This autocomplete suggests domains that don't exist. You can just type garbage and it will suggest something, but then if you go there, there are no records.
It seems like one purpose of an autocomplete box is help you avoid typos, so that makes it less useful.
Yes, it will autocomplete the most popular domains first, then known registered domains, and then fall back to appending popular TLDs.
The reason is that it's impossible to enumerate all domains. gTLDs (like .com, .info) mostly publish their zone files, but ccTLDs (like .cn, .de) usually don't.
But maybe you're right. After reading these comments, it's clearly confusing to people.
There are limits to how accurate you can make this. Zones like .xyz renew at 18%, so you'd expect 0.2% of those domains to expire without renewal each day. The tranco list is based on a 30 day look-back and I've seen a small percent of those domains lack name servers, even for the latest list.
Similarly, domains can be registered since the last time you downloaded your lists. So you never have a complete or accurate list.
I think it's perfectly reasonable to suggest names that have recently expired or domains that could have been recently registered. The alternative requires an NS lookup.
The autocomplete seemed to suggest real domains to me. If you typed in garbage it suggested the garbage plus a bunch of common TLDs - that seems like a reasonable choice to me. In any case, if your issue is the list of domains to suggest, its trivial to change that to a different list.
This behavior seems to happen only when there is no more words in the list that can complete the current string. Otherwise, it shows only existing domains.
I guess this behavior helps to autocomplete the domain extension for less-known domains that are not in the search list.
That's also the old behaviour of browsers and the current one of text browsers. I fail to see how this is wrong.
$ lynx doesnotexist2
Looking up doesnotexist2 first
Looking up www.doesnotexist2.com, guessing...
Looking up www.doesnotexist2.edu, guessing...
Looking up www.doesnotexist2.net, guessing...
Looking up www.doesnotexist2.org, guessing...
Can't Access `file://localhost/home/user/doesnotexist2'
"That's also the old behaviour of browsers and the current one of text browsers."
As a daily text-only browser user for over three decades, using it as a "daily driver", I'm no doubt biased. I was a lynx user before links was released. I would never go back to using lynx. When I see HN comments suggesting lynx as an example of text-only browsing I question whether the commenter is actually using it as "daily driver", or looked at the source code and compared it with alternatives
links doesnotexist2
does not trigger a DNS lookup
Unlike elinks, w3m, etc.
One of the many reasons I prefer links (no graphics)
I use links as an offline HTML reader; I am opening local HTML files every day
That said, it's relatively easy to modify these programs to behave according to personal preference; they are relatively quick and easy to compile
One could remove the unwanted DNS lookups, for example
Contrast this with the so-called "modern" browser that almost zero users have the time and resources to modify and compile themselves
Instead, we see dissatisfied browser users beg browser vendors to make changes
To no avail
There are other ways to prevent unwanted DNS lookups besides modifying the source code
For example, I run own authoritative DNS on the loopback, with custom root.zone
All lookups by default return the same address: the address of the local forward proxy
The proxy's memory contains DNS data precollected periodically in bulk from a variety of sources. At startup/restart, the proxy loads the data into memory from a plain text file (a "map")
The map would not contain any DNS data for "doesnotexist2"
There is no way for DNS lookups to leave the loopback nor any way for HTTP requests to be sent to destinations other than the ones I approve
No Javascript engine in text-only browser
For use with "modern" (read: uncontrollable) browsers the proxy can deny requests for URLs ending in .js, add a Content-Security Policy response header that disables Javascript and, if the CSP were to be ignored, can rewrite pages to comment out or delete inline JS
What I mean is if "modern" browser vendors change their popular browsers to make it more difficult for users to control others' Javascript, we can use URL blocking, URL rewriting and/or page rewriting to modify or delete inline or sourced Javascript and/or substitute the user's own Javascripts
I hate this behaviour because it means if I typo an internal hostname suddenly it gets broadcast to everyone's DNS resolver, and wastes some time in the process (and in the worst case, takes me to a random website I never intended to visit). I always have to rummage around in about:config or whatever to disable it, because that behaviour is never what I expect nor intend.
Using keyup makes no sense and is inconsistent with user expectations. For triggering actions (which includes normal typing), you only ever use keydown. (Well, there’s one exception for reasons unclear to me: activating a button by pressing Space. That triggers on keyup like how clicks are on release, while Enter triggers on keydown.) Keyup is limited to things where you’re constantly reacting to the state of a key, as is common in games.
This affects the functionality, too. It is in fact introducing latency by using keyup instead of keydown. Feels bad.
For triggering actions (which includes normal typing), you only ever use keydown.
I think you have this wrong, actions generally occur on button release, until that you can move the mouse cursor to a different target, tab to another control, use [ESC] to cancel, and so on. Typing, moving a slider with the cursor keys, and similar things that make use of key repetition while holding down the key are the exception to this.
I am speaking specifically of keyboards. For mice, it’s release. (When I made a similar comment two months ago on Lobsters when this article came up there, I mentioned this, but decided to leave it out this time, thinking the context was clear enough. I think I was wrong.)
If it matters, it seemed pretty clear to me. You're also explicit about it, so I'm not sure how one would think you were referring to mouse input. Unless people type by clicking on a virtual keyboard? Benefit of the doubt can only stretch so far.
Can you give some specific examples? In my experience (which I also verified to make sure), all of the things you mention act on keydown. The exceptions are on multi-key combinations, which follow a special key that is held down while modifiers are entered, where the action is completed by releasing this special key (ALT + TAB [...TAB | SHIFT]) being a common example.
TAB is otherwise immediate, as is ESC, and the examples of typing, as well. So I'm left wondering, how do they have it wrong?
Non latin characters are sometimes inputted by holding down a key. For example on my keyboard holding down z shows ź,ž and ż with a corresponding number to select which one I would like to type. I haven't checked but one would assume if you're detecting keydown the input at that time will be z, and keyup would record ž (or whatever was selected).
Physical keyboards and virtual keyboards are different. Traditionally, ź and ż (my keyboard doesn't have ž) are inputted on keydown X or keydown Z while Alt modifier is active.
Key down is used for events that can be duplicated by holding the key. Eg in a text editor, when you want multiples of the same letters, you’d press and hold the key.
Key up is used for when it’s important to only have one occurrence of that event.
Technically you could write code that made key down only react once. But the logical separation makes some sense.
Examples where key down is used as an input event?
Basically anything that requires text input from OS-native text input fields to IDEs, word processors, spreadsheets, web browsers, terminal emulators and so on and so forth.
Your message was ambiguous to me. But I will admit that I was also quite hungover so it’s possible that the error was mine.
In answer to your question: I’ve only really used keyup for either one time events (like keyboard shortcuts for UI elements) and for ending the status of repeated events (eg movement in games, dragging UI elements, etc).
I think if you were to search for key up events in GitHub, most of references might likely be games and other visual interactions rather than text entry.
As a user, I’ve not thought too much about this before now. I agree with you mostly, but the keyup on space behaviour actually feels so innate I’d hate any change to it. Keydown on space is “jump”; nothing else.
I’d just like to hear an explanation of why it is. Because if you handled “is key currently pressed” events with polling rather than events (which is how people almost always consume them logically), it would be the only thing in the entire world that I can think of that would/should use keyup.
Edit: I think people are misunderstanding me. I’m asking for an explanation of why, when you have a button focused, pressing Space activates on key up, rather than key down like literally everything else on the keyboard, including Enter on a focused button. (Mouse activates on release.)
Because in those contexts, the Space key is being used as a physical mirror of the on-screen button, and on-screen buttons are UX'd to perform action when released.
These two interactions should behave identically:
- Hover over a button, press mouse button 1, release button 1.
- Tab over to a button, press spacebar, release spacebar.
Why on-screen buttons are UX'd to behave this way is a logical follow-up question. I'd wager that it gives a means to bail out of the clicking (e.g. by moving the mouse out of the on-screen button before releasing mousebutton 1, or by pressing Esc while having spacebar still depressed).
Yes, the bailing-out explanation is right. With a mouse it's intentional that you should be able to push the pointer off the button with the mouse button still pressed to cancel.
Under Linux, Firefox, Chromium and GNOME Web (which is WebKit-based; a.k.a. Epiphany) all have Tab cancel but not Esc.
The GTK 4 behaviour feels very much like a bug. I can reproduce it in gtk4-demo and in gtk3-demo. It leads to obviously-nonsensical behaviour, including messed up window focus. I’m tempted to file a bug report, but haven’t had such productive experiences with filing things on GNOME historically as with Firefox or even Chromium.
Why should Enter being a different key be a different discussion? They’re both used to activate buttons, and one is behaving like every single other keyboard interaction, while the other isn’t. (Related: if a link is focused, Enter activates it, Space doesn’t, just doing the normal scroll-down-one-page.)
It just matches real life. What about lmb to shoot? You don't fire a gun on trigger release, you fire it on trigger down. Walking? Would be weird moving only on keyup.
This raises the question of why real life buttons and virtual buttons behave differently. My unsubstantiated guess is that clicks act on release to give the opportunity to slide off the button to abort, and/or because the button would disappear while pressing if it ex. submits a form.
Games use keydown because it is more responsive and control responsiveness is vital to dynamic gameplay feeling good. There are some exceptions - if action can be charged, then it is on keydown. And then on gamepad sometimes actions are on release when developer has too many actions and not enough buttons, so it can be keyup vs hold.
I may be wrong on this, but I think that pressing esc while space is held down cancels the button press, like moving the mouse off the key while holding down the click does. I'd need to test this, though.
That’s where the input event is really useful as it allows for more input options than just keyboard and then you don’t have to deal with figuring out if it’s down or up.
On my own computer I can open a context menu and select an action with a single click. Whenever I need to use MS Windows, I always need two and I find that infuriating every time.
Unfortunately this approach doesn't feel that great down here in Australia, definitely a function of latency.
I think you could get a lot closer by framing this as an optimization problem, where you use the full alphabet dictionary, but add a residual prediction which aims to cover as much of the remaining domain name tree as possible weighted by popularity. This tree could then be pre-baked and stored with the same system. This would probably get you p99 0ms even in Australia.
I think most people search for domains they own, which doesn't correlate with Tranco popularity ranking. Treating 'popularity' not as a function of visitors, but as a function of number of known domain names with that prefix could work, though.
Why not just trigger the fetch on keyDown and show it as soon as the response arrives, as usual?
The time it takes to press a key is a reasonable target to aim at for API latency I suppose, but it is still an arbitrary target. Waiting to display until keyUp just adds more latency if your API is faster. Having it synced with keyUp doesn't make it feel more immediate to me.
If you’d like to reduce the network latency further you can store each trie node as a file, naming it conveniently the prefix path to that node. Then dump the few hundred million files onto R2.
Now the traversal can be done completely via CDN lookups!
Cloudflare R2 doesn't distribute files globally. They only cache requested files in PoPs. So then it would function the same as an API behind the Cloudflare CDN.
It would replace the Cloudflare to server network hop with a cloudflare PoP to R2 hop. That covers approximately the same geographical distance and results in similar latency.
3B existing URLs extracted from CommonCrawl, with instant search results. It was a fun project but didn't serve much real-world purpose besides curiosity and discovery. So I eventually ditched it, primarily because the link DB was a whopping 500 GiB in size, too much to just keep hosting.
I just used SQLite FTS5 as the backend search engine. Just a few lines of code, but immediate response from a 0.5 TiB DB. SQLite is amazing.
The perceived latency starts from keydown, not keyup. Redefining latency to start at keyup reduces measured latency, not perceived latency, and delaying the visual display to keyup makes perceived latency strictly worse, not better. Even sticking with the keyup definition, just displaying the result as soon as it is available gives the possibility of negative (defined) latency.
I think you don't fully understand. They aren't just doing a search based on the key being pressed down. They already did a search based on the previous characters that returned results for all possible next characters. So by the time you type a second character it just checks results locally from the search that had likely already been returned from when you had typed the previous character.
There are a large number of edge cases where rendering it on key down would not be desirable (user types Ctrl+v, how accented characters are enetered, how some Asian languages get handled, backspace/delete, user holds a key down, etc). This is getting into the point of optimization where it simply does not matter and can actually harm things
Holding a key would indeed require both on key down and up. I'd still argue that handling rerender on key down (or to fully match whatever OS/browser semantics, oninput) makes more sense than key up. Pressing Ctrl+V for example produces feedback as soon as "V" is pressed, not when Ctrl and/or V is released.
Interestingly Ctrl+V on OP's website does update immediately when Ctrl+V is pressed.
What autocompletes are you testing? I set my repeat delay to be very long so it doesn't interfere with the test, and then went to google and held down some keys for half a second. It types the letter and autocompletes long before I let go.
Every UI I can find types the letter on keydown, and I don't know why you'd want typing and autocomplete to use different timings.
Waiting for release is for mice and touch, not keyboards. And that's only because they have to deal with dragging.
I just typed a random sequence of the characters, long enough to be certain such domain doesnt exist. No only, the browser send an autocomplete request for every keystroke but for each request it returned a set of proposed domain names (which I'm 100% certain doesnt exist). At this point, how do I understand which results are legit and which are fake?
Also, it would be nice to highlight the typed part in the result set so I can visually see what matches exactly.
For the nonexisting domains, it seems like it only autocomplete with the possible domain extensions, (e.g .com, .org, ....) as the search list is non-exhaustive. But it could indeed be improved by not sending autocomplete requests anymore.
You're not the first to mention this, so clearly users expect something different from what I've designed it for. I think I'll drop the TLD postfix suggestions.
On a technical level yes but once you factor in monitor refresh rates then you can get to levels of optimization where it simply doesn't matter because you are constrained by waiting for the refresh rate anyways.
It seems like one purpose of an autocomplete box is help you avoid typos, so that makes it less useful.
Yes, it will autocomplete the most popular domains first, then known registered domains, and then fall back to appending popular TLDs.
The reason is that it's impossible to enumerate all domains. gTLDs (like .com, .info) mostly publish their zone files, but ccTLDs (like .cn, .de) usually don't.
But maybe you're right. After reading these comments, it's clearly confusing to people.
Similarly, domains can be registered since the last time you downloaded your lists. So you never have a complete or accurate list.
I think it's perfectly reasonable to suggest names that have recently expired or domains that could have been recently registered. The alternative requires an NS lookup.
Can't you just append any TLD to anything and get a domain that "might" exist? Or am I missing something?
I guess this behavior helps to autocomplete the domain extension for less-known domains that are not in the search list.
speedtest.me.uk exists, but typing speedtest.me does not suggest it, and speedtest.me.u suggests speedtest.me.us first (doesn't exist).
(.me.uk is a second level domain category in .uk, intended for personal domains.)
As a daily text-only browser user for over three decades, using it as a "daily driver", I'm no doubt biased. I was a lynx user before links was released. I would never go back to using lynx. When I see HN comments suggesting lynx as an example of text-only browsing I question whether the commenter is actually using it as "daily driver", or looked at the source code and compared it with alternatives
does not trigger a DNS lookupUnlike elinks, w3m, etc.
One of the many reasons I prefer links (no graphics)
I use links as an offline HTML reader; I am opening local HTML files every day
That said, it's relatively easy to modify these programs to behave according to personal preference; they are relatively quick and easy to compile
One could remove the unwanted DNS lookups, for example
Contrast this with the so-called "modern" browser that almost zero users have the time and resources to modify and compile themselves
Instead, we see dissatisfied browser users beg browser vendors to make changes
To no avail
There are other ways to prevent unwanted DNS lookups besides modifying the source code
For example, I run own authoritative DNS on the loopback, with custom root.zone
All lookups by default return the same address: the address of the local forward proxy
The proxy's memory contains DNS data precollected periodically in bulk from a variety of sources. At startup/restart, the proxy loads the data into memory from a plain text file (a "map")
The map would not contain any DNS data for "doesnotexist2"
There is no way for DNS lookups to leave the loopback nor any way for HTTP requests to be sent to destinations other than the ones I approve
No Javascript engine in text-only browser
For use with "modern" (read: uncontrollable) browsers the proxy can deny requests for URLs ending in .js, add a Content-Security Policy response header that disables Javascript and, if the CSP were to be ignored, can rewrite pages to comment out or delete inline JS
What I mean is if "modern" browser vendors change their popular browsers to make it more difficult for users to control others' Javascript, we can use URL blocking, URL rewriting and/or page rewriting to modify or delete inline or sourced Javascript and/or substitute the user's own Javascripts
This affects the functionality, too. It is in fact introducing latency by using keyup instead of keydown. Feels bad.
I think you have this wrong, actions generally occur on button release, until that you can move the mouse cursor to a different target, tab to another control, use [ESC] to cancel, and so on. Typing, moving a slider with the cursor keys, and similar things that make use of key repetition while holding down the key are the exception to this.
TAB is otherwise immediate, as is ESC, and the examples of typing, as well. So I'm left wondering, how do they have it wrong?
Because they listed the exceptions to the "act on keyup" rule
Key up is used for when it’s important to only have one occurrence of that event.
Technically you could write code that made key down only react once. But the logical separation makes some sense.
Basically anything that requires text input from OS-native text input fields to IDEs, word processors, spreadsheets, web browsers, terminal emulators and so on and so forth.
(I confess I’m getting frustrated at the ridiculously large fraction of messages in this thread that have somehow completely misunderstood things.)
In answer to your question: I’ve only really used keyup for either one time events (like keyboard shortcuts for UI elements) and for ending the status of repeated events (eg movement in games, dragging UI elements, etc).
I think if you were to search for key up events in GitHub, most of references might likely be games and other visual interactions rather than text entry.
But for reference browsers fire the keypress event on keyup
https://w3c.github.io/uievents/#event-type-keypress
Edit: I think people are misunderstanding me. I’m asking for an explanation of why, when you have a button focused, pressing Space activates on key up, rather than key down like literally everything else on the keyboard, including Enter on a focused button. (Mouse activates on release.)
These two interactions should behave identically:
- Hover over a button, press mouse button 1, release button 1.
- Tab over to a button, press spacebar, release spacebar.
Why on-screen buttons are UX'd to behave this way is a logical follow-up question. I'd wager that it gives a means to bail out of the clicking (e.g. by moving the mouse out of the on-screen button before releasing mousebutton 1, or by pressing Esc while having spacebar still depressed).
You’re only answering half of the inconsistency: because Enter activates buttons like any other key, on keydown.
Qt: Esc cancels. Tab cancels.
Gtk-4: Spacebar auto-clicks before keyup (probably due to auto-repeat).
Win32: Esc sends Esc to the window. Tab cancels.
Firefox: Esc does nothing. Tab cancels.
So, mileage does vary.
Re. Enter key, that's a different key, so different discussion.
The GTK 4 behaviour feels very much like a bug. I can reproduce it in gtk4-demo and in gtk3-demo. It leads to obviously-nonsensical behaviour, including messed up window focus. I’m tempted to file a bug report, but haven’t had such productive experiences with filing things on GNOME historically as with Firefox or even Chromium.
Why should Enter being a different key be a different discussion? They’re both used to activate buttons, and one is behaving like every single other keyboard interaction, while the other isn’t. (Related: if a link is focused, Enter activates it, Space doesn’t, just doing the normal scroll-down-one-page.)
> if a link is focused, Enter activates it, Space doesn’t
It seems that you answered your own question: different keys, different purposes.
Enter is for "execute the focused action". Spacebar is for "operate the button".
So this could be the oddity
I think you could get a lot closer by framing this as an optimization problem, where you use the full alphabet dictionary, but add a residual prediction which aims to cover as much of the remaining domain name tree as possible weighted by popularity. This tree could then be pre-baked and stored with the same system. This would probably get you p99 0ms even in Australia.
I think most people search for domains they own, which doesn't correlate with Tranco popularity ranking. Treating 'popularity' not as a function of visitors, but as a function of number of known domain names with that prefix could work, though.
The time it takes to press a key is a reasonable target to aim at for API latency I suppose, but it is still an arbitrary target. Waiting to display until keyUp just adds more latency if your API is faster. Having it synced with keyUp doesn't make it feel more immediate to me.
Now the traversal can be done completely via CDN lookups!
(I'm the author btw)
3B existing URLs extracted from CommonCrawl, with instant search results. It was a fun project but didn't serve much real-world purpose besides curiosity and discovery. So I eventually ditched it, primarily because the link DB was a whopping 500 GiB in size, too much to just keep hosting.
I just used SQLite FTS5 as the backend search engine. Just a few lines of code, but immediate response from a 0.5 TiB DB. SQLite is amazing.
> And on keyUp (the user releases the key), we render the suggestions.
Interestingly Ctrl+V on OP's website does update immediately when Ctrl+V is pressed.
Every UI I can find types the letter on keydown, and I don't know why you'd want typing and autocomplete to use different timings.
Waiting for release is for mice and touch, not keyboards. And that's only because they have to deal with dragging.
The letter appears instantly, then after a moment a lot more copies of the same letter start showing up rapid-fire.
What are you calling flawed about my test?
This happens on both Windows 10 and Linux Mint, in Firefox.
I never meant to infer that the keyup event is specifically what should be listened to
How do you know if it's necessary? If Tranco returns less than eight?