In John Boykin's article, Stop Helping Me Google Calendar AutoSuggest #WTFUX, published by UXMag.com, he states unequivocally that Google Autosuggest is a ridiculous tool to include in Google Calendar. His assessment appears to be an opinion piece based on personal preference and emotion. Luckily for the rest of us, Boykin does not work for Google and is only one very prolific stakeholder with an extensive online following. I disagree with Boykin's statements, but his article makes an excellent example of informal evaluation in the summative phase of assessment. The language used in this artifact is written for a broad audience, likely meant to provoke emotion and drama so that readers will comment and would not be considered an academic or facts-based article.
Since we are discussing usability vs. evaluation this article also shows one perspective of user experience. This user is frustrated with a product and is providing public feedback which may be useful for Google developers.
Usability is the act of making EIT easily usable and a process efficient for users. Accessibility is the process of making sure EIT is accessible to all users regardless of their disability status. Evaluation is analyzing the feedback and information attained from those two fields (and others) to make informed judgement(s) that are defensible to stakeholders.
I've given my feedback and Boykin wrote his feedback. If I where a paid elevator to do a formal usability assessment/evaluation of this product, I would do user research, testing, and interviews to ascertain a wider array of thoughts regarding this product. Since the feature Boykin is complaining about is not actually impeding his use of the product, he just doesn't like the way it functions, my gut reaction would be to offer a button to easily turn on/off this functionality. That being said, Google already has enough hidden features and is cluttered by functionality so maybe this feature could integrate local resources and commonly typed places so that eventually this functionality would be useful to Boykin. More informed recommendations could be made with an evaluation and usability testing.