[sc34wg3] Possible TMRM issue

[Robert Barta]

On Tue, Mar 21, 2006 at 10:39:13AM +0100, Lars Marius Garshol wrote:
> >They are like fingerprints (I actually use MD5 over the proxy  
> >content here). So the current model differs from what you describe  
> >above: [...]
> 
> Is it not possible that if there is a function label(proxy) we still  
> get the same problem, but that it's now hidden inside the label  
> function?

Good thinking, but no. Once the proxy is 'flattened out' (we only have
in it there labels and primitive values) the 'computation' of the label
is trivial. No complexity here in terms of climbing down something or
following up references.

The only complexity is then the lookup

  I have a label in my hand....what is the proxy for it?

And this is a key lookup (row in table, element with a particular key
in an XML document, key in a hash/dictionary). This is implementation
field, here it depends that you compute this label effectively and
then appropriate the reverse lookup.

So here TMRM has the same complexity as TMDM. Which is not surprising
as every TMDM item can be mapped into one (or more) proxy.

\rho