new draft of <infrastructure> 0.94 [message #13] |
Wed, 20 August 2003 16:26 |
Matthias Hengartner
Messages: 57 Registered: August 2003
|
Member |
|
|
Hello,
on http://www.theband.ch/matthias/railml/infrastructure_V094_04 .zip (user:
railml, pwd: line) you can find my new draft of the infrastructure schema.
The changes can be found
in the <startPoint>/<endPoint> elements: The x- and y- coordinates are no
longer attributes of these elements, but encapsulated in a child element
<position> (analogous to <lineElement> and <stationVis>).
Greetings from Switzerland
Matthias Hengartner
---
Noch eine Frage v.a. an Nikolaus & Ulrich:
Mein Schema basiert immernoch auf dem Entwurf von Ulrich. Dani hat mir
Nikolaus' Arbeit inkl. Schema weitergeleitet. Sehe ich das richtig, dass
Eure beiden Schemen noch zusammengeführt werden? D.h. ich muss damit
rechnen, dass grundsätzlich eher die Struktur von Ulrichs Schema gültig sein
wird und dazu die neuen lineElements aus dem Schema von Nikolaus kommen?
Im Übrigen ist es schön, einmal eine ausführliche Doku in den Händen zu
halten, so erübrigt sich einige Fragerei. Trotzdem noch eine kleine Fage:
Wann benutzt man eine <connection> (anstatt eine <switch>)? Ist dieses
LinienElement für den Anfang oder das Ende eines Tracks vorgesehen (ohne
Weiche oder Kreuzung)?
|
|
|
|
Re: new draft of <infrastructure> 0.94 [message #16 is a reply to message #14] |
Tue, 02 September 2003 06:32 |
Vasco Paul Krauss
Messages: 27 Registered: July 2003
|
Junior Member |
|
|
Liebe Kollegen,
die untenstehenden Nachricht sende ich im Auftrag von Herrn Thomas
Kauer (SBB, Schweiz), der wegen Firewallproblemen nicht direkt in
Forum schreiben kann.
<english>
I post this message on behalf of Mr. Thomas Kauer (SBB, Switzerland).
</english>
On Thu, 21 Aug 2003 13:17:11 +0200, "Matthias Hengartner"
<mhengart(at)studentethzch> wrote:
> I have an additional comment on my proposed schema:
>
> I've added an ID-attribute to all lineElements for the unique identification
> (which is necessary e.g. for visualisation) of a lineElement within one
> track. So we can identify a lineElement globally using the combination of
> lineId, trackId and ID.
>
> Alternatively we could also define that the ID of any lineElement has to be
> globally unique.
>
> I think that the former idea of identifying lineElements by
> lineId-trackId-pos has the disadvantage, that there can't be 2 or more
> lineElements on exactly the same position (or is the occurrence of this
> situation not feasible?). That's why I added the ID-attribute.
We are also discussing the problem of a global id.
In our actual systems we know the problem of different elements at the
same
position and had to introduce an artificial offset - so the postion is
not
always helpfull.
Furthermore we have the goal to make the difference between different
states of an element so that we can manage the whole lifecycle:
- first introduction of an element within studies (many parallel
versions
for the same line!)
- refinement (of one or more versions) for concrete projects
- final construction plans
- in construction
- build (not yet in use - testing)
- in use
- deconstruction
- history
So to identify an element we should be able to address the state we
are
looking for. Each state should have a time of validity.
An other problem when using the position of an element for it's ID:
During it's lifecycle the postion of an element can be changed
(normally by
a small amount) without the need of changing it's ID or usage.
Best regards
Thomas Kauer
|
|
|