OverviewTeaching: 10 min
Exercises: 10 minQuestions
What are the key differences between research software and ‘normal’ projects?
What does ‘done’ look like for a research software project?
What are the goals of this class?Objectives
Compare and contrast research software projects and ‘normal’ software projects.
Name and explain key features of a mature research software project.
- Scientific disciplines across the board are becoming more compute intensive
- modeling, simulation, data analysis, data management
- Few scientists have any formal training in programming
- Much more to a research software project than just code:
- We should think about programming not as instructing a computer what to do, but as communicating with other humans what we want the computer to do (paraphrased from Donald Knuth)
- This class will discuss:
- What a research software project needs to achieve its goals
- How to go from ad hoc solo project to reliable small-team project
Key differences between research software and “normal” projects
- Requirements may be either:
- Discovered as we go along (exploring)
- Relatively stable (engineering)
- Problem is subtle as well as complicated
- Control flow may be very simple…
- …but figuring out exactly what coefficients and indexes to use can be very hard
- Staff are often research software engineers (RSEs)
- Combine expertise in programming with an in-depth understanding of research
- Typically have:
- An advanced degree in a research discipline
- Lots of experience developing software (often acquired while working on their thesis)…
- …but little or no formal training in software engineering
- “The astronomers who build telescopes so that others can study stars”
What does “done” look like?
- Software can be used by people other than original authors
- Reproducibility meaningless without this
- Reasonably confident that results are correct
- As trustworthy as physical experiment
- Small changes and extensions are easy
- I.e., researchers can safely make small changes to code
- Fast enough to be useful
- Can be maintained by people other than its original author(s)
- I.e., it is sustainable
- Has a citable publication described in an easy-to-find way so that users can credit the project in an academically-digestible form
For our purposes:
- 3x6: three people for six months
- Contributors are frequently time-slicing other projects
- “Everybody makes coffee”
Create a Value Proposition for Your Project
Fill in the template below for your current project.
- For [description of target users]
- who want to [statement of their need(s)],
- [project name]
- provides [statement of key benefits].
- Unlike [name of alternative solutions(s)],
- our project enables users to [key differentiator].
Describe How Your Project is Managed
Write a short point-form description (5-6 bullets) of how your current project is managed:
- Who uses the software?
- How do they find the software?
- How do they set it up?
- Who decides what to change and when?
- How are decisions and changes circulated?
Requirements for small research software projects are typically emergent.
Research software is ‘good enough’ when people other than its authors can use it with confidence and extend it with reasonable effort.