Multiple "Add Tag" rules problem

I was playing around with Hazel and Mavericks today and encountered this strange behavior.

1) Rule 1 marks any file added to the desktop with the tag "Inbox".

Works fine.

Then...

2) Added Rule 2 to desktop. If Name starts with "Screen Shot," then add tag "Screenshot".

PROBLEM:

Rule 2 will not run if Rule 1 is active. If I disable Rule 1, then Rule 2 works.

I guess my question is, are you not allowed to add tags in two separate rules for one location?

PS when I'm creating Rule 2, the "Screenshot" tag already exists but Hazel gives me a "Create tag..." prompt instead of displaying the existing tag "Screenshot" during autocomplete.

EDIT: I also noticed that when entering a tag in the Rule Screen, the "All Tags" link in the auto-complete drop-down does not open the tag window as it does within the Finder.

Only one rule can match a particular file so you need to combine and order them properly (this is not tag specific nor new behavior). As for the UI issues, those should be fixed in the next patch.

Oh OK, I didn't realize this. I don't see how I can do this by combining into one rule or ordering them properly. Thanks.

EDIT: My rules work if I do something to the screen shot file, e.g. rename it. If I don't, the other rule won't work.

Hi!

Now I know why my rules aren't working the way I want them too :-(

I have been scanning hundreds of documents that have been sitting waiting in boxes for years. I have a rule which detects and tags the year, then a rule which *should* detect and tag as bank statements, etc.

I was hoping they would be like mail rules which fall through to the next one unless a stop action is added. This would make setting up rules a lot easier as I don't have to set up the bank statement rules for every year.

Is there any chance of adding this functionality?

Many thanks,
dm.

It specifically does not work like Mail because it actually is more confusing that way. People tend to write their rules in isolation and do not expect them to fall through.

There is a feature on the list to have a "Continue evaluating rules" type action. I can't say if and when it will be done as it does introduce a few complications.

So, the only real way to make something like this work is to set up a chain of folders and have rules do the tagging (etc.) and pass them onto the the next folder until all the tagging and desired actions are performed.

This is not exactly less confusing than the way Mail.app does it. I think most of the users who buy Hazel are pretty advanced so I don't really see why they would be entirely confused with such functionality.

When you state that people tend to write their rules in isolation, I think you need to realise that there are a boat load of articles about going paperless that mention Hazel. People are going to start thinking of more and more advanced workflows. We can already see so many posts about embedding Applescripts, etc.

I myself wanted the fall through (which I assumed was the default process and other rules wouldn't continue if the file was no longer there due to a rename, etc.) so that I could actually just download a bank statement or invoice or tax document (or whatever else) and have pretty much 99% of my filing done automatically. Anything not processed would be put into a folder so I could analyse the document and tweak the rules. Nirvana!

In conclusion, please, please, please(!) reconsider implementing this functionality. Even it is enabled like the hidden debug switch.

Thanks for listening,
dm

I don't think you are going to be able to convince me to make this the default behavior. Note that I may be dealing with a considerably larger sample size of users than you.

As I mentioned before, I am willing to add the ability to do this but it won't be the default and it probably won't happen tomorrow since it does introduce complications into the engine (Hazel's engine is a good deal more involved than Mail's).

Also, in general, when making feature requests, the best thing you can do is just state your personal use case and nothing else. Any extrapolation to other users or interpretation of what the right solution is, should be left to me.