[sc34wg3] Review of N0393

27 Apr 2003 12:06:47 +0200

* Jan Algermissen
| 
| Well, in order to advance N0393 it would be usefull if you would
| list as many problems as you see. 

I'm sure it would, but I have other commitments to the SAM, XTM, and
TMQL, first and foremost, and I also try to support the work on HyTM,
CXTM, and TMCL as best I can. For me, the RM has no priority at all.
I did this review because I felt I ought to before we recommended that
it be taken out of 13250 altogether.

| Sure we will try fix the 'many more problems' once you point them
| out.

My main problem is really that N0393 has not made up its mind whether
it wants to be a technology or a set of guidelines.

* Lars Marius Garshol
|
|  - N0393 describes a particular model, but does not make it clear
|    whether, or how, that model enables any form of interoperability.

I explained that in other recent postings to this list.

| How does N0396 achieve interoperability?

That should be clear, I think.

* Lars Marius Garshol
|
| - It is not clear that the notion of the conformance of a Syntax
|   Deserialization Definition is useful. What does it mean for an SDD
|   to conform to this document? What value does it add for SDDs to
|   conform? 

* Jan Algermissen
|
| Suppose SDDs would be defined in an XML syntax (propably looking
| close to XSLT) then generic syntax processors could be implemented
| that would be driven by SDD documents. In fact, syntax processors
| could fetch SDDs from registries over the Web based on the document
| type they are to process (e.g. XTM, NewsML, NITF etc.).
| 
| Is that a benefit?

It certainly is, but doing it on the RM would be very painful, and it
would require the RM to be tightened up considerably. Meanwhile,
Ontopia has already done based on the SAM. So to us there does not
appear to be any benefit whatever to do this with the RM. We don't
need the RM to do this, and it does not seem very well suited to the
purpose. 

* Lars Marius Garshol
|
| What is to be done about SDDs that do not conform?

* Jan Algermissen
|
| You mean: N0393 does not say whether or not to report an error?

N0393 does not even say properly what an SDD *is*, and what an error
would be and who would report it is something you can't really form an
opinion on on the basis of N0393.

| You are right then, N0393 should say to report an error.

I don't think that makes any sense.

| Why is it insufficiently defined? In fact, exactly this I think is very
| clearly defined. What is missing from your POV?
| 
| [Suppose for a minute you were to implement N0393, what information
| is missing?]

Some of the problems were pointed out in more detail further down, but
you haven't quoted those.

| See my question on the differences between N0396 and N0393. What of
| N0393 is any more complex than it needs to be.

Section 4.2.1 and the rest of section 4 is the prime culprit at the
moment.

| In fact, I don't think you can get less complex than N0393.

In that case there's not much we can say to one another.

| In my own words:
| "A topic map consists of topics that surrogate subjects. Properties on
| the topics are used to define what particular subject a given topic
| surrogates." 

And what about assertions? Are they expressed with properties, or are
they not? That's just about the most basic question about the RM you
can imagine and N0393 does not even answer that. The whole thing is so
underspecified that it is difficult to take it seriously. I'm sorry to
use so strong language, but a more neutral phrase would not convey the
extent of my disappointment with N0393.

| <humor>
| So, does that imply that once the 'underspecification' you mention is
| fixed and thus the sand turned into a solid rock, do we then have
| a TMM that is an impregnable fortress, built for eternity?
| </humor>

I very much doubt it.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >