[Openicc] colord 0.1.0 released!

[Chris Murphy]

On Jan 19, 2011, at 6:06 PM, Hal V. Engel wrote:
> 
> 
> I don't really care about the "larger ICC market" in this context.  I am much 
> more concerned about doing this in a way that actually works on our open 
> systems.  We could end up being the leaders here with the "larger ICC market" 
> following us out of the wilderness.

I'm with you on that. The "larger ICC market" issue is that this is where the tools come from to do what we want and those optional tags mean "we aren't going to do it" to the vast majority of the tooler makers.

> Yes that is the basic idea.  The only real difference is where the settings 
> data is stored.  Is it part of the profile (in which case there is no need for 
> a database to be searched) or is it in a separate database?

I am skeptical of using the existing ICC profile format for a workflow that seems to want a relational database. Maybe it could work, but I'd think you'd still need caching and/or indexing.


>  Either approach 
> can be simple and elegant for end users but I think embedding the information 
> in the profiles is also elegant from an implementation point of view since it 
> actually makes things very simple for everyone particularly since we already 
> have most of the pieces to put it all together.  All we really need is to have 
> a well documented architecture so that everyone is working toward a defined end 
> state.

I think the "how we want it to work" should come first, and then "how to make that happen" comes second. Maybe existing work with ICC profile metadata is sufficient, but I'm pretty much over it when it comes to how the ICC profile format forces workflow, rather than the other way around.

Chris Murphy