Thank you! I started to bring more structure to my documents with attributes using agents. How much of agents can I have till it will affect performance?
It depends on (a) how many notes you have, and (b) whether your queries are reasonably efficient.
For example, a very complex regular expression is burdensome when you have a lot of text. Similarly, a query that needs to look only at the note in question is easier than a query that needs to look at a lot of other notes. (Compare "Is the price over $100?â to âIs the price greater than the average of all prices?â)
Even a very complicated query will be no problem if you have dozens or hundreds of notes.
People worry a lot more than they should about agents, actions, and performance. First: performance used to be a problem, because when agents were running you had to sit and wait. Those days are long ago, but the memory remains.
Also, some people who do really complicated things with really immense documents do have real performance concerns. These are, it turns out, often the people who visit the forum most often, and so you hear more about performance than you ought. (Even if you have a real performance issue, the worst that typically happens is that agents donât update promptly, or documents take a while to load.)
Also, some people spend a lot of time on airplanes or in other places where they rely on laptop batteries. Agents use power. Not a lot of power, but it can matter if youâre Tinderboxing your way across the Pacific Ocean. That, too, is a diminishing worry; lots of planes provide power now, and Apple Silicon Macs have big batteries.
General advice: donât worry about performance until you have a problem. If you do have a problem, there are usually ways to make it go away. Odd are, you wonât have any trouble.
1 Like
Iâve just substantially updated the section of aTbRef on Taggers, including some new child notes (see list at the end of the main article).
Less to do with performance than keeping the aisles clear for work, I try to get in the habit of deleting agents I donât need. If the query (and/or) action used in the query is complex its easy to store it in a note using a built-in Action or Code prototype. Itâs very easy to get in the habit of always adding agents, e.g. just to quickly find something thatâs hard to define in a Find panel but then now delete it when done. Suddenly you are overrun with agents who purpose youâve forgotten and which arenât contributing meaningfully to the document. such clean-up also reminds one that agents are easy to (re-)make so hanging onto unneeded agents is only making a mess.
Donât take any of this too literally. There is no hidden âtoo manyâ number for agents. Find the balance that suits your needs. Also, this article is probably pertinent Taggers and agents.
Do you use text expansion for quiries code?
I also thought like that, but then I came upon thought that maybe I can just keep these agents to add structure according to defined rules to new notes. Thatâs why I asked about performance. So I will follow the advice of @eastgate for a while. But also will try to get your habit of creating short-term agents.
This one useful! Looks like I missed it before or maybe it is new? Thank you!
Good questions all, but there really arenât simple hard-edged answers.
Iâm not exactly sure of the context here. I definitely use autocomplete for attribute names and code when writing queries in places like the the query inspector. As a dyspraxic non-touch typist I need all the help I can get. As to getting queries right, as with so much in life the more I practice the better I seem to get aided by trips to aTbRef (there are too many configurations for one to remember everything all the time.)
Performance? Donât worry about. The useful thing to keep in mind is that if you are one of the few who needs to consider effects of scale be aware of the things you do in the app that might contribute to any issue you find. Unless a project will stay small (small is hard to define) I would review things if I found my document was heavily dependent on or only used agents. Itâs highly likely such a pattern could be revised/ How? Well, that would depend on the task and subject matter. We take-away is being open to the fact we can often improve on our first plan (guess) of structure and happily Tinderbox is very forgiving of such activity.
Taggers and agents? Itâs not new but I doubt many knew of it as Iâd forget to link it from its parent articles. Sorry! Iâve not used the technique, but only as my work doesnât have that need. But for someone whoâs always looking for specific terms (noting that for user-defined terms there is no regex or stemmingâ ) via agents it might prove a fruitful task. As with most exploratory tools in the box here, make a small text TBX, small enough so you can check the outcome by eye. Then youâve a sense of how best to use a given tool (or, indeed, not!)
â . To my eye sensibly so. If we want a text concordance, there are tools specifically for that task, plus Tinderbox has other affordances like the Get info sub-tabs for repetition, similarity, etc.
1 Like