See related Albert issue.

I was surprised to find out that the trigger does not require a whitespace to work. Took me a minute to figure out why searching Albert for "Software Manager" only showed "Stack Overflow" instead; I expected that I would need to type so<space> to search Stack Overflow, not just any string starting with so. That's the behavior on Alfred, Chrome, Firefox, DuckDuckGo, and pretty much everywhere else.

Actually I really like how, say, Firefox and Chrome work here, where I type the trigger (gi for google images), press space, and then a little tag ([Google Images]) shows up at the left of the search bar to show me what "search engine" I'm using.

Websearch is violating the principle of least surprise here out of the box. I agree the flexibility of using a custom delimiter (e.g. :) is nice. Could the delimiter be a configurable option, defaulting to whitespace? (Including possibly "no delimiter", the current setting, though I'm not totally convinced that's useful.)

The space should not be a part of the keymap, if this is by design, this is anti-human.

It took me a few days to finally find out that the pre-installed websearch config has a space after the term gg or so ...

I think this is the problem.

Of course, it is your code, your repo, and you have the decision to not change the design.

albertlauncher works perfect, if it does not have the whitespace special design.

@githorse thank you sharing your idea, but I think I'd better close this PR.

according to albertlauncher/albert#321

I've switched to Ulauncher, it does almost the same thing.

And the most important thing, it is user friendly. It does not have anti-human config design.

Why don't you just add a space to the trigger?

I have at least two websearches that do not use space. Do you think forcing people to use a space is better?

Why don't you just add a space to the trigger?

I have at least two websearches that do not use space. Do you think forcing people to use a space is better?

the problem is not about the PR.

this PR is wrong.

because it try to fix a thing already broken by design.

the trigger should not need a space. this is the problem.

we want the trigger be trigger , not trigger(space).

just like the below example: the tirgger keyword is @gemini, not @gemini(space). people will type space themself.

Your own design is inhumane and does not conform to regular usage habits. Why not reflect on it yourself instead of asking others?

Since this is your own open-source project, I won't say much more.

You can refer to https://support.google.com/chrome/answer/14886647?hl=en

Note that if your design is only for your own use, feel free to do as you like.

If it's for public use, making something inhumane is not right. However, this is understandable since it's your personal project. You have the final say and can decide your preferences.

Please refrain from commenting any further.

keep it simple, keep it stupid.

The design is not broken.

The freedom to choose any trigger, even eg three spaces, is intended.

I just wanted to make sure that I did not fundamentally misunderstand something.

What do you suggest to chnage such that it becomes humane but does not force users to have to add a space for eg a "=2+2" query?