Hi there.
yes internally the software is aging and we mostly have problems to correct the handling with libraries. And since years no new features. I am using blender plugins as well, but I think the intuitive idea of MakeHuman (which is more or less a morphing system) should stay. We have some things where I was always unhappy.
1) we have a base mesh, just one. I replaced it by the same + wings and fishtail at the end and we (Elv & me) were able to do e.g. fantasy characters, I will not do the error to throw this mesh away, but with a simple folder structure I wanted to add more than one mesh. I could even add my horse or cat ... the principle would be the same. So I want to use it as a base.
Since we have all the morphs this will not change too much. With one exception of course: we use the same targets often ... (check github, same length, same morphs) ... when I tried to make them symmetric it was a lot of work to figure out where I need to work and where I only need to copy the file.
2) I wanted to separate the view-mode similar to blender, in a simple mode one to create the character and one to render it. But of course only with OpenGL shaders and some light settings, perspective and a transparent background for those, who use the system without blender but insert characters in GIMP or other graphical programs.
3) bvh poses can also be a sequence (it works with MakeHuman now), Some people proved it by saving 250 times the same frame, Makehuman always loaded all but of course it only showed first one

.. if an animation would work? Well I tried with kivy (instead of QT, just an idea atm) this weekend and they rotate Blender Suzanne. I hoped I could do it with my gumby .. I was happy that I was able to rotate the head of my character. The problem is, that OpenGL has some limits ... It must be buffered, when more than 64k indices/vertices are there. So this code I will re-use in a way.
4) the file-caches I want to exchange with python builtin sqlite ... so database knowledge would help, it is not an oracle or mariadb/mysql ... we need to keep the project small, otherwise we will have new dependencies. But also for selection of clothes we would have a better method with a database. I made a concept in the developer forum (a part invisible for normal users) how this could be solved. So a user can download an asset and add additional information for his own choice. A tiny bit I started with the tag manager in makeclothes already to have at least some common keywords.
5) bone-constraints I wanted to add. Ricardo: not only knees and elbows. I got my Anissa character in unreal engine (the engine "normally" does not support keep volume) but the lower arm will be extremly thin when she shows the palm of her hands and when she sits down ... well ... buttocks are gone

. So in both cases "distance bones" driven by other bones might help. In unreal engine/metahuman there are some extra bones, but these are not animated by user, these are normally driven by constraints. For the knees I tried a double joint .. so one small bone where the kneecap is. The distance bones also can be used for jiggling btw.
6) and here a few other things. Did you ever notice our characters are a bit cross-eyed? Or how the small cubes for bone-endpoints will behave when we morph the character. Why is our waist-line going up and down and it is hard to create pants for both genders? Why are custom morphs only one-directional? Why can't we add a new nose-slider not to the other ones etc. etc.
Hey Salty, change that to e.g.
- Code: Select all
#!/usr/bin/python3
strict is not known to python. Python always complaints. It complaints so much that people give us a stack trace instead of "cannot open file xyz.obj" ... yes I am also perl programmer but I try to get rid of it, because it mostly assumes things (like empty strings, or values of 0 per definition) which sometimes creates weird things (like empty database entries).
github can be used for that, for programming it is good enough I think.
Jira with epics and stories ...

I try to avoid that in my company where I work. We have like 20 agile project managers, scrum masters.. and ehm ... well 1 and a half programmer where one is in vacation ... and the half one is unfortunately blocked by dailies, sprints, sprint plannings, reviews and retrospectives ... that reason why CASE was invented to make this guy working with goals, since his productivity is so small because his own bad organization

Sorry could not resist to write that. But yes after a brainstorming we should agree on what we want to realize. As you can see, my ideas were not that different.