The following content is discussed in scrum today:
-
- Design document alpha version review, corrected the design of the Character Model
The previous design of the character model is as follows: the user inputs an image,ProgramExtract information such as the beginning, trunk, and limbs, and map them to the predefined character model. Because we have never done similar work before, and this solution looks quite simple. However, after Chen Zihe discussed this solution with Xu yingqing today, he felt that it could be extremely difficult to implement it:
- How to extract information from various parts of a user's image is a difficult research topic, which is almost impossible within 8 weeks.
- Whether the user is willing to spend so many twists and turns to play a game, or the technical difficulties of pattern recognition, even if the physical information is extracted (possibly specified by the user ), how to map them to the character model? Because users can only provide two-dimensional, some angle photos, and other angles may appear in the game, it may be very difficult to directly hard map the past.
For the previous design, we used to only get YY in our minds. Now we have to consider these technical problems before we can find it very difficult. It seems that programmers are born to be happy. Fortunately, the definition of the program architecture is clear, and the game module only needs a series of key frames for each scheme. It doesn't matter how these key frames are obtained. After discussion, I think the more reliable method is to take photos. In the Alpha stage, you can take key frames of several actions and put them in the program to let the characters move. In the beta version, you need to consider how to smoothly combine the User-Defined Avatar and Avatar with the body.
- Next Plan
Because the design has changed, you need to re-develop the plan. Now, you do not need to create character modeling and recruitment data, because the main part of the game is how to manage images. After discussion, I think there are three urgent tasks:
- Familiar with DirectX
This is required, because we have never touched DirectX before. However, I believe that after the previous C # And Silverlight, everyone's ability to learn new technologies has been fully trained, and DirectX should not be too problematic. One person should have a deep understanding of DirectX and become a technical expert. Anyone who has any questions about DirectX can discuss it with him, which can reduce the burden on everyone.
- FrameworkCode
Because the current design is still in the document stage, the description is still not specific enough, and test cannot be started, it is necessary to write the Framework Code as soon as possible, at least to define the interface through code.
- Take a photo
In the initial stage, you can take a small number of cute photos and add them to the program to observe the actual effect, and then enrich the style.
- Summary
|
today |
tomorrow |
Fu Hao |
try to define the character model, code for writing the character model generator |
familiar with DirectX, we will discuss with you tomorrow on scrum |
Tian Fei |
conception of the Generation Method of the Character Model |
familiar with DirectX, write your own image processing demo |
Chen wenlin |
conception of the Generation Method of the Character Model |
familiar with DirectX, write your own image processing demo |
Peng Xiulian |
be familiar with program design and design a test framework |
familiar with DirectX, write your own image processing demo |
Chen Zihe |
discussed the feasibility of the original solution with instructor Xu yingqing |
familiar with DirectX, write your own image processing demo |
Shi jinhao |
How to create a character model |
Familiar with DirectX and write framework code |
-
- TFS report