Re: prefixing interface identifiers (X...)

[Michael Hoennig "(mi)" <mi sun com>]

Hi Torsten,

>   1. it is absolutely no argument for me that OO possible has to rename
>      their existing interfaces without a X-prefixed Bonobo. OpenOffice
>      API has to change anyway if it will be founded on Bonobo. And if it
>      will be backward compatible, we will have to provide a wrapper
>      layer anyway.
>      Again, the pain of huge changes in OpenOffice code base was _not_
>      my studlycaps argument, too.

I disagree. First of all, we cannot really make a wrapper. It would blow 
up the size of OpenOffice way too much - at least if it really affects 
ALL interfaces; not so, if only a few interfaces are affected. Second we 
do not really have to change large parts of the API - despite the naming 
conventions and scope. There are actually three areas of APIs:

1. basic component UI-less handling stuff
	=> here both APIs do heavily clash
2. UI-bases component handling stuff
	=> here OpenOffice has little, Bonobo has much
3. special office suite stuff
	=> here Bonobo has little, OpenOffice has much

This means, with OpenOffice naming conventions, there are only limited 
parts to change - though unfortunately nearly all objects implementat 
those basic interfaces.

>   3. I think the decision to prefix interfaces in OO was made only to
>      have a obvious distinction between the meta-types of type names.
>      But _then_, why aren't all other type names prefixed, too? (like
>      'enum YMyEnumType {...};' and struct ZMyStructType {...};') Or if
>      you want to distinguish interface names and service names, why
>      don't prefix services?

This actually is a good question. Especially between structs or services 
and enums or constant groups we are actually having naming problems once 
in a while.

	Michael