I have been working on Flash games professionally for two years now and somehow, having our artists producing assets the right way is one of our biggest challenge.
More precisely, it is very hard to have them following any kind of structure and/or standards, nor taking into consideration performance. I would say also the most of our issues concerns UI and related animations.
Our current workflow is (on a Facebook hidden object game) :
- Artists produce PSD and animate prototypes in Flash
- Artists re-organize their FLA files to be a bit more "programmer friendly"
- Programmers retouches assets until they have the right structure and export classes inside a SWC, from Flash
- Programmers try to improve performances, sometimes degrading the quality of game graphics
Our main idea is to hire somebody dedicated to prepare assets for programmers but I am really looking forward to improving the pipeline.
I was wondering if you guys have tips of any kind to improve this workflow, whether it be team organization, training, tools or tips with Flash. Any explanation on your asset pipeline is well appreciated too.
EDIT :
Thanks for your replies. I do agree with you that tools and processes will never replace collaborative people. However, I feel like we got stuck into this "I am doing my own thing" mentality as the integration process caused a lot of friction. As part of our Scrum process, the team clearly acknowledge its own quirks and was looking at immediate solutions for the next project : better rules, better attitude, better understanding of each other.
My team have specialized artists (UI, character drawing, character animation), but no leading art director (we use existing licenses/brands), and programmers (both front and back end). That's why having a very versatile person in between (designer, integrator, technical artist or however you want to call it) is invaluable in my opinion - so, I am looking for innovative solutions.
Usually, the main specs/requirements for assets are :
- Naming (of classes, textfields or various visual objects that will be used by the code, taking in account that visual needs can cause deep nesting of objects)
- States (that we implemented with frame labels)
- Alignement (always build from 0,0 - use round values)
So, while adopting "good practices" can sure help, there always is this little part of feature-specific requirements that has to be decided. Basically I was looking for a an efficient way to communicate these specs (even to non-technical people) and avoid the back and forth to verify assets are as expected (this is where my idea of intermediate validation format comes from).
I understand there is no clear answer and I created this post to get people feedback and good ideas (it's working so far :)