[usmt dev] refresh my memory...

Jason Pyeron dev@smime.pyerotechnics.com
Wed, 14 Feb 2001 18:08:25 -0500 (EST)


<Addr, Type, Reserved>

where Addr=valid address
where Type=key type
where Reserved=NULL

rational:

obviously addr/type are needed
but there might be a case where a third dimension on the index is needed.

for example

i use jason@pyeron.com for my outgoing mail both at work and for personal
mail.

but i have roles enabled in pine.

I can have my work role (9-5) sign messages with my companies sig yet in
the pm i can have it sign with my personal sig.

it would be useful if one could index on <jason@pyeron.com, cert, work> vs
<jason@pyeron.com, cert, personal>


btw, if this were the case i would include a header called role which
would look like

x-cert-role: work 

or 

x-cert-role: personal

-jp

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
-                                                     -
- Jason Pyeron              Pyerotechnics Development -
- Projects Manager          Suite 101/2937 St Paul St -
- 410 467 2266              Baltimore, Maryland 21218 -
-                                                     -
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


On Tue, 13 Feb 2001, Mike Schiraldi wrote:

> I'm starting to wonder why we don't just make <Addr, Type> our primary
> key. It seems that there would only be a small number of cases where a
> person would have more than one object of a given type. Usually when that
> happens (when a person "wears two different hats") they use a seperate
> email address for each one.
> 
> Expanding past <Addr, Type> opens up a Pandora's box. If we make the key
> <Addr, Type, Foo> to support people who want to have records with the same
> Addr and Type, what happens when someone wants to support different
> records with the same Addr, Type, and Foo?
> 
> So unless someone can demonstrate that the extra functionality of allowing
> multiple records to have the same Addr and Type is worth making things a
> lot more complicated, i'mgoing to dump that support.
> 
>