Documentation Style Guide

Thank you for taking an interest in contributing to HPE Machine Learning Data Management’s docs! 🐘 📖

This style guide provides editorial guidelines for writing clear and consistent HPE Machine Learning Data Management product documentation. See our contribution guide for instructions on how to draft and submit changes.

Audience

HPE Machine Learning Data Management has two main audiences:

  • MLOPs Engineers: They install and configure HPE Machine Learning Data Management to transform data using pipelines.
  • Data Scientists & Data/ML Engineers: They plan the development of pipelines and consume the outputs of HPE Machine Learning Data Management’s data processing to feed AI/ML models.

Be sure to provide links to pre-requisite or contextual materials whenever possible, as everyone’s experience level and career journey is different.


Voice & Tone

HPE Machine Learning Data Management’s voice in documentation should consistently convey a personality that is friendly, knowledgeable, and empathetic.

The tone of voice may vary depending on the type of content being written. For example, a danger notice may use an urgent and serious tone while a tutorial may use an energetic and instructive tone. Make sure the tone of your documentation aligns with the content. If you aren’t sure what tone to convey, ask yourself: “What is the reader likely feeling when presented this content? Why are they here?” and adjust your language to the most appropriate tone.


Language & Grammar

The following guidelines are to be followed loosely; use your best judgment when approaching content – there are always exceptions.

Use Active Voice

Write in active voice to enforce clarity and simplicity.

👎🚫 Don’t 👍✅ Do
The update was failing due to a configuration issue. The update failed due to a configuration issue.
A POST request is sent and a response is returned. Send a POST request; the endpoint sends a response.

You can break this rule to emphasize an object (the image is installed) or de-emphasize a subject (5 errors were found in this article).

Use Global English

Write documentation using Global English. Global English makes comprehension easier for all audiences by avoiding regional idioms/expressions and standardizing spelling words using the US English variant.

Put Conditional Clauses First

Order conditional clauses first when drafting sentences; this empowers the reader to skip to the next step when the condition does not apply.

👎🚫 Don’t 👍✅ Do
See this page for more information on how to use this feature. For more information, see How to Use Console.
Enable the CustomField setting if you want to map custom fields. To map custom fields, enable the CustomField setting.

Write Accessibly

Be mindful of how you describe software behavior and users; in particular, avoid ableist language. Use generic “they/them” and “you” when describing users or actors; avoid the use of “obviously”, “simply”, “easily” – every reader has a different level of expertise and familiarity with key concepts/tools.

👎🚫 Don’t 👍✅ Do
To start, simply enter the following command: To start, enter the following command:
Configuring this setting just requires a simple API call. Make an API call to configure this setting.
The results, without this setting enabled, might look crazy. Your results may be inconsistent or unreliable without this setting enabled.

Formatting & Punctuation

Markdown

All documentation is written using Markdown syntax in .md files. See this official Markdown Cheat Sheet for a quick introduction to the syntax.

Code Blocks

Use ``` to wrap long code samples into a code block.

This is a code block.

Headers

Use title casing for all header titles.

  • Capitalize the first and last word.
  • Capitalize adjectives, adverbs, nouns, pronouns, and subordinate conjuctions.
  • Lowercase articles (a, an, the) and coordinating conjunctions (and, but,for, nor, or, so, yet).
👎🚫 Don’t 👍✅ Do
How to develop sentient ai How to Develop Sentient AI
How to use the pachctl cli How to Use the Pachctl CLI

Use meaningful link descriptions, such as the original article’s title or a one-line summarization of its contents.

Examples:

- See the [HPE Machine Learning Data Management Technical Documentation Style Guide](../docs-style-guide)
- Use the [official HPE Machine Learning Data Management style guide](../docs-style-guide).

Lists

  • Use numbered lists for sequential items, such as instructions.
  • Use unbulleted lists for all other list types (like this list).

Commas

Use the serial or Oxford comma in a list of three or more items to minimize any chance of confusion.

👎🚫 Don’t 👍✅ Do
I like swimming, biking and singing. I like swimming, biking, and singing.
I only trust my parents, Madonna and Shakira. I only trust my parents, Madonna, and Shakira.

UI Elements

Bold UI elements when mentioned in a set of instructions (a numbered list).

  1. Navigate to Settings > Desktop.
  2. Scroll to Push Notifications.

Organization

Use Nested Headers

Remember to use all header sizes (h1-h6) to organize information. Each header should be a subset of topics or examples more narrow in scope than its parents. This enables readers to both gain more context and mentally parse the information at a glance.

Publish Modular Topics

Avoid mixing objectives or use cases in one article; instead, organize and separate your content so that it is task-based. If there are many substasks (such as in a long-form tutorial), organize your content so that each major step is itself an article.

Examples:

The below outlines are 4 articles, with the parent article linking to each modular sub-topic.

  • How to Locally Deploy HPE Machine Learning Data Management
    • MacOS Local Deployment Guide
    • Linux Local Deployment Guide
    • Windows Local Deployment Guide

Images & Diagrams

Visualizations are helpful in learning complex workflows and UIs; however, they can expire quickly and take a lot of effort to maintain. Ask yourself the following questions when deciding whether or not to add visualizations:

  • Is the user interface complex enough to warrant a screenshot?
  • Can a diagram convey this concept more efficiently than words?
  • How frequently does this visual need to be updated?

Also, remember to add alt-text to your visualizations for screen readers.

Mermaid

You can build diagrams using mermaid.js in our documentation by:

  1. Adding mermaid: true to the frontmatter of an article.
  2. Drafting a diagram, like so:
    ```mermaid
    graph TD;
        A-->B;
        A-->C;
        B-->D;
        C-->D;
    ```
graph TD;
    A-->B;
    A-->C;
    B-->D;
    C-->D;
  

Want to make an update to this style guide? Select Edit on Github and leave a suggestion as a pull request!