115 lines
5.0 KiB
Markdown
115 lines
5.0 KiB
Markdown
# Hermes
|
|
|
|
### Statement
|
|
Hermes is a multi-client chat program akin to IRC written in Racket. Building
|
|
Hermes was interesting as it exposed us to various design problems namely networking,
|
|
synchronization, scheduling, GUI design, and component design.
|
|
|
|
### Analysis
|
|
> Will you use data abstraction? How?
|
|
|
|
TCP communication has been abstracted away, so that we deal with Hermes
|
|
definition of a message.
|
|
|
|
> Will you use recursion? How?
|
|
|
|
The server continually loops waiting for connections from clients. The clients
|
|
are always on standby to receive input.
|
|
The GUI continually loops to handle input from the user,
|
|
as well as to keep the canvas it writes the messages on updated.
|
|
|
|
> Will you use map/filter/reduce? How?
|
|
|
|
Map was used for dealing with input area of clients, and iterating over a list
|
|
of open ports to send messages. Filter was used to find the recipient of
|
|
a whisper.
|
|
|
|
> Will you use object-orientation? How?
|
|
|
|
Keeping count of the number of clients required working with objects that are able to
|
|
increment and decrement the number of users. We handled a list of connection
|
|
ports, messages similarly.
|
|
We also keep the GUI in an object so the many moving parts of the
|
|
user interface are packaged in one place.
|
|
|
|
> Will you use functional approaches to processing your data? How?
|
|
|
|
The communication part of Hermes is over tcp which uses a lot of functional
|
|
approaches e.g. you start a listener which you can call tcp-accept on.
|
|
The result of tcp accept are two pairs of ports which we can then bind to some
|
|
variables. Functional approaches are exemplied in most of the code base.
|
|
|
|
> Will you use state-modification approaches? How? (If so, this should be encapsulated within objects. `set!` pretty much should only exist inside an object.)
|
|
|
|
State-modification was used e.g. keeping count of logged in users requires
|
|
state modification via set! to maintain the true user account, managing the list
|
|
of open connections and messages required state-modification.
|
|
The user interface also needs a few states that it needs to keep up to date.
|
|
|
|
> Will you build an expression evaluator, like we did in the symbolic differentatior and the metacircular evaluator?
|
|
|
|
We allowed the use of a few commands through the user interface. The most notable ones
|
|
are the /whisper to send private messages to a user, /list count and /list users
|
|
to view user statistics , and the /color command to allow
|
|
the user to change the color of their text.
|
|
|
|
### Deliverable and Demonstration
|
|
There are two big deliverables for this project. Code for the server
|
|
, and the clients which not only has code for interacting with Hermes,
|
|
but also a GUI for interactivity with a user.
|
|
|
|
We are going to demonstrate Hermes by running the server code on a remote machine.
|
|
We will connect to the server via our PCs running client code. We will ssh into
|
|
the remote machine to see the server running. Since Hermes is a multichat anyone
|
|
can join in the demonstration by connecting their computers to the remote
|
|
machine!
|
|
|
|
|
|
|
|
### Evaluation of Results
|
|
Evaluating Hermes was very simple. Can at least two clients hold a meaningful
|
|
conversation remotely? If Client A speaks at 11:01 am, and client B does so at
|
|
11:01 plus a few seconds, Hermes has to convey this state correctly. Is the GUI
|
|
intuitive for current irc users? We successfully met these questions, and more.
|
|
|
|
|
|
## Architecture Diagram
|
|
|
|
#### Preliminary design
|
|
![Architecture](https://github.com/oplS17projects/Hermes/blob/master/ext/arch_diagram.png)
|
|
|
|
|
|
#### The Game plan
|
|
![Diagram](https://github.com/oplS17projects/Hermes/blob/master/ext/architecture_diagram.png)
|
|
|
|
|
|
## Schedule
|
|
The first step in our project was to setup a system to get data from one machine to another. What data exactly wasn't directly important and the other machine didn't really need to display it in a pretty manner, it just needed to relay that it has recieved the correct information.
|
|
|
|
Next we needed to create a user interface that looked nice. Some way to control the connection and display information in a convient and readable format.
|
|
|
|
After we finished the user interface and connecting the machines, we needed to merge them together and begin expanding the utility if time permits.
|
|
|
|
### First Milestone (Sun Apr 9)
|
|
Get two different machines to relay information meaningfully.
|
|
|
|
### Second Milestone (Sun Apr 16)
|
|
Get a GUI that looks professional and uses the correct format.
|
|
|
|
### Public Presentation (Mon Apr 24)
|
|
Merging the GUI and information relay together into one program. If time permits we also plan on adding additional features.
|
|
|
|
## Group Responsibilities
|
|
|
|
### Douglas Richardson @Doug-Richardson
|
|
I have written the code for the GUI.
|
|
It presents the user with a simple readable format for displaying the information
|
|
that the server provides. For the most part the program only interacts with the user
|
|
through the GUI.
|
|
|
|
### Ibrahim Mkusa @iskm
|
|
I have written the networking code i.e. code that allows communication between
|
|
clients through server. I wrote scheduling code responsible for queueing
|
|
fairly the client messages and broadcasting to the rest of connected
|
|
clients. Implemented the logic for handling /list, /whisper commands.
|