Color commentary follows:

On Monday 04 February 2008 11:17, Richard Spindler wrote:
> Hi,
>
> We had a little discussion about UI stuff in Cin3, some points raised
> were rather interesting, so I'd like to put a little attention on
> those. :-)
>
> The first Part of the Mail is a little ranting, so scroll down, if you
> are not into flamebaits. ;-)
>
> Some of the suggestions of different People were into totally opposite
> directions, like for example:
>
> Stefan de Konink <[EMAIL PROTECTED]> says:
> > If in your perception we are making an 'office like' application that
> > edits video, I might be wrong. But we probably can stick with what
> > already exists. I am hoping we will design a GUI that is thought
> > 'out-of-the-box'. As a toolkit and not some application, something you
> > will focus on 100% when you use it.
>
> Artem Baguinski <[EMAIL PROTECTED]> says:
> > A "Serious" reason to reuse widgets wouldn't be lazyness of
> > programmers, but satisfaction of users. The less skills user has to
> > forget each time s?he starts another application - the better.
> >
> > What you really want is to make sure the same things are done the same
> > way, similar things are done similar way, and things that are unique
> > to your application are kept to minimum.
>
> I am with Artem in this case, because I don't think we should try to
> outsmart everyone else by making something totally crazy and
> unconventional. I think we are going for a solid, stable and reliable
> tool, and I don't think that this is the place to get sidetracked by
> experimenting with to many new ideas.
>
> I am open to apply novel and innovative interface concepts in isolated
> well defined use-cases. But for the overall application and Interface, I
> strongly vote for and suggest to go the way of boring, well proven and
> familiar concepts.
>
> Other than that, and this is just my opinion and observation. I do not
> think that Stefan de Koninks Comments are moving us forward very much.
> Stefan, while I value your willingness to discuss, and to contribute
> your thoughts and Ideas, I also hope that I can point out some flaws in
> your arguments, not to offend you, but to encourage you to make even
> better suggestions. :-)
>
> This is going to be a serious App, and somehow compromises are
> necessary. I am not saying that there is no room for freewheeling and
> unconventional ideas, but lets assume that we do have some limitations
> both in the resources that we have and in the problems that we want to
> solve. Try to think inside that box. ;-)
>

I agree with you that the "stock" UI should be as you describe - nothing 
fancy, but functional.  That said, if the "core" is designed carefully, the 
flexibility will be there to allow someone like Stefan to go "out-of-the-box" 
with a custom front-end.  Might even want to leave some capability for some 
scriptable CLI interface for bulk processing.

> Stefan de Konink <[EMAIL PROTECTED]> says:
> > I'm sorry developer time is more expensive? Where do you come from? We
> > are not starting a NewCinelerra company and probably want to have the
> > best (not the optimal) solution for the system.
>
> Yes, developer time is actually expensive, why? Because we only have
> very little of it. I guess most of us are hobbiests, working in our free
> time. So better use that time efficiently, and not waste it.
>
> Sure we do not have any deadlines, we can take as much time as we wont
> to, but if it takes 10 years to write a better cinelerra, then I don't
> think this is going to make anyone here happy. Think about it, if we can
> come to a resonable implementation in 1 Year, we have 9 more years to
> play around with crazy innovative concepts. ;-)
>
> The Point is, that we will use an existing well proven toolkit, because
> it will save us Time and Headache. There is no serious alternative. BUT:
> This does not mean that there is no room for fancy stuff. We just have
> to deal with it that all the basics are standard of the shelf Widgets,
> Like: Buttons, TreeViews, TextEdit Fields, File Open Selectors, and so
> on.... While it might look like trivial stuff, creating all those basic
> things will take time, and if you want to do it right, it will take a
> long time! So better don't do it ourselves.
>
> So I repeat myself:
> GTK as well is Qt are well up to the task provide a
> Widgetset for a Video Editor. Period. So don't even think about
> recommending your personal favorite niche Toolkit.
>
> ===========================================================

Again, I think this should be designed so that there is as little interaction 
with the GUI as is absolutely necessary, and that those interactions be 
through well-defined and thought-out interfaces.  Then you can make a 
front-end using whatever toolkit your heart desires.

Maybe video processing is something that might be too cumbersome to effect a 
reasonable separation between the GUI and "core", but I think it's something 
that should have a serious look.
> ===========================================================
>
> So after all that Ranting, I'll move on to less Fluff, more Stuff. ;-)
>
> Ichthyostega <[EMAIL PROTECTED]> says:
> > When working with the current "state of the art" in any real world
> > project, I constantly have *too much* tracks, *too much* effect
> > windows open, *too much* sliders, *too much* knobs, I need to scroll
> > around all the time, poke through endless tabs/lists/bins, while
> > actually working at any time with only a limited focus set of
> > entities. The problem being, this focus is not constant, it doesn't
> > fit with the systematic hierarchy and is not known in advance.
> > We should try to improve things here!
>
> This of course is a problem that is very disconnected from the "Core"
> Video Editing Problem, and therefore an ideal candidate to develop
> individually. But it is also a very difficult Problem, because it is
> what I like to call a "Workflow" Problem. The Problem varies widely
> between people, because everyone has a different Workflow, some people
> might shoot lots of footage for a particular project and want to
> organize that information, others already have a large pool of stuff
> that they tend to reuse again and again in a remixing fashion, others
> might have lots of takes of the same scene and want to single out the
> best shot, etc.... There are plenty of use cases that I can think of
> here, and all of them are different.
>
> This is also a problem that is not so technical but much more about
> Perception, and Information Categorization and Organisation.
> Implementation will probably not too difficult, so whoever wants to take
> up on that task should be willing to do lots of Screen Mockups and
> Detailed Workflow Examples and Explainations. ;-)
>
> I suggest to look at the following applications to get some inspirations
> about footage organization:
>
> http://picasa.google.com/
> http://macslow.thepimp.net/?p=125
> http://f-spot.org/
>
> If you have more links, please add them. If you are interested to work
> on that problem, send in your proposals. :-)
>
> Cheers,
> and have fun,
> -Richard
>

As you may have guessed, I actually envision several front-ends to a (whatever 
it's actually called) Cin3 back-end core.  These front-ends would be tailored 
to the different workflows that exist out there in the wild.  Start simple, 
but allow the flexibility to adapt the GUI interace(s)/plugins later.  But 
have a rock-solid core to it.

Anyone have any thoughts on the utilization of Cg and/or SIMD in the 
processing functions/methods?

My two-cents.

Dave

_______________________________________________
Cinelerra mailing list
[email protected]
https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra

Reply via email to