Proofing comments: was Re: [sc34wg3] TMDM FDIS draft

Patrick Durusau sc34wg3@isotopicmaps.org
Sun, 30 Oct 2005 16:59:40 -0500


Lars,

Some comments on the latest TMDM draft.

Aside to the SC34WG3 list: Despite the length of my comments, realize 
that I think the latest draft is very strong and my comments are an 
attempt to help it say what it wants to say as precisely as possible.

1 Scope, second paragraph, last sentence:

Reads: "All of these standards fall outside the scope of this document, 
however."

Comment: The "..., however" is unnecessary. Simply state the fact and 
move along.

Suggest: "All of these standards fall outside of this document."

2. Normative references

Reads: Omission of XTM

Comment: Noted the omission of "XTM" despite its appearance in 5.1 
General, first note, second paragraph, second sentence. For consistency 
I think it should either be a normative reference or included in the 
bibliography. Probably more a normative reference than a bibliographic 
item.

3. Terms and definitions

Comment: Note that sort name is omitted from the listing, but is the 
head of section 7.4 Sort names

Suggest:

sort name

a form of topic name used to sort topic

(from 7.4 Sort name, first paragraph, first sentence)

3.11 reification

Reads: "making a topic represent the subject of another topic map 
construct in the same topic map"

Comment: Although the definition of topic map construct includes "topic 
map," which satisfies the need to reify a topic map, it seems odd that 
reification says "another topic map construct in the same topic map." 
Reasoning that a topic map doesn't really "contain" itself.

Suggest: "making a topic represent the subject of another topic map 
construct in the same topic map or the topic map itself"

Comment: May be unnecessarily fussy as I don't see real harm in the 
present definition but it does raise the question of a topic map 
containing itself. I assume that foreign topic maps are treated as all 
other subjects.

5.2 The topic map item, note, first sentence:

Reads: However, while the topic map does not represent anything it may 
be reified in order to make statements...."

Comment: Uncertain about the language "...the topic map does not 
represent anything..." being followed by statements on how to provide 
traditional infomration for something that "does not represent 
anything..." I think I understand the difficulty, particularly following 
the last sentence of the preceeding paragraph, which reads: "The topic 
map itself has no meaning or significance beyond its use as a container 
for the information about those subjects."

Suggestion: I would drop the sentence: "The topic map itself has no 
meaning or significance beyond its use as a container for the 
information about those subjects."

Suggestion: Amend the "...does not represent anything..." sentence to 
read: "A topic map may be reified in order to make statements about the 
topic map (that is, the collection of topics and associations) as a whole."

Comment: Reasoning that the comments about a topic map having no meaning 
or significance are gratuitous and certainly untrue if the topic map is 
reified.

5.3.2 Identifying Subjects, Example, first paragraph, second sentence:

Reads:

***
Consider the IRI |http://www.iso.org|. If given as the subject locator 
of topic A this would mean that topic A represents the information 
resource identified by this IRI. However, using it as the subject 
identifier of topic B would mean that B represents what is described in 
that information resource. At the time of writing this would seem to be 
the organization known as the International Organization for 
Standardization. (Note: the /organization/; the real-world institution 
known by that name. This is different from the subject of A, which is 
the web page itself.)
***

Comment: Not sure the second sentence is consistent with the definition 
of subject identifier.

subject identifier: locator that refers to a subject indicator

subject indicator: information resource that is referred to from a topic 
map in an attempt to unambiguously identify the subject of a topic to a 
human being

I don't see "...attempt to unambigously identify the subject of a 
topic..." as being the same as " ..B represents what is described in 
that information resource."

It isn't that I don't appreciate the difficulty of illustrating the 
principle of subject identifiers but on the other hand, I think the 
example should be consistent with the definition given in the standard. 
To put it another way, I think within the topic map community we share 
an unvoiced assumption about subject identifiers that makes this 
sentence make sense.

The same criticism applies to the "...it does not unambiguously indicate 
a single subject" in the following paragraph. It really is a matter of 
judgment as to the subject the topic is meant to represent as to whether 
to a particular user a subject identifier is adequate to the task of 
"unambiguously" indicating a single subject.

To be honest I have tried for some time to think of a way to fix the 
example and came up empty. A real example would require more detail that 
I think would be appropriate in such a note. Since examples are meant to 
assist in explaining the standard and this one appears to depart from 
it, I would suggest simply dropping the example.

5.3.3 Topic characteristics, second paragraph, first sentence:

Reads: "The properties of topic items that do not represent topic 
characteristics are not statements about the subject; they are 
statements about the topic."

Comment: Since "topic" is a defined term, the final clause: "...they are 
statements about the topic." seems incorrect or at least confusing.

Suggest: "The properties of topic items that do not represent topic 
characteristics are not statements about the subject; they are 
statements about the topic item."

Comment: Which makes it clear that the statements are about a topic map 
construct.

5.3.4 Scope, note:

Reads: "This is the first Topic Maps standard to define scope as being 
"all subjects", hence older topic maps may not conform to this."

Comment: Granted this is the first Topic Maps standard to define scope 
as being "all subjects" but then the reference to older topic maps 
leaves the reader wondering what other topic map standards had different 
definitions of scope.

See the comment on the undocumented reference to XTM, which could be 
used here to fix this problem.

Suggest: "This is the first Topic Maps standard to define scope as being 
"all subjects." Topic maps written using the [XTM] standard may not 
conform to this definition.

5.3.5 Reification, first paragraph, second sentence:

Reads: "For example, creating a topic that represents the relationship 
an association represents is reification."

Comment: Since this is explanatory material, shouldn't this be in a 
note? (as opposed to the current note, see the following comment on the 
current note)

5.3.5 Reification, first note:

Reads: "Note that the use of the term 'reification' in this document is 
not to be confused with its use in philosophy."

Comment: I am not sure how useful this note will be to readers in light 
of the precise definition that begins this section.

Suggest: drop

5.4 Topic name items, first paragraph, second sentence:

Reads: "It is only the topic name which is a topic characteristic; the 
base name and variant name are only parts of the topic name characteristic."

Comment: Uncertain what "only" adds to "...are only parts of the..."

Suggest: "The topic name is a topic characteristic; the base name and 
variant name are parts of the topic name characteristic."

5.4 Topic name items, first note, last sentence:

Reads: "Essentially, a base name is a specialized kind of occurrence."

Comment: The word "essentially" appears to qualify the statement in some 
unstated fashion. Why not simply make the statement?

Suggest: "A base name is a specialized kind of occurrence."

5.5 Variant items, note, final sentence:

Reads: "This document does not constrain the process by which this is done."

Comment: Granted but if so, need we say anything at all? What is not 
prohibited is permitted sort of reasoning.

Suggestion: drop

5.6 Occurrence items, first paragraph, fourth sentence:

Reads: " Occurrences are essentially a specialized kind of association, 
where one participant in the association shall be an information resource."

Comment: Question of why qualify with "essentially" again.

Suggest: "Occurrences are a specialized kind of association, where one 
participant in the association shall be an information resource."

6.2 Merging topic items

Comment: I puzzled over the absence of the absence of any updating of 
the [roles played] property until I remembered that it is a computed value.

Since the computed properties on topic item are only [roles played], 
[reified], and [parent], will all other instances being [parent] 
properties, I would suggest a note here to cover the updating the [roles 
played] property under 6.2.

Suggest: Note: The [roles played] property is a computed value and 
therefore is not explicitly covered under the procedure for updating 
values when merging two topic items A and B.

7 Core subject identifiers, 7.1 General, second paragraph:

Reads: "All core subject identifiers defined by this document are 
distinct, that is, topics representing these subjects canot be merged 
with one another."

Comment: Yes, the core subject identifiers are distinct, but what's 
important is that the subjects they represent are distinct. That subject 
identifiers are distinct is no barrier to merging  them, provided they 
identify the same subject.

Suggest: "The subjects represented by the core subject identifiers 
defined by this document are distinct, that is, topics representing 
these subjects cannot be merged with one another."

7.2 The type-instance relationship, first paragraph, second sentence:

Reads: "Any subject that belongs to the extension of a particular topic 
type is known as an instance of that topic type."

Comment: I tend to read "...the extension of a particular topic type..." 
to mean "subtype" and not "instance." Not to mention that the sentence 
is mixing topics and subjects. That is that to be an instance of a topic 
type is a characteristic of an information item, not the subject it 
represents.

Suggest: "An information item with a particular topic type is known as 
an instance of that topic type."

7.2 The type-instance relationship, last paragraph (Same comment and 
suggestion applies to 7.3 The supertype-subtype relationship, last 
paragraph before example):

Reads: "Scope applies to this association type just the same way as it 
does to any other."

Comment: What drew my attention was "...as it does to any other." Agree 
that making the application of scope clear is a good idea, question 
whether this reaches that goal.

Suggest: The rules for applying scope to associations are defned in 5.7 
Association Items.

(Both occurrences of this sentence.)

7.4 Sort names, third sentence:

Reads: "Implementations may use other sort orders for datatypes other 
than those defined in this document."

Comment: In other words, sort order for the string datatype is fixed? 
Granted that the variant name item [value] property is not fixed as 
string and so an implementation could choose a sort order on some 
datatype other than string for variant name.

7.4 Sort name, second paragraph:

Reads: " Sort names can be represented by variant items whose [scope] 
property contains a topic item...."

Comment: Is the TMDM allowing the possibility that a sort name could be 
represented by an association? That seem to be the implication of the 
"...can be..." language.

7.5 Subjects for defined terms, 
http://psi.topicmaps.com/iso13250/association

Reads: "Usage: the type of all associations."

Comment: Isn't this going to be confusing with associaton [type] property?

Same problem occurs with: association-role, association-role-type, 
association-type, occurrence, occurrence-type, subject, 
subject-indicator, topic-map, topic-name, topic-name-type, topic-type.

This statement is absent from variant-name.

Since I am not certain what is meant by "the type of all (insert term) 
types" I am not able to suggest a correction.

Speculate what is meant is a type that includes all of the instance of 
each particular type, but I am not sure why we would need a subject for 
those.

Not questioning the need to have representatives for each of these 
subjects, but am questioning the meaning of the "Usage" statements as 
listed.

As I said at the outset, a very strong draft and one that I think, with 
some minor tweaking merits approval as an FDIS draft. (Speaking in my 
personal capacity and not as representative of the US National Body.)

I don't see any of my comments as show-stoppers nor my suggestions as 
necessarily the correct solutions to those comments. I hope these 
comments will be helpful in terms of preparing the final version of the 
FDIS draft.

Hope you are having a great day!

Patrick

Lars Marius Garshol wrote:

>A proposed FDIS draft of the TMDM has just been uploaded to:
>  <URL: http://www.isotopicmaps.org/sam/sam-model/ >
>
>The idea is to spend the weekend proof-reading, and then submit the
>final draft for ballot on Tuesday. Comments very much welcome.
>
>  
>

-- 
Patrick Durusau
Patrick@Durusau.net
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model
Member, Text Encoding Initiative Board of Directors, 2003-2005

Topic Maps: Human, not artificial, intelligence at work!