# Introduction

Generally, there are a lot of different ways and opinions how to teach people something new. However, the significant majority of them suggested hands-on experience as one of the best techniques to make human brain to remember. Learning must be fun. Some kinds of knowledge are more suitable for this practical type of learning than others, programming is fortunatelly one of the better ones. Possibility of trying things in real time is just amazing.

This knowledge is needed to reflect also in university education system. In computer science, there are several specific requirements -- the code must be efficient and easy to read, maintain and extend. Correctness and efficiency can be tested automatically to help teachers save time for their research, but checking for bad code, habbits and mistakes is really hard to automate and requires manpower.

Checking programs written by students is very timeconsuming and also boring. First idea of automatic evaluation system comes from Standford University profesors in 1965. They implemented such software, which evaluates code in Algol submitted on punch cards. In following years, many similar products were written.

There are two main ways how to automatically evaluate source code -- statically (check the code without running it; safe, but not much precise) or dynamically (run the code on testing inputs with checking the outputs against reference ones; needs sandboxing, but provides good real world experience).

This project focuses on the machine-controlled part of source code evaluation. First, problems of present software at our university were discussed and similar projects at other educational institutions were examined. With acquired knowledge from such projects in production, we set up goals for the new evaluation system, designed the architecture and implemented working solution. The system is now ready for pruduction testing at our university.


## Current solution at MFF UK

Ideas presented above are not completely new. There was a group of students, who already implemented an evaluation solution for student's homeworks back in 2006. Its name is [CodEx - The Code Examiner](http://codex.ms.mff.cuni.cz/project/) and it is used with some improvements till now. Original plan was to use the system only for basic programming courses, but now it is used much more widely and in different conditions than intended.

CodEx is based on dynamic analysis. It is a system with web-based interface, where supervisors assign exercises to their students and the students have a time window to submit the solution. Each solution is compiled and run in sandbox (MO-Eval). The metrics which are checked are: corectness of the output, time and memory limits. Supported languages are C, C++, C#, Java, Pascal, Python and Haskel.

Current system is old, but robust. There were no major security incidents during its production usage. However, from today's perspective there are several drawbacks. The main ones are:

- **web interface** -- The web interface is simple and fully functional. But rapid development in web technologies opens new horizons of how web interface can be made.
- **web api** -- There is only very limited XML API based on outdated technologies in current CodEx. This locks users from creating custom interfaces like command line tool or mobile application.
- **sandboxing** -- MO-Eval sandbox is based on principle of monitoring system calls into operating system and blocking the bad ones. This could be easily done only for single-threaded applications. These days parallelism is very important part of computing, so there is requirement to test multi-threaded applications too.
- **instances** -- Different ways of CodEx usage scenarios requires separate instances (Programming I and II, Java, C#, etc.). This configuration is not user friendly (students have to register to each instance again) and puts unnecessary work to administrators. CodEx architecture does not allow to share hardware between instances, so not trivial amount of additional hardware is occupied.
- **task extensibility** -- There is a need to test and evaluate complicated programs from Parallel programming or Compiler principles classes, which have more difficult evaluation chain than simple compilation/execution/evaluation. CodEx is only capable of the simple solution without possibility of easy extension.

After considerring all these facts, CodEx cannot be used anymore. The project is too old to just maintain it and extend for modern technologies. Thus, it needs to be completely rewritten or another solution must be found.


## Analysis of related projects

First of all, some code evaluating projects were found and examined. It is not a complete list of such evaluators, but just a few projects which are used these days and can be an inspiration for our project.

### Progtest

[Progtest](https://progtest.fit.cvut.cz/) is private project from FIT ČVUT in Prague. As far as we know it is used for C/C++, Bash programming and knowledge-based quizzes. There are several bonus points and penalties and also a few hints what is failing in submitted solution. It is very strict on source code quality, for example `-pedantic` option of GCC, Valgring for memory leaks or array boundaries checks via `mudflap` library.

### Codility

[Codility](https://codility.com/) is web based solution primary targeted to company recruiters. It is commercial product of SaaS type supporting 16 programming languages. The [UI](http://1.bp.blogspot.com/-_isqWtuEvvY/U8_SbkUMP-I/AAAAAAAAAL0/Hup_amNYU2s/s1600/cui.png) of Codility is [opensource](https://github.com/Codility/cui), the rest of source code is not available. One interesting feature is 'task timeline' -- captured progress of writing code for each user.

### CMS

[CMS](http://cms-dev.github.io/index.html) is an opensource distributed system for running and organizing programming contests. It is written in Python and contain several modules. CMS supports C/C++, Pascal, Python, PHP and Java languages. PostgreSQL is single point of failure, all modules heavily depends on database connection. Task evaluation can be only three step pipeline -- compilation, execution, evaluation. Execution is performed in [Isolate](https://github.com/ioi/isolate), sandbox written by consultant of our project, Mgr. Martin Mareš, Ph.D.

### MOE

[MOE](http://www.ucw.cz/moe/) is grading system written in Shell scripts, C and Python. It does not provide default GUI interface, all actions have to be performed from command line. The system does not evaluate submissions in real time, results are computed in batch mode after exercise deadline. Used sandboxing environment is Isolate. Parts of MOE are used in other systems like CodEx or CMS, but the system is generally obsolete.

### Kattis

[Kattis](http://www.kattis.com/) is another SaaS solution. It provides pretty nice web UI, but the rest of the application is too simple. Nice feature is usage of [standartized format](http://www.problemarchive.org/wiki/index.php/Problem_Format) for exercises. Kattis is primarily used by programming contest organizators, company recruiters and also some universities.


## ReCodEx goals

From the research above, we set up several goals, which a new system should have. They mostly reflect drawbacks of current version of CodEx. No existing tool fits our needs, for example no examined project provides complex execution/evaluation pipeline to support needs of courses like Compiler principles. Modifying existing project is also not an option, because of specific university environment. To sum up, existing CodEx has to be completely rewritten, with only small parts of adopted code (for example judges).

The new project is **ReCodEx - ReCodEx Code Examiner**. The name should point to CodEx, previous evaluation solution, but also reflect new approach to solve issues. **Re** as part of the name means redesigned, rewrited, renewed or restarted.

Official assignment of the project is available [here](http://www.ksi.mff.cuni.cz/sw-projekty/zadani/recodex.pdf) (only in czech). Most notable features are following:

- modern HTML5 web application written in Javascript using suitable framework
- REST API implemented in PHP, communicating with database, backend and file server
- backend is implemented as distributed system on top of message queue framework (ZeroMQ) with master-worker architecture
- worker with basic support of Windows environment (without sandbox, no general purpose suitable tool available yet)
- evaluation procedure configured in YAML file, compound of small tasks connected into arbitrary oriented acyclic graph


## Terminology

Official terminology of ReCodEx which will be used in documentation and within code.

* **Exercise** -- Exercise is a template of programming problem including detailed text description, evaluation instructions, sample implementation and reference inputs and outputs. Author of exercise is mostly lecturer of a programming class.

* **Assignment** -- Assignment is basically instance of exercise which was assigned to a group of students by their supervisor. Supervisor can alter predefined restrictions for resulting code (execution time limit, etc.), deadlines and maximal amount of points for correct solutions.

* **Reference solution** -- Solution of exercise provided by author. This solution should pass all test cases and could be also used for auto-calibration of the exercise. One exercise could have more reference solutions, for example in different programming languages or with various level of complexity.

* **Submission** -- Submission is one student solution of an assignment received by ReCodEx API. Submission can contain submitted source code and additional information about assignment, exercise or submitter.

* **Job** -- Piece of work for worker, generally corresponding to evaluation of one submission. There are also other types of jobs like benchmarking submission for memory and time limits configuration, but this classification has no effect for evaluation. Internally job is set of small tasks defined in exercise configuration. Job itself is transfered in form of archive with submitted source codes and configuration file written in YAML. 

* **Task** -- Atomic piece of work defined in job configuration which can execute external program or some internal command. External program execution is (mostly) performed in sandboxed environment, internal commands are executed directly. For example, one task could make a new directory, copy a file or compile source codes using GCC.

* **Test** -- Test is a piece of work to check correctness of a program. There are multiple tests inside job, which together checks validity and correctness of all aspects of exercise solution. In easiest case, testing is done by providing reference inputs to the tested program and results are compared with reference outputs. One test consists of multiple tasks.

* **Judge** -- Judge is a standalone comparision program which compares sample outputs against output from tested program.

* **Limits** -- Tasks executing external programs are usually executed in sandbox, to limit these programs predefined limits are used. These limits are specified in job configuration and can limit execution time, consumed memory, used disk space and some other worth-to-watch values.

// TODO: which one, the above or the below?

Tasks executing external programs are usually using sandbox with defined limits on run time, consumed memory, used disk space and others. The term 'limits' in this context means all these restrictions for program execution together.

* **Hwgroup** -- Hardware group is set of workers with similar hardware capabilities. Each group has unique string identifier and every worker in particular group has that identifier inside its configuration. Hardware group management is done manually by system administrator. Jobs are routed to the workers according to hwgroup, limits are also tied up with specific hwgroup.