[Kolab-devel] Kolab Support for the Toltec Connector

[Joon Radley]

Hi Rici,

> > What Bo means is that standard names for folders and then a 
> translation
> > by the client might have been the better solution, because 
> then clients
> > with different language settings could work on the same folders 
> > without having a huge map of all translations of the folder names.
> 
> Absolutely. 100% in agreement.
> 
> As someone who has administered a variety of multilingual mail
> systems, I concur with the judgement that the MS solution is
> moronic.

:-( <*bows head in shame, moron, moron, moron*>
 
> Consider the case where you are bulk creating new mailboxes.
> You don't know what language the user will want to use. Way easier
> to create everything with a standard name, and count on the client
> to translate the standard names to something which makes sense
> for the user.
> 
> The problem is that you have to standardise the names, and there
> is no convention for this. Suppose the client (mentioning no names,
> but a moronic client does this) automatically creates folders 
> if it doesn't find the one it needs. Then, even though you have
> correctly created a sent folder, the client automatically creates
> a folder called, say, "enviados". Now any server-side scripting
> you might want to do is going to fail. It's a right PITA.
> 
> I would personally favour putting translation entries in the
> folder itself, sort of like ISO9660 translation. There could,
> for example, be a folder called something like !XLATE, which
> contained individual messages with headers X-Standard-Name and
> X-Presented-Name. If everyone standardised on a system like
> this -- which is dreaming in technicolour -- then I could
> even see correct folder names on a borrowed computer.

I would love to see a standard like this myself.

But what about sub folders? When a user creates a sub folder with a arbitrary name, how will this be translated? Babel Fish or will users only be allowed to create sub folders from a list of approved folder names?

Until such a standard becomes reality each client is going to implement its own solution. Take the model put forward by Bo. Each folder must be manually mapped to folder type and translated. This must be done for each install and reinstall.

This create a few problem for me. 

1)If I have to tell a system administrator that if a virus strikes his organization he will have to reinstall, map folder types and rename folders for 2000 clients, he will not buy my product. ( Actually he will die laughing. )

2)I have to adhere to the limitation set to me by Outlook. I need to now what type a folder is when creating it. Normal users will not know how to map this folder and so they will phone the administrator. ( As an administrator you can image handling 3000 calls like this. Been there, done that. )

We can argue about the best/ideal model to do this. This, however, does not bring me closer to supporting the Kolab clients.

Regards

Joon "Flame Proof" Radley