> "Some groupings have icons and get inset, while other groupings don’t have icons and don’t get inset. Interesting…again I wonder what the rationale is here? How do you choose? It’s not clear to me."
The rationale is clear, they are choosing to use icons only when a widely-recognized icon is available. This makes sense, and it answers the author's concerns perfectly about icons being used arbitrarily when they don't convey anything.
To be honest I find the whole motion of this blog post quite confusing. The user starts with a bad example, people using icons randomly that nobody could recognize without the text, which is evidence of the fact that the icon itself doesn't convey much information.
Then he shows an example where someone doesn't do the thing he complained about, they actually did use icons with a rationale. At which point he asks the question, "What is the rationale" but does not actually attempt to answer it..
To me though, there is a much more interesting paradox beneath all this. If we grant that it only makes sense to show an icon when it's meaning is widely known. How are new icons ever going to be introduced? Presumably every well known went through a period where it was used with text because it was still not well known. So while it might be bad UX to use an icon that is unfamiliar to users, over the long term using such icons has the benefit of creating a shared visual language that we all understand. I guess the litmus test for when to put an icon should then become: Is this functionality widespread enough in other applications that I can imagine this icon becoming standard in the future?
The rationale is clear, they are choosing to use icons only when a widely-recognized icon is available. This makes sense, and it answers the author's concerns perfectly about icons being used arbitrarily when they don't convey anything.
To be honest I find the whole motion of this blog post quite confusing. The user starts with a bad example, people using icons randomly that nobody could recognize without the text, which is evidence of the fact that the icon itself doesn't convey much information.
Then he shows an example where someone doesn't do the thing he complained about, they actually did use icons with a rationale. At which point he asks the question, "What is the rationale" but does not actually attempt to answer it..
To me though, there is a much more interesting paradox beneath all this. If we grant that it only makes sense to show an icon when it's meaning is widely known. How are new icons ever going to be introduced? Presumably every well known went through a period where it was used with text because it was still not well known. So while it might be bad UX to use an icon that is unfamiliar to users, over the long term using such icons has the benefit of creating a shared visual language that we all understand. I guess the litmus test for when to put an icon should then become: Is this functionality widespread enough in other applications that I can imagine this icon becoming standard in the future?