Answer: YES! Version control systems do not belong to programmers ONLY. They are very useful for writers, designers, analysts and data teams, and the type of file will determine which tool is the right one. Git works well for text and code, version control history as a built-in feature of any application works well for casual documents, and Git LFS, Perforce and DVC work well for large binaries and data. Don’t write to what’s popular, write what you version.
Version History Isn’t Version Control
Most of the non-developers already have a bit of history with versions. As often as to explain Git to some non-programmers, a consultant does: “Google Docs and similar solutions are revision history trackers, which means they will write down what you type at any instant. Instead, git monitors the entire project. Think of Google Docs, where you write a book, but would get a version history for the document as well as each chapter and the book itself.
Collaboration differs too. In a word processor, teams may write a single document or sends documents with names such as final-draft3-revisions(2).docx via email. Git allows everyone to work on the same project, without destroying other’s work, and the group chooses how things are coupled together.
What Works and What Does Not
Version control systems were originally conceived for text files. One University course description states that they can store binary files, like images, but that they will not be able to merge changes in such files as they do merge changes in text files. That is a restriction that gives rise to most of which this article treats.
It also gives an explanation as to why large binary assets create problems. Teams that spend a lot of time working with large binary assets may require Git LFS, file locking or something else.
The File Type Fit Chart
The File Type Fit Chart (Original Comparison)
This is an editorial chart and not a benchmark chart and is reflective of published descriptions.
| You | Good starting point | Why |
| Code, Markdown, plain text | Git | Merges text well and has a vast ecosystem |
| It is important to note that binary formats don’t merge like text: word documents and spreadsheets | built-in history or a document tool like Simul. | |
| design files and large media | The design tool’s history is stored here or via Git LFS / Perforce | Big binaries require locking or special storage system |
| DVC | Manage versions of data and models along with code | |
| Database schemas | Migration tools like Flyway | Traces Database Schema changes between the environments. |
| Game Assets | Perforce or Unity Version Control | For large binary assets. |
Version Control for Data Teams
Data Teams is a version control for data.
DVC is an open-source tool for data science and machine-learning projects. It is run in parallel with Git. This means the code itself and DVC’s small metadata files are versioned by Git and DVC’s large data and model files are stored in remote storage like S3, Google Cloud, Azure or SSH server. The pipelines tie together code and data, and one may be able to visualize how a particular model was developed from given data. It is designed to replace spreadsheets and shared documents, which teams use to keep ledgers.
Writers and Researchers
Git can be used by writers to work on text-based files. A university course page explains how a writer or group of writers could have multiple branches each containing their own histories, for example, separate drafts. Plain text or Markdown works a lot better than Word files as it is possible to compare line by line. If your editors are the type that only use Word, keep to the document’s built in history and a sort naming system.
Non-Developer Starter Path
Original Framework – Non-Developer Starter Path
This is a style guide, not an official guide.
- Pick one folder. Choose one project, now is not the time to choose your entire drive.
- Prefer text. Use Markdown or other text format if possible.
- Commit small. Take a snapshot after every significant change accompanied by a brief note: what was changed, and why.
- Back up remotely. Move to a hosted solution to maintain the history even if a laptop is lost.
- Keep binaries separate. Store large files in storage specifically designed for the large files and keep track of them using a reference.
If the command-line isn’t for you, you can use a graphical Git client. Remember that the commands don’t matter as much as the concepts.
Version Control Fit Checklist
Version Control Fit Checklist (Linkable Asset)
- Opened and listed the file types I want to version
- Differentiate between separated and binary files
- Select tool for each file-type
- Experienced a small project up front
- Committed messages that discussed reasons for
- Set up a remote backup
- Decided upon need for file locking
- Captured process for co-workers
- Prevented storing sensitive data and credentials in the repository
- It is setup and has been reviewed after 1 Month.
Guest Post Angles
The following topics are appropriate for publications related to productivity, data sciences and creative industry:
- It’s git for writers – a markdown workflow to replace final_v7.docx.
- DVC for beginners: Versioning Data without overloading your repository.
- So why Google Docs history is missing version control: what you lose.
Version One Folder This Week
This week, it’s Version One Folder!
If you haven’t yet been using a version control system, don’t read anything about it for another month. Select a single simple project folder, containing predominately text documents, and use it to create a repository, with five well described commits and push to a remote backup. After a week, you’ll be able to determine if version control systems are a good fit for your work, and which type of files are suitable for another tool.
