[Fenfire-dev] Re: spatial view vs. content view

[Benja Fallenstein]

Hi!

On Wed, 23 Mar 2005 08:51:20 +0200 (EET), Matti Katila
<address@hidden> wrote:
> 19:34 < majukati> what's the difference between spatial and content view?
> 19:37 < majukati> content is text, image and spans but not page image
> 
> 21:29 < benja_> majukati: why not page image?!? it should be page image too
> 21:30 < benja_> content is what's "in" the node, spatial is what's "around" 
> the node
> 21:31 < benja_> you can think of it like this: the content view renders
>                 the text, image, etc. that 'represents' the item
> 21:31 < benja_> e.g. the name of the person (where the node is the person) 
> etc.
> 21:31 < benja_> the spatial view renders a context for the node
> 21:32 < benja_> e.g. a calendar with other events that happened around the 
> same time
> 
> No, the idea with spatial view was it to be spatial.

The idea was that it is a *spatial* *context*. Exactly those words.

True, there is also the structural context, i.e., the buoys. But the
spatial view is definitely meant to render a (spatial) context (which
contains the node itself).

It's quite obvious that the page span cannot be a spatial context
view, but must be a content view, btw: Otherwise, you could not place
a page span on a canvas, because the canvas is a spatial context view.

> But when we though
> this spatial vs. structure thing I was not thinking this problem at all.
> 
> I propose that we create ContentView or RepresentView

Hey, we already have ContentView :-)

> and from that
> inherit [Text,Image,Page]ContentView/RepresentView. And for predicate
> lobs use something else than these.

Follow-up in the other mail:

> Well, I don't think it would be nice if there's a huge document
> representing a predicate.

I really don't see the point. Sure, predicates normally have short
labels. But what does this have to do with code? Predicates *are not*
documents -- the classes are disjoint -- so the only way a predicate
could be shown as a huge document is if the user explicitly specified,
"Show this node (the predicate) like this (huge document)."

If a predicate has a short label, e.g.

    predicate rdfs:label "linked to".

then the content view will only show a short label.

> Second thing is that we need to do
> internationalization for predicates and this functionality is easy to do
> in one place.

We need internationalization for other nodes as well and we need it to
work exactly the same way, so this is no argument for using a
different view for predicates.

- Benja