Folks … I am a newbie in fixlet authoring so apologies in advance if this has already been answered.
When you have multiple relevance in a fixlet, how can one determine which relevance(s) is
really
evaluating to true for a given client. I suppose I could use fixlet debugger to check what each relevance evaluates to on my authoring machine but it wont work for all my clients. Another way for me would be to manually check each relevance statements and see if it applies to the client - this might work if I have one or two relevance entries but will soon become cumbersome if I have more than 10-15 relevance coded.
If the Fixlet is reporting relevant, then it is true… you can double-check a few clients like you mentioned to make sure… are you worried that the results are not accurate?
I am sure results are accurate but I am thinking of cases when I need to know why my fixlet is showing relevant when I am certain it shouldn’t be. In those cases I’ll have to check all my relevances one by one on the client to see which of them is returning true to figure out the issue. It would have been nice if I can see in console which relevances are evaluating true (probably color of relevance changes or some sort of text saying “these relevances are evaluated true for this client …”)
For scalability and performance reasons, we don’t collect the details of every sub-clause for every Fixlet for every agent… But… we think it would be a cool feature that we can add if you were able to click on a computer and have it remotely evaluate them for you… Hopefully we can get to this feature someday soon (it is a pretty large amount of work to make it work right).
I almost never write fixlets without first putting the various relevance elements into an analysis for exactly this purpose. It helps to find any edge cases where some relevances may be returning errors or unexpected return values.
Now that we’re able to write custom fixlets and tasks that have separate relevance statements, I’m wondering if it’d be worthwhile to write a script that could take an exported fixlet/task and create an an analysis (with all the relevance as separate properties, named Rel1, Rel2, etc. And do the reverse as well, take an analyses, collect all the properties and create a empty action fixlet with either the relevances OR’d or AND’d. Hmmm.
This would mean that you could work either way (starting with a fixlet or analysis) and it’s not double the work when you feel like creating the other one for testing.