This wiki page, a work in progress, describes best practices for documenting research with the goal of enabling others to understand that work and to reproduce experiments / results, and assess the data resulting from that research.
This document is intended to be a collaborative effort, drawing on the expertise of many contributors. Because the goal is accuracy and adherence to sound scientific methods, not all perspectives may survive the peer-review process.
In general, scientific research should be well documented. The process does not dictate the document timeline but it does require accuracy and sufficient detail for others to clearly understand and duplicate that research (assuming they have the skills and resources available.) The final document should include 1) an Abstract, 2) a discussion of measurements, procedures, observations and techniques, 3) details about the data, accuracy, resolution, etc., 4) interpretation and analysis of the observations and data which includes ownership of it's limitations and 5) discussion of conclusions which have been proven by the observations.
When any or all of these basic fundamentals are missing or lacking in detail or scope, the author is effectively asking the reader to take what they say on the basis of assumption, not the scientific method based on facts and observable measurements. Resolving such omissions can be resolved either by clarification and added detail / data or by retraction until additional work can be performed which supports the prior assumption.
Abstracts should generally be just a short paragraph, at the top of the document, and should quickly outline 1) the focus of the work, 2) the reason for the work, 3) the nature of the experiments and observations and 4) a broad hint at the conclusions derived from that work.