single
|
|Matching case and missing quotation
|<fs id=PKAB027> [should be lower case]
|
|Missing quotation marks
|<f name=siblings>
|
|16.7
|
|Suggest a bit of rewording for following paragraph:
|
|However, the <f name=number.of.bathrooms> element in this example can be factored out of the alternation, and a
|<vAlt> element used instead to represent the alternation of just the feature values, as follows. 
|
|Suggest shortening to the following:
|
|However a <vAlt> element may be used instead to represent the alternation of just the feature values, as follows. 
|
|Missing quotation marks
|<f name=extras org=set>
|
|In the example of the representation of feminine gender, accusative case, plural number noun form suggest introducing whitespace between the <fsLib> <fLib> and <fvLib> for readability.
|
|Suggest specifing 
|<f name=...> 
|as 
|<f name="case">
|
|Missing dot:
|<alt id="kg.nska.np" targets="kg.ns kanp">
|should read
|<alt id="kg.nska.np" targets="kg.ns ka.np">
|so that the idref "ka.np" matches the id "ka.np"
|
|16.8
|Missing quoation marks:
|<f name=auxiliary>
|
|Suggest eliminating the =/= sign  and the == sign to indicate equivalency between pairs since it suggests the PCDATA is permissible in the place. Suggest splitting the pairs by the explanation. For example.
|
|Given this <valRange>, then the following pair of specifications is distinct.
|<f name="auxiliary" rel="eq">
|   <plus/>
|</f> 
|<f name="auxiliary" rel="eq">
|   <any/>
|</f> 
|In this situation, the <any> element is equivalent to the alternation of the <plus> and <minus>.
|
|Given the same <valRange>, then the following pair of specifications is also distinct.
|<f name="auxiliary" rel="eq">
|   <minus/>
|</f> 
|<f name="auxiliary" rel="eq">
|   <none/>
|</f>
|The <none> element is equivalent to the negation of the alternation of the <plus> and <minus>.  
|
|
|However, if the auxiliary feature is declared to take only the <plus> value, then the specifications below are equivalent.
|<f name="auxiliary" rel="eq">
|   <plus/>
|</f> 
|<f name="auxiliary" rel="eq">
|   <any/>
|</f>
|
|If the auxiliary feature is declared to take only the <plus> value, then the specifications below are not equivalent; in fact, the specification is invalid. 
|<f name="auxiliary" rel="eq">
|   <minus/>
|</f>
|<f name="auxiliary" rel="eq">
|   <none/>
|</f>
|
|-------
|Missing quotation marks
|<f name=impossible>
|
|Suggest removing == and replacing with line break in the last pair of binary and boolean specifications.
|
|Text reads "section" before 26, Feature System Declaration
|should read "chapter"
|
|Missing quotation marks:
|<f name=gender>  (2 places)
|<sym value=feminine>
|
|Suggest changing 
|<f name=...>
|to 
|<f name="gender">
|
|Missing quotation marks:
|<f name=number>
|<f name=category>
|
|Need <fsConstraints> wrapper for <cond> example
|
|Case mismatch:
|verbagreement in the example
|verbAgreement (2 places)in the exmplanatory parapraph following example
|
|Missing quotation marks:
|<f name=adjInflection>
|<f name=verbInflection>
|
|16.9
|Missing quotation marks:
|<f name=case>
|<sym value=genitive>
|
|The examples following the introductory paragraph are identical.
|Should one of them not specify a value for the rel attribute in order to illustrate an interpretation of the default value? i.e. <f name="case">
|
|Suggest insertion of whitespace between the two examples.
|
|16.9.1
|Missing quotation marks:
|f name=case (3 places)
|
|
|16.9.2 
|Suggest insertion of whitespace between the examples (otherwise it may be interpreted as non well-formed XML).
|
|16.9.3
|Suggest insertion of whitespace between the examples.
|
|Missing quotation marks:
|<f name-person>
|<f name=number>
|<fs rel=eq>
|
|Should the following sentence not refer to a relation of inequality i.e. <fs rel="ne"> ?
|
|One situation in which it may be preferable to specify <fs rel=eq> is when
|the feature structure has many optional features and it is known that none of them occur. 
|
|-----
|
|Case & quotation marks
|<fs id=NP3NS> (2 places)
|
|16.9.4
|Suggest insertion of whitespace between the examples.
|
|16.9.5
|Suggest insertion of whitespace between the examples.
|Remove rel="sb" in the second example <fs rel="sb"/> since the preceding prose indicates "In the second example below, the rel attribute has not been specified, since it has the value sb by default on <fs> elements."
|
|16.10
|Suggest rewording introductory sentence from
|
|In this section, we present two illustrations based on one text of how to associate feature structures and their components with textual elements.
|
|to 
|
|In this section, based on one text, we present two illustrations of how to associate feature structures and their components with textual elements.
|
|Suggest changing the type of element contained in  
|<f id="j37iltr> from <str> to <sym>  
|<f id="j37ilr> from <str> to <sym>
|since when one verifies the link to the text one doesn't find a string equivalent to the declared feature. One does find strings indicative of the meaning found in the features.
|
|linkGrp corrections:
|
|Change
|<link targets="j37pc mds09"/>
|to
|<link targets="j37pc mds06"/>
|Retain one of the <link targets="j37pc mds09"/>
|
|Change
|<link targets="j37pc mds11"/>
|to 
|<link targets="j37pc mds12"/>
|
|Change 
|<link targets="j37r mds25"/>
|t0
|<link targets="j37ilr mds25"/>
|
|Missing Quotation marks
|id=mds0905  (2 places)
|id=mds0906  (2 places)
|
|Thanks!!
|
| 
|Francois Lachance, Scholar-at-large
|	 http://www.chass.utoronto.ca/~lachance/ivt.htm
|per Interactivity ad Virtuality via Textuality
|