[sc34wg3] Typo in 5.4.2 Identifying Subjects

Patrick Durusau sc34wg3@isotopicmaps.org
Mon, 17 Nov 2003 08:02:17 -0500


Lars Marius Garshol wrote:
> * Patrick Durusau
> | 
> | Note that in the example, the URI http://www.topicmaps.org is said
> | to be a "subject identifier" for topic B. In the following
> | paragraph, that changes to make it a subject **indicator** for topic
> | B.
> That's true, but notice what the second paragraph says:
>   The information resource in question is a subject **indicator** ...
> So the second paragraph is talking about the information resource,
> while the first is talking about the URI. The first does not really
> make it clear that it is referring to the URI, however, so we'll fix
> that. I think that will be clear.
Yes, it was the shift between saying URI and then saying information 
resource where I got tripped up. (In first paragraph, "subject 
indentifier for B", in the second, "subject indicator for B". I just 
skipped over the URI vs. information resource qualification.)

> | Not sure what it means to say that the "information resource" in
> | question is ambiguous? May be poorly written, confused, etc., but
> | hasn't the distinction (a la Pepper's "Curing the Web's Identity
> | Crisis) already been made by the choice of subject identifier or
> | subject indicator?
> Good question. What we mean is that the information resource is
> ambiguous in the sense that it does not clearly indicate a *single*
> subject, but could reasonably be interpreted as indicating several
> different subjects. So it's not really poorly written or confused, but
> just fails to work well as a subject indicator, for the simple reason
> that it was never meant to be one.

Hmmm, but isn't that interpretation controlled by my use of the subject 
identifier in the topic map? That is to say, if I give that URL as the 
subject identifier for topic B, which leads to a subject indicator, 
haven't I already resolved the issue by saying it indicates the subject 
of topic B?

Ah, but you are saying that if I have made a poor choice in subject 
identifiers that lead to subject indicators (which could indicate more 
than one subject), then the subject of my topic is ambiguous.

Hmmm, not sure that is the case for all information resources, whether 
drafted as subject indicators or not.

Let me try an example that may help illustrate my point:

Topic is Prolog, subject identifier is: 

Now it is true that the subject indicator has a variety of information 
about Prolog but in my topic map I have said this indicates the topic of 

I guess the real issue is whether it will be necessary to recast 
existing information resources in order for them to function as subject 
indicators. Seems like we would want to leverage on as many existing 
information resources for that purpose as possible.

> I think we should explain this in the example, since you're unlikely
> to be the only person wondering about this.
> | Not altogether happy with the "guaranteed to be stable" language but
> | not sure I have a useful alternative to offer. PSI's are "supposed"
> | to be stable but not sure we should indicate that stability should
> | be assumed by users. How to even test for stability dips into
> | implementation issues outside the scope of this document but it is
> | something we will need to think about in PubSubj.
> The thing about "guaranteed to be stable" is only there to make
> readers consider the issue of "if I use page X as an SI for this
> subject, do I really know that it will identify the same subject later
> on?" Whether it's really appropriate to have it there I'm not sure. We
> thought it was. Does anyone think otherwise?
Not inappropriate. Suggest something along the lines of: "Users must 
consider stability, both in terms of the subject indicated and 
availability, when choosing information resources to act as subject 

Alerts readers that for mission critical subject indicators that the 
usual criteria for reliance apply, single point of failure, stability, etc.

May also need to add some language that say to pick information 
resources for subject indicators that are not ambiguous.

Hope you are having a great day!


Patrick Durusau
Director of Research and Development
Society of Biblical Literature
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model

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