Re: [Gzz] OrthoCoordsys???, reverting

[Benja Fallenstein]

Tuomas Lukka wrote:

> <me>: 
>  
>
> > Ok, to avoid a worse flamewar, let me vent my feelings a little:

Good point. :-) I took a break partly to relax, which always helps 
against flamewars. :)

> >         Yes, I AM annoyed that I discover days before a release that
> >         an API has been changed incompatibly without regard for the OpenGL
> >         implementation. I'd have much preferred if it had been a branch, or
> >         first a proposal to the mailing list.
> >
> > 	But: I'm not blaming you for the breakage of the OpenGL stuff - 
> > 	quite simply the lack of unit tests makes it impossible to know
> >         that it's broken.
> >    
>
> And of course, part of that is because I feel strongly about the OpenGL stuff,
> because it's kind of "my baby".

Yes, I understand. And this is one reason why I need to get it to run: 
so that I can fix it myself. (We should look into that next week, when 
I'm back behind a DSL line-- and back at school, too, unfortunately.)

I think being able to run it is even more fundamental than being able to 
run the unit tests: I *knew* that it was broken insofar as it didn't 
have a map for interpolation that takes the parent coordsys into 
account-- but there was simply no way to fix this without running the 
client (or breaking even more while doing a "blind fix").

> Based on all this discussion, I'm beginning to think that we may
> not reach a conclusion soon enough for release - 
>
> Benja, would you mind a lot if I asked you to revert the hierarchical
> coordsys change and put it in lava/ or something?

Yes, I would. Please, don't.

Ok, it took me some time to figure this out, but I think the reason I 
resent this so much is purely psychological. For a long time over the 
last year, there was no development at all in Gzz, and on my part, 
despite time constraints, the reason for that was simply that I felt 
that there was no way to make progress with it-- I lost faith. 
Currently, I'm feeling quite good about the progress with the client-- 
the interfaces aren't terribly dirty, most of the very basic 
functionality is in place, it's not slow to the point of being unusable, 
and we're *making progress*. It feels like the point where it's a really 
nice little program is coming closer.

However, that feeling is still fragile. This became very clear to me 
when you said the above today-- I felt like running into a wall, 
stopping all forward movement instantly. Yes, I understand that if we 
put it into lava/ now, we can develop it there until GL can handle it to 
and then put it back... but that doesn't help: having to maintain the 
old stuff seems so heavy a burden that I'm afraid I'd lose faith again, 
and I don't want that now.

The actual approach for hierarchical coordsys we take isn't that 
important (as long as it is able to express the things we currently use 
it for: cell padding, and skewing cell content so that the insertion 
cursor is always visible). I may disagree with you on what the cleanest 
way to implement this is, but as always, you're making the final 
decision there and it's ok if we disagree. But please, let's not move 
*backward*. I think it should be possible to find *some* way to 
implement this in GL quickly for now... we can always change it later.

[I'll reply to the other mail with technical aspects later.]

- Benja