Hacker News Reader

> Showing me where the link points when I hover the mouse over it is something so fundamental to browsing I just can't figure out why it was ever removed

Firefox 57 (and maybe previous versions, I can't remember) pops up a "temporary status bar" (for lack of a better phrase) with the link destination in it when you hover on a link (https://imgur.com/a/Vo7AZ).

> and maybe previous versions, I can't remember

Yes, this has been the way it's done for a very long time. It's such a basic feature that I don't think there was ever a version of Firefox without this feature in one form or another.

Hi, Screenshots developer here –

We've gotten similar feedback about the Save and Download buttons, we added the cloud icon in part to make that clearer. I feel like there's a ticket about it as well, but I not able to find it. We're planning on adding a Copy button, and we'll probably do some user research on the buttons generally to make sure the choice is clear.

We did have a full page screenshot feature, but it didn't make the cut for Firefox 56. It's back in Firefox 57.

If you have any issues or feedback on Screenshots, we track most of our work in GitHub and welcome new issues there: https://github.com/mozilla-services/screenshots

Thanks for the reply. Adding a "cloud" icon to a button called "Save" does not make it clearer, it makes it contradictory (does it save it or does it upload it?). It's very clear based on the design of the interface that your primary goal is to get people to upload their screenshots to your servers. The "save" button is the only one with a label. The others are icons only. The "save" button is 3x as big as the other two. The "save" button is the only button with any color (the others are black & white). Also note that the current "save" feature (which uploads the screenshot to mozilla servers) actually does the opposite of saving them because after 2 weeks they will be deleted.

I'm not convinced "user research" is the best way to go here. If the majority of your selected users think "save" should mean "upload to mozilla" does that mean they're right? And when they hit "save" in microsoft word or whatever, they think the document is going to microsoft servers?

"Save" means put it in my local storage.

"Download" means copy some file from a remote server onto my local storage.

"Upload" means copy it from my local storage to a remote server.

I also just realized that clicking the "save" button clobbers whatever is in my clipboard.

https://screenshots.firefox.com/oFTVibyVnGx7LPAl/news.ycombi...

After using it to actually upload that shot, I just realized it captured and shared much more than just the screenshot. It captured and shared the page title and domain of the page I took a screenshot of. That seems like quite the security risk given that I wasn't notified such was going to be shared. Imagine if I'm working on a dev server or a private place on a public server. It seems that when people want to share a screenshot, they want to share just the screenshot in many cases.

> I'm not convinced "user research" is the best way to go here. If the majority of your selected users think "save" should mean "upload to mozilla" does that mean they're right?

More or less, yes, if more people understand one design than another, then the one that more people understand is the right one. We're generally shooting for more than a majority though, we'd want most people to understand the interface, and would also take negative reactions into account.

In our past studies we found people generally understood the buttons well enough, and specifically it was obvious after first use what was happening. But we only compared the current buttons to one other option (which didn't compare well), so it wasn't a very thorough test.

There's more overhead in updating this interface than some others, since we have to get updated translations (to more strings than just the button) and change all our onboarding, which is why we've kind of left it be until we can look at the full set of buttons more closely.

Haha I am completely dismayed. So that means if I use File -> "Save Page As..." I can look forward to that eventually sending my entire page to your servers ;)

Again thanks for your replies and hopefully they can be more clear in the future (without being incorrect). And just in case I have sounded like a jerk, I really do love Firefox and I have been using it since it first came out as Phoenix. I appreciate the work you all put into it even if my above comments didn't make that clear. I actually used the Mozilla Suite before that, too. I still have, wear, and love my dinosaur head t-shirt I ordered from the mozilla store soooo long ago. https://upload.wikimedia.org/wikipedia/commons/thumb/7/74/Mo...

This seems to be one of those "dark patterns" in design, where they try to persuade someone to perform a certain action.

At the very least, change it to "Save to Cloud" and "Save to Disk." But to be really clear, "Upload and Share" is definitely the way to go.

> and specifically it was obvious after first use what was happening.

Sure, it's obvious after the first use, but if somebody is taking a screenshot of sensitive information, is the best way for them to find this out to show them that they've already uploaded said sensitive information to Mozilla's servers?

One of Mozilla's major value propositions is a privacy and security focus. Lots of people are interested in Firefox because it doesn't stream tons of data about browsing sessions upstream. It should always be obvious when a user is uploading information to the internet; the intuition on snipping/screenshot tools is that they perform actions locally.

I take screenshots of internal graphs, admin pages that may contain pseudo-private endpoints and admin account information all the time ("Logged in as $ADMIN_USER", account numbers with vendors, etc.). Sharing those in a private chat or email is a big part of my screenshotting use case. People screenshot a lot of private data.

I don't think Mozilla is wise in making "send your image to a public bucket protected only by the obscurity of its URL" the default behavior, at least not without making it obvious that this is what's happening. This sneaky exfiltration of private data is a large part of what people hate about Google and Facebook.

Maybe there could be an opt-out on first use? I get that your average grandma isn't going to understand the distinction and will want something she can share (but she's also probably just trying to get to email or Facebook anyway), but this is a real potential security problem for your users.

> if more people understand one design than another, then the one that more people understand is the right one.

Something something Henry Ford and faster horses. It's still bad design even if 50%+ε understand save as 'upload to some server' and not the old meaning that predates Windows 95 which is to save it to local storage. You shouldn't be taking liberties whenever you are about to upload any bit of information from your users, instead you should obtain informed consent from all users and not just from better informed ones that can parse the meaning from the cloud icon. The UI is wrong until only a moron in a hurry[1] would be mistaken about what your software is about to perform. Otherwise how are you any better than that other company that starts with a G and ends with oogle.

[1] https://en.wikipedia.org/wiki/A_moron_in_a_hurry

"More or less, yes, if more people understand one design than another, then the one that more people understand is the right one. ", he posted on the one website in the world that still holds on to the original meaning of the word 'hacker'.

I'm on Ubuntu on a Thinkpad that has the trackpad disabled. Plain Jane Microsoft wired USB mouse plugged in.

I use the right-click/context menu a fair amount. Since FF screenshots was added, it gets triggered frequently when I'm popping the right/click menu on a page -- without my making any click after the right-click to open the context menu. And with the trackpad disabled and using a mouse with a physical button and a rather distinct click, I can say this with considerable confidence.

So, then I have to to hit the escape key to kill the FF screenshot functionality. Which is becoming muscle memory, at this point.

I too found this dialog to have bizarre choices for function and iconography. My main gripe with it is that there are now two "Save" actions in Firefox that have completely different functionality (Save a page vs Save a screenshot). I found this to be a very odd choice.

I was pretty pissed it automatically uploaded the image. "Save" refers to saving to your local machine. "Upload" means uploading to a remote server. Fix the damn button.

On your second point, I'm running the 57 beta and am not sure what you're talking about.

It's true that there is no "always on" status bar at the bottom. However, when I hover over a link, its URL does appear in the bottom-left corner. Exact same behavior as Chrome.

Thanks to you and others for pointing this out. I've been using status 4 evar addon since they removed the status bar. I think the original behavior when they removed the status bar was to have a tooltip popup with the URL, but looks like it has changed since then. Does it popup immediately or is it like the tooltip where you have to hover for a couple of seconds before it shows up? I'll try removing my addon to check. Note that the addon allows me to configure how long the delay is (current setting=0ms) and how long it stays there after I move my mouse.

It takes a fraction of a second to pop up (maybe a few hundred milliseconds).

Yeah just confirmed by removing the addon. The url shows up but only after an annoying delay and an animated fade-in. Ugh.

Flickering in and out instantly as you move the cursor across a page and happen to pass through a bunch of links would be a lot more annoying.

Alternate implementation would be to make sure the cursor has actually stopped, which still needs a frame or two before you act on it.

I could definitely do without the fade-in, that's inconsistent with the rest of the OS and it feels clunky. I wonder if the designers are on a different OS where this is normal for tooltips?

Just did some browser comparisons:

Safari has a fade-in and triggers it immediately, so you do get a pop-up as you move your mouse around. But the fade and delay on closing smooths it out enough that it doesn't flash on and off. I don't love that it pops up instantly when I don't want it to.

Chrome and Firefox feel pretty similar. There's a delay on fade-in and a (longer?) delay on fade-out. Looking at it more closely, I think it does check that the cursor has stopped. Moving left to right across a link won't trigger the popup even if you're over it for a long time. So is it just the fade that makes opening feel slow?

I'm not totally sure whether I'd like it better without the fade. Since I don't see a way to turn it off there's no easy way to find out.

I don't know what your delay is set to, but for me it's faster than I can move my eyes from the link to the bottom of the screen.

What model of computer are you using (and what operating system)? For me it is practically instantaneous.

was super shocked when the button uploaded to the servers.

The straightforward way most people use screenshots nowadays is to paste in Slack/Messenger/Google doc. Seems like a no-brainer for the default to be "copy to clipboard", but instead it's "upload to servers with no indication of this on the first click".

Thank you for pointing this out. This is a bug IMHO. No security-focused browser to do that.

honest question, why would anyone use an in-browser screenshot tool when screenshot tools already come pre-installed in basically every operating system?

can Firefox's screenshot tool be disabled? I'd like to switch back from Chrome due to Big Brother concerns but not if Firefox is making questionable choices like this.

Because that allows you to capture a complete, scrollable website as a single png image. OS tools only let you capture the currently viewable area.

ah! thank you, that makes a lot of sense. I use Evernote's web clipper for similar functionality.

Well, once you learn the backwards behavior of the buttons, nothing will get sent to mozilla (click "download" and not "save"). I thought it would be cool to save the entire page, even the parts that aren't in view (difficult with operating-system screenshot tools). (I couldn't figure out how to make this work.) And yes, you can right-click on the screenshot button and remove it from the toolbar (I don't think this would disable the entire feature, maybe there's a keyboard shortcut to activate it as well).

If you consider Firefox Screenshots from the perspective of a Support tool, this makes alot of sense. Customer wants to file a support ticket or bug report, needs to include a screenshot. Screenshot taken, they paste the link into the ticket, done. They never leave the browser.

Outside of the browser, fire up the screenshot tool, crop whatever you want to show, save it, upload it to imgur or website of choice, attach or paste it.

I would liken the tool to Fedora's amazing fpaste tool, which you invoke via the following:

    foo | fpaste

fpaste takes the output of the `foo` command, sends it to a log file, automatically uploads it to a fedora server, and returns a minified URL that can be copied and pasted to the developer. It is an amazing tool, provided you know what it's doing.

fpaste also takes several safety precautions such as static URLs that delete themselves after 24 hours.

So I can see Screenshots being very useful for that purpose. It just needs to be better about disclosing what it's doing.

    cat verysecret.csv | cut -f 1 -d ' ' |fpaste -z -x -a

Wait, I meant paste.

You can try it for yourself in the browser: https://paste.fedoraproject.org/

If you accidentally fpasted something sensitive, you're given a deactivation link to remove it immediately.

This is one of the best features of new Firefox for me. Normally I had to start up gimp, select an option, set few seconds timeout, grap the screenshot, and then edit it to show only the part I wanted.

Screenshot in firefox is aware of the page, you can select the appropriate html tag visually.

Actually I thought they removed it in recent Nighlty and was looking for it, after few days I found it in the triple dot, (just next to the start and pocket icons).

Note that there are a ton of ways to grab screenshots on linux that don't require you to start gimp. (I'm assuming you're using linux since you mentioned gimp; yes I know it runs on other platforms, but those platforms also have other, much easier ways to capture screenshots). On linux workstations I usually have one of those quake-console things so I hit whatever key combination triggers it (on my mac it's double-ctrl, brings down iterm 2 window). From any console you can use the "import" command (part of imagemagick package). This lets you grab the whole screen, a single window, or an arbitrary rectangle. `import foo.png` Gnome has its own screenshot tool, and there are likely dozens of other packages that do the same thing.

Screenshots, what.. ? Is this even if you're not signed into Mozilla?

I remember recent HN submissions about Mozilla sharing information with Google...

>it could screenshot the entire page

I saw that too, but it seems the only way to do it is the old ctrl+shift+i and use the dev screen shot tool after enabling it in the toolbox. This method does always save locally though.

I've tried Developer's Edition, but gave up when I saw that Vimperator would not work on it.

Hopefully someone from Mozilla reads this and takes notes. I completely agree with you on this one.

File a bug.

I would not get my hopes up on the second part. Mozilla seems to be living in a "we know better" bubble these days, alongside Gnome and certain other big name FOSS projects...