Re: Opcodes and bfd & plugins.

["nshmyrev" <nshmyrev at yandex dot ru>]

>A plugable interface is inevitably slightly slower, so most people
>would not want to use it.  There are very few people who would benefit
>from a plugable interface--basically just people in your position.

I don't think that dissasembler should take much time to load shared
library.

>Hmmm, good point.  Of course I can't imagine who would buy a processor
>for which they didn't get documentation, and I can't imagine who would
>want to use the GNU tools if they didn't get source.  But you're right
>that it might be a bad idea to even let some vendor start down that
>path.

I know, that there will be problems with proprietary software, but you 
can't avoid them. May be, plugins in scripting language can be solution
 in this case.

But, plugins will solve another development problem -- one can concentrate
on developing library, instead of removing obsolete configuration. Look at
this mailing list, obsolete wars are main discussion subject. With plugins 
there will be no problem with new architecture or obsolete -- one can just 
keep code of plugin or write a new plugin for new architecture. if you afraid, 
that there will be no free plugins, I am sure they will be available.