[DAS2] hrefs and xml:base

[Andrew Dalke]

Gregg:
> If we want to fully adopt the xml:base standard (which I think we do),
> then my interpretation is that xml:base is lexically scoped and within
> its scope allows any attribute that requires a URI to be potentially
> specified as a relative URI (relative to the xml:base) instead of an
> absolute URI.

That's what I figured, and what I did in my code.

I'll edit the docs to flag which fields are expanded by xml:base.

Now for part 2 of the question, in the <PROP> tag will
the values ever take relative URLs?  Eg, suppose some
server has an SVG description of how to draw the type

     <PROP key="depict"  value="/svg/tRNAscan.svg" />

In a related topic, the TYPES example we have says

<TYPES
      xmlns="http://www.biodas.org/ns/das/genome/2.00"
      xmlns:xlink="http://www.w3.org/1999/xlink"
      xmlns:bg="http://www.bioperl.org/biographics/properties"
      xmlns:das="http://www.biodas.org/ns/das/genome/2.00"
      xml:base="http://www.wormbase.org/das/genome/volvox/1/type/">
...
     <PROP key="bg:glyph"     value="box" />
     <PROP key="bg:bgcolor"   value="white" />
     <PROP key="bg:fgcolor"   value="black" />
     <PROP key="bg:key"       value="tRNAs" />
     <PROP key="bg:citation"  value="tRNA predictions made using Sean 
Eddy's tRNAscan program." />
     <PROP key="bg:fontcolor" value="slateblue" />
     <PROP key="bg:height"    value="3" />

Is the "bg" in the XML namespace "bg" supposed to be inserted
into the key name as a string?  If so, how?  Using universal
names ( http://www.jclark.com/xml/xmlns.htm ) like
  "{http://www.bioperl.org/biographics/properties}height"

Or is the xmlns:bg definition there in error?

This is a problem for me because the parsers I use treat
XML namespaces as things that only apply to tags and attributes
and throw away the namespace scope.

There aren't that many people using DAS nor will there
be.  We could just say that certain prefixes are reserved
and that people may pick a unused namespace prefix as desired.


					Andrew
					dalke at dalkescientific.com