# Meeting Notes 2/8/2019
## Discussing what we need to do next...
Basically, we started talking about the algorithm, and what is considered to be plagiarism. However, we *bikeshedded* it, because of the short duration of this meeting (45 min). Instead...
## Structure of the system
- We can build out a structure of the system, without the actual plagiarism detection in it.
- Obviously, there will need to be some basic functionality that we know we need, such as moving a zip file from the client server to the processing server
## How to code for the server
- Use GitHub. Push to the repo, then update the server by pulling from the GitHub repo.
- Make sure to have a proper configuration file
## What language to code the servers in
- We agreed on C#, since it is basically Java, and doesn't use the JVM
- Also, it is easy to use for server schtuff
## Created Git Issue for initial task
- See "Set Up Basic Structure of System"
## How to structure working on things
- The systems can be worked with individually
- We should define the communication between components first, that way when we split up to work on the components, it will be done the right way.
## Other Stuff that was talked about
- The client server should also scrub out the common source files (i.e. node_modules)
- Inter-language detection is not within the scope (If they spent the time translating, they probably know it).
## Servers while developing
- Processing Server will be Dom's server
- Client Server will be... our computer?
- - This is difficult, because the client server needs
- - So we will probably just have EVERYTHING set up on our dev computers
- - - (i.e. Client Server, Processing Server and Client Portal will be on the same computer)
## How will we be sending the results back?
- Can either have both servers with ports open
- - Less secure
- Or we can have the processing server store the results
- - And the client server will request the results when they are needed
- - But this is kinda against what Bockus said? Not really, he just said it won't have any "Client data". So if we store the results, that should be fine, since it is just which files and lines match. Filenames will be scrubbed, so that should be fine.
## What are we working on this weekend?
- Do the hackmd document stuff, describing the communication between components