Re: Updated proposal for Issue 302

 Gudge, others,
 I didn't spot it the first time, but it seems that mandating xsi:nil on
array accessors does not make your amendment text below work. The
problem is that in arrays you cannot say that there is a missing edge in
the array because if it's missing, it just isn't there so the following
edges' positions change and you've got a different array.
 I think the smaller of the evils we are choosing from (I agree it's not
broken enough to hold us now) is the editor's copy version with
inbound-only and outbound-only edges. I think it will be understood,
although a different approach would be (IMO) considerably easier on the
 So, I don't strongly disagree with the notions of inbound-only and
outbound-only edges, I strongly disagree with the amendment text below
because it breaks on arrays and it implies a schema.
 I'm sorry I'm so lagging behind in the discussions - I'm trying to
 Best regards

                   Jacek Kopecky

                   Senior Architect, Systinet Corporation

On Mon, 2002-09-16 at 13:53, Martin Gudgin wrote:
> After the e-mail thread resulting from my initial proposal[1] for
> closing Issue 302[2] I thought it might be useful to update the proposal
> based on the feedback. Here goes;
> >From the ensuing discussion it seems that people are happy with;
> 	2. Amend clause 4 of[5] to read:
> 	Certain graphs may sometimes contain a given edge and at other
> 	times that edge will be missing. Such missing edges can either
> be
> 	omitted from the 	serialization or can be encoded as an
> element
> 	information item with an xsi:nil attribute information item
> whose value
> 	is "true". 
> The above effectively covers ( or makes unnecessary ) the 'outbound
> only' case
> We still need to tackle the 'inbound only' case. Noah is correct in[3]
> when he draws a graph with an inbound only edge. This is one way to tell
> the story and one I am perfectly happy with.
> The other way would be to state that encodings start with a node rather
> than an edge. This would require us to modify the language concerning
> EIIs and whether they represent edges, nodes or both. Currently EIIs
> always represent an edge and may also represent a node. We would need to
> add a case where an EII represented just a node. This would mean
> re-writing 3.1.1[4] and at the moment I'm struggling to come up with a
> reasonable and concise formulation. 
> Thus unless anyone STRONGLY objects to the notion of an inbound only
> edge, I suggest we leave in the text about 'An edge MAY have only a
> terminating graph node, that is be inbound only.' in the current
> edcopy[6]. Remove the text about 'outbound only' edges from [6] and
> amend clause 4 of[5] as described above.
> Gudge
> [1]
> [2]
> [3]
> [4]
> dnodes
> [5]
> [6]

Received on Tuesday, 17 September 2002 03:56:12 UTC