[topicmapmail] BeerTM

[Lars Marius Garshol]

* Martin Bryan
| 
| You might like to split retailer down into "distributor" and
| "outlet" for two reasons:
| 
| 1) Many beers are distributed outside the immediate area of the
| brewery through one of the standard distribution chains to outlets
| that range from shops (true retailers direct to end-customers) to
| tied pubs (who sell the beer on behalf of a different brewery: it is
| the second brewery that is the customer of the first, but does not
| sell to end-customers directly.)
| 
| 2) Beer festivals are not reatailers directly. They take money at
| the gate for which you get so much free beer, without directly
| buying the beer.

That's a good point. I've updated the page accordingly. I think
"distributor" and "outlet" are subclasses of "retailer", so I propose
modelling it that way.
 
| Also notice that license type will need to be scoped by country.
| Even Scottish rules cannot be applied in England and Wales, so you
| need 3-digit country codes for the scopes :-)

Right. :-)

Seriously, we can just reuse the OASIS GeoLang PSIs for countries, so
that one is easy. Maybe we should put in an explicit statement to say
that the geographical aspects of the ontology will connect to the
GeoLang PSIs rather than reuse them.
 
| Its a pity that the interesting facets, such as conditioning and
| whether or not the beer needs to be pressurized prior to or during
| distribution were not included. 

We were consuming beer as we brainstormed, so our productivity
descreased after a while. That's the main reason for the omissions,
basically. 

| A key factor is length of time in barrel. Your draft beer v lager
| beer is inadequate. You need to be able to distiguish stouts,
| porters, ales, beers, pale ales, lagers and alcohol free products at
| the very least.  

Oh, I completely agree. Under that the whole "beer style" concept is
hiding, I think. The question is whether to have a beer-style topic
type with instances like pilsner, stout, imperial stout, trappist, and
so on, or whether to create a class hierarchy under beer-product where
these are the possible subclasses. What is your opinion?

| Germans will also want to distinguish true lagers (brewed using the
| rules of the German legislation) from "foreign" lagers, and most
| countries distinguish between native beers and "imported" beers.

Hmmm. The question is how to model it. I guess if you represent the
ingredients directly you can infer from that whether it follows the
Reinheitsgebot or not. As for the other, I guess we could have a
country-of-origin association, both for beer-products and for
beer-styles. 
 
| Much thought and testing of the ontology is needed. I suggest field
| trials to which all members of this discussion group are invited.
| I'm happy to host the first such meeting here in the UK :-)

I think this is an excellent idea. I suggest we have our next meeting
at the London conference in May. There's actually a group of topic map
people already who would be looking for interesting beer during the
conference, so I think this makes good sense. Obvious members are Geir
Ove, Steve P., Naito-san, and possibly also Marc de Graauw. Peter
Flynn might also join, but I don't think he will attend the conference
itself. 

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >