- Avoiding writing for outputDo not include formatting information in your DITA content topics. 	     
- Stem sentences, glue text, and other transitional informationTransitional information, such as stem sentences, is largely 	 redundant in DITA's minimalist, topic-based, semantic mark-up architecture. The 	 role of transitional information in  	 contextualising 	 the content has been eliminated through the use of smaller chunk sizes, 	 meaningful titles, information typing, structure, and demarcation labels 	 automatically-generated in the output.     
- Punctuation in listsOrdered and unordered list items should have no terminating 	 punctuation marks, such as commas, semicolons, and full stops, except when all 	 items in a list are complete sentences. This approach  	 maximises 	 re-usability.     
- Titles and headingsTitles are used in a number of mark-up structures in DITA. Titles 	 are a mix of content and metadata. The mark-up in titles should be kept as 	 simple as possible, and workarounds to achieve specific formatting outcomes 	 should not be used.     
- Crafting paragraphsA paragraph should contain one unit of thought or action, or main 	 point, and should be typically three to eight sentences in length. All of the 	 sentences in a paragraph should relate to its main point.     
- Different languagesThe  	 xml:lang attribute stores the human language of the content, 	 and can be applied to any DITA element.     
- Quotation marksQuotation marks should not be included in DITA content. The marks 	 should be added, if required, during the publishing process.