[Zope3-dev] mild ramble: the object hub and site definition by services

[Gary Poster]

Casey Duncan wrote:
> On Tuesday 17 December 2002 04:29 pm, Gary Poster wrote:
...
>>It is worth noting, I think, and by the way, that just because the 
>>object hub may need to be unified for *all* parts of a site, indexing 
>>clients of the object hub can still only index subsections of the site.
> 
> Yup. Perhaps this is good enough, at least for now. And perhaps we could 
> create the illusion of multiple hubs, but in reality just keep everything in 
> one big honking BTree. Maybe the "child" hubs could actually be indexes on 
> the root hub.

One big honking BTree, one ObjectHub to bind them and one to rule them 
all, is my preferred choice.  Nice 'n' simple.  The hubid can be treated 
as a guid, which saves a *lot* of hassle and worries, and makes for 
prettier clients.

And if I were forced at gunpoint to write a "site with occasionally 
unified subsites" application, then the child index pseudo-hubs that you 
describe--really, they are more like object hub filters, only giving a 
subset of the hubid mapping in the top object hub--might be my first 
choice of approach.  It seems reasonable and not too unpleasant to write.

So I guess that's two votes for making the following official, 
documentable design decision:

You Can Treat HubIds Like Guids, At Least Until We Change Our Minds.

This would affect the current event service refactoring discussion too.

Any nay-sayers?

Gary