Make the non-recommendation recommendations more intuitive

this’ll be my last UI-enhancement issue before whichever comes first of Doctor V1 launch or us clearing all the current issues - but I think this is an important one. Edit: Conor seems to agree 😃

As explained in Andreas’s talk, we deliberately show all the articles on what steps can be taken as tabs so a user can browse and isn’t prevented from learning about a possible problem if Doctor has for some reason not picked up on it. See also https://github.com/nearform/node-clinic-doctor/issues/39

The current implementation, however, is potentially very confusing for new users. For a user to understand what the different tabs mean, they have to figure out that a red icon next to a tab means “This is a problem you have” while no red icon means “This is a problem we don’t think you have, but feel free to read about it anyway” - which is particularly confusing since the text for these is written in a way that makes it sound like it is a problem you have (e.g. “The process may have a memory management issue”). It certainly confused the heck out of me when I was exploring Doctor - before Andreas’s talk which explained it.

Here’s my proposed plan to fix this:

• By default, hide tabs that don’t have the .detected class, with a show/hide button with an intuitive label so a user who wants more to read can see they’re available. Something like “Show other possible issues” - but this wording needs a little work to make it clear it’s issues that exist in the wider world that Doctor doesn’t see any evidence that you have, rather than issues where Doctor thinks there is a little evidence that you might have it. Feel free to make suggestions, and I’m planning to get some input from Joy here. Design details will depend on the length of the wording.
• Add a class that can be added to tabs like “Unknown” and “No issue” so they aren’t shown by the show/hide button and only shown if they are the result of the analysis (.detected). A user who wants to learn about potential Node performance issues isn’t looking for the wording of our “No issue” message! 😃
• Reword the notice added to .detected issues to be a bit more explicit. Currently it says “Potential [name of issue] issue detected: [bullet list of signs and symptoms]”, I’m thinking something like “Doctor has found evidence suggesting a [name of issue] issue: [bullet list of signs and symptoms]”
• Add an equivalent notice to the non-detected issues setting a “hypothetical” context to the rest of the text. Something like “Doctor has not found any clear evidence of a [name of issue] issue. If there was such issue: [bullet list of signs and symptoms]”

These should be quite simple changes, and they should mean the non-detected issues can still be easily found by an interested user but don’t confuse or distract a user who just wants to see Doctor’s recommendations.

[AlanSl]

Good suggestions and reasoning, I like the “[Verb] issues [noun]” structure, and “archive” does a good job of getting across that this is the full roster of inactive issues that detected issues are plucked from. Only issue is that “explore” and “archive” imply a large, searchable resource, when we’re just showing a couple of extra tabs… and with archive buttons increasingly common in apps, “archive” sounds like a feature that archives a user’s previous Doctor profiles and recommendations. I was thinking “browse” for similar reasons you gave to “explore”: both get across that it’s no longer Doctor leading you by the hand.

Not sure about “diagnosis archive” since a diagnosis is something individually tailored that you give to one patient: I’d definitely expect a diagnosis archive to be an archive of the diagnoses I’ve personally been given. Also I think we’re keen to emphasise that the good doctor is referring the patient to a specialist for further testing at this point, not giving a definite diagnosis.

Doesn’t help that all the good words for this already have a technical usage: directory, repository, cache, registry, library… “catalog” nearly fits but sounds too much like shopping, “encyclopedia” and “compendium” feel pretentious, “knowledge base” would be ideal but it’s a horrible expression 😕 )

For now I think “Browse undetected issues” is the most accurate three-word description of what we’re doing.

[AlanSl]

Fixed in https://github.com/nearform/node-clinic-doctor/commit/1fb5d33b81b3a4cf2073963f8b043ec9a30ff720