RE: [xsl] Re: Re: Network diagram - node set intersection

Subject: RE: [xsl] Re: Re: Network diagram - node set intersection
From: cknell@xxxxxxxxxx
Date: Mon, 23 Feb 2004 15:50:02 -0500
I've looked over the DTD, I hope I won't get slapped on the wrist for being off-topic, but I think the answer to one more question will settle the matter of graphML for me. It appears that an edge can have one source and one target. If true, that means that I can't use it. 

What appears to be ambiguous is, I suspect, merely the fact that in the model I am representing a node can have more than one predecessor node and more than one successor node. The nodes represent tasks and the edges represent a temporal relationship between them. The temporal relationships are Start-Start, Finish-Start, and Finish-Finish. 

This means that a task may have a Finish-Start relationship to more than one successor task. That is to say in my original diagram, Task B and Task C cannot begin until Task A is complete, but there is no temporal relationship between Task B and Task C. Similarly, Task D cannot start until both Task B and Task C are complete.

My review of the DTD seems to preclude using graphML to represent such a model. Do I understand this correctly?
Charles Knell
cknell@xxxxxxxxxx - email

-----Original Message-----
From:     Dimitre Novatchev <dnovatchev@xxxxxxxxx>
Sent:     Mon, 23 Feb 2004 21:37:17 +0100
To:       xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject:  [xsl] Re: Re: Network diagram - node set intersection

<cknell@xxxxxxxxxx> wrote in message
> Thank you for your helpful and speedy reply. I understood two of the three
variables you defined right away, but this puzzled me for a while.
> <xsl:variable name="sibling-cnt" select="count($other-predecessor-nodes[.
=$this-predecessor-nodes])" />
> Is the select expression saying, "Count all the nodes in
$other-predecessor-nodes where ..., no, I'm still puzzled. Can you please
put this expression in English so I can get a handle on it?

Count all nodes from $other-predecessor-nodes that have a value, equal to
one of the values of $this-predecessor-nodes

> As to your remark on the topic of graphML, are you comparing it favorably
to my ASCII art representation of the graph that got mangled at the top of
the post or to the XML sample I showed?

Of course I cannot compare flies to elephants. Your source xml did not
express very well the graph -- the terminology is strange and imprecise, the
format is also ambiguous (A could have been specified to have two successors
instead of one)
> The XML was developed strictly as a represenation of the problem. The
original dialect is one I am developing to represent project management data
and is far more involved than the stand-in I posted. The production of this
network diagram is but one relatively minor use for the XML. In order to use
graphML I would have to transform the relevant portions of my project
management language document to graphML and then transform that to SVG. What
is the benefit in my situation of doing that conversion?

The benefit will express itself when any graph problems have to be solved.
Expressing a graph in a convenient and standard way facilitates
understanding and solving the problem. Not to speak that there may be
already tools (to be honest, I don't know of any...) operating on the
standard format.
> Like most self-taught programmers, I am weak in the area of computer
science. I have a basic idea of nodes and edges, but I wonder if a jump into
graphML would put me in over my head. Could I get a handle on graphML
without a sound foundation in the fundamentals of computer science?

The graphML DTD is available at:


Dimitre Novatchev,
FXSL developer, -- the home of FXSL

 XSL-List info and archive:

Current Thread