Firefox does it again



  • Chrome, Firefox, Windows . . . . . . Wow.  Is there some sort of agreement among all software developers that they are going to keep making their products worse and worse?

    So, a few days ago my wife says "My Firefox is broken".

    OK, what's wrong.

    "When I right click on a link, there's no "Send Link" for me to click on."  (Automatically opens up your email client and pastes in the link)

    I look at it, and she's right.  Oh well, I don't know what to tell you.  (Not the answer she wanted to hear.)  At the time I didn't know she had upgraded to the latest version.

    Then I found this:  https://bugzilla.mozilla.org/show_bug.cgi?id=239307  Originally submitted in 2004 and finally implemented in November of this year.  Despite a significant amount of feedback from people saying that they really liked this feature, the Firefox developers, in their infinite wisdom, declared "we need to clean up the context menus and 'sophisticated' users rarely uses this."  And "Chrome and IE don't have this feature, so what's the problem?".  And they suggested an extension people could use to get this function back, except for the one little problem that it doesn't work.

     



  • Just tell her she "fears change". That's what all the cool kids do.



  • Well, in some way they are right on trying to clean up vanilla Firefox since it has a pretty good plugin support, althought I don't see why this particular feature could bring so much trouble, specially when there's people specially interested enough on it to open a ticket and discuss it online.

    Anyway, now you have a weekend project!



  • I'm just wondering why they didn't simply add an option to the config to enable/disable this?



  • @Rhywden said:

    I'm just wondering why they didn't simply add an option to the config to enable/disable this?

    Because "it's a lightweight browser". You can't even change the keyboard shortcuts without an extension.



  • It's madness

    Somebody just filled a bug asking the developers to remove functionality, without any reasoning about why they should do so, assigned it to a developer without permission, and it gets removed!

    That is very funny:

    "<label for="priority" accesskey="i">Importance</label>:

      --
       enhancement<span id="votes_container">
      with
      10 
            votes
    (vote)</span>"</p><p>There were more than 10 people complaining about it when they decided to implement the "enhancement". <br></p>


  • Well, I admit I'm one of the people who hates the clutter on the context menus, and the Send Link / Send Image in particular, *especially* because they are placed immediately next to Save Link / Save Image, and it's easy to click the wrong one. 



  •  To be fair, I never even noticed them there. But as far as clicking on the wrong menu item goes, creating a new email for you to send is at the harmless end of the scale. Now that it's removed, there is probably something worse at the same place.

     Anyway, I was complaining about the process, not the result. A bug report claiming that you must remove a feature because it confuses the users, because it puts a burden on developmen, or because it is just plain ugly... all of those are ok. A bug report claimingthat you must remove a feature because, well, I can't think about any reason, but here goes a link to a forum where I posted that it must be removed, and everbody disagreed... That should be only laughing material.

    I'm now tempted to troll the Firefox bugzilla, reporting something like:

    Remove unnecessary DHCP requests

    importance: enhancement

    description:

    The requests the browser makes every time the user visits a page.

    There is no loss of functionality, since resolving the page URL can be done by the following alternative procedure:

    1 - Open a terminal

    2 - Type the command "host pagename.example.com"

    3 - Select the resulting IP address and copy it with CTRL + SHIFT + C

    4 - Paste the address into the address bar of the broser and press ENTER

    Here is a related bug outlining the Firefx Development Team policy of fighting bloat at all costs: https://bugzilla.mozilla.org/show_bug.cgi?id=239307.

    Reproducible: Always

    Steps to reproduce:

    1 - Type an URL into the address bar using the site name instead of the IP address.

    2 - Press ENTER.

    3 - Cry as your bloated browser take hundreds of miliseconds just to translate it into an IP address.

     

    There, it is more complete than the real one, but I think the Firefox people won't mind it. The only reason I won't troll the bugzila is that there is no originality in doing that. Somebody else already had the same idea.



  • Except Mozilla has actual usage data, and the people griping on the bug report do not.

    @Mozilla said:

    item %people_used %mean_among_users %median %top5pct

    sendimage 3.34 1.93 1 3

    sendlink 0.89 1.4 1 4

    sendpage 0.89 2.93 1 31

    @Rhywden said:

    I'm just wondering why they didn't simply add an option to the config to enable/disable this?

    Because every checkbox you add to a program doubles the QA time. That said, Firefox is open source and therefore probably not QAed at all, but in theory if it were, this would be a smart move.



  •  Being a web site developer, I look at the page source frequently. A few versions ago they removed the "View Page Source" from tehe menu it was on. Took me a while to find it down in a context pup-up menu. And on some pages it's hard to find a context that shows it (e.g. if the entire page is a video). Why? Who knows? Who else cares? But now you can see why I don't want monthly updates.



  •  Wow, this menu item's removal dominates the google results for mozilla context menu usage stats. (at least tthe results the all-knowing-search-engine thinks are relevant to me)

     I was searching for the source of those stats, but then, meh, I gave up.

    I just wanted to know what to expect of the fate of other options I personally find useless: Back, Forward, Reload, Bookmark This Page, Save Page As...

    I simply can't be bothered to go ahead and customize the menu myself.



  • @AndyCanfield said:

     Being a web site developer, I look at the page source frequently.

    and you don't use ctrl+u to do that? ccc some developer you are.



  • @AndyCanfield said:

     Being a web site developer, I look at the page source frequently. A few versions ago they removed the "View Page Source" from tehe menu it was on. Took me a while to find it down in a context pup-up menu. And on some pages it's hard to find a context that shows it (e.g. if the entire page is a video). Why? Who knows? Who else cares? But now you can see why I don't want monthly updates.

     

    Don't mind. Firefox's view source doesn't actualy show you the page source anyway.

     



  • It clearly says it was removed from the context menu, not removed altogether.

    Check your firefox menu.

    Firefox 17

    I hate the changes Mozilla have made to the UI since Version 6, but for once, I agree with their decision on this.


  • Trolleybus Mechanic

    @LoremIpsumDolorSitAmet said:

    Check your firefox menu.
     

    Not the same. That emails the link of the current webpage. If I recall correctly, the other context-menu item emailed the link of the item clicked on.

    Or I could be wrong.



  • Hmm... Perhaps. But then you could just click the link and repeat the above.

    Besides, I personally prefer having the absolute control and discretion of copying and pasting URLs manually. Anyone who uses websites' own 'share/email this link' tool are asking to get spammed. Yes, you should be able to trust the browser not to do this, but old habits die hard.

    PS. Hey, I used to think people made it up, but yes, CS does sometimes delete two characters when I press backspace once!



  • @blakeyrat said:

    Because every checkbox you add to a program doubles the QA time.

    Not being familiar with proper QA (please don't ask *sniff*): why? Is it because then you need to test everything with the checkbox turned off, and everything with the checkbox turned on? The math seems solid but surely there are better ways to go about testing than try every possible inputs?


  • Trolleybus Mechanic

    @LoremIpsumDolorSitAmet said:

    Hmm... Perhaps. But then you could just click the link and repeat the above.
     

    You could-- but you've just doubled the amount of work an end user has to do, AND forced them to load a page when they don't have to (which requires them to go BACK a page to continue whatever they were doing).

     



  • @LoremIpsumDolorSitAmet said:

    Hmm... Perhaps. But then you could just click the link and repeat the above.
    But what if the link you want to send is to goatse? Have you thought of that?

    @toon said:

    Not being familiar with proper QA (please don't ask *sniff*): why? Is it because then you need to test everything with the checkbox turned off, and everything with the checkbox turned on? The math seems solid but surely there are better ways to go about testing than try every possible inputs?"
    No, it doubles QA time for that particular functionality. So, instead of simply checking if the menu item is there, you'd also need to check for the menu item being there when set to visible on, and not when set to not visible. Hence, double QA time for that menu item.

     

    Or if you weren't checking for the menu item being there in the first place, because, you know, no one cares, and now you'd have to check both cases: that's infinitely more QA time.

    Seriously though, you'd be going from:

    - check if the menu item is there

    - check if it's actually working

    to:

    - check if the checkbox is there and correctly setting the corresponding preference

    - check if the option is there, or not, depending on the preference

    - check if the menu item is working

    That's approximately double the QA time, arguably more, allotted for that menu item.


  • ♿ (Parody)

    @Zecc said:

    @LoremIpsumDolorSitAmet said:
    Hmm... Perhaps. But then you could just click the link and repeat the above.

    But what if the link you want to send is to goatse? Have you thought of that

    I think people sending goatse should absolutely be forced to look at it every time they try to email the link.



  • @boomzilla said:

    I think people sending goatse should absolutely be forced to look at it every time they try to email the link.
     

    In order to reinvigorate them, no doubt.



  • @dhromed said:

    @boomzilla said:

    I think people sending goatse should absolutely be forced to look at it every time they try to email the link.
     

    In order to reinvigorate them, no doubt.

    I suddenly wonder if the guy who built that site back in the day, made a huge buck out of that, and if so, whether or not he's flaunting it.


  • Trolleybus Mechanic

    @toon said:

    I suddenly wonder if the guy who built that site back in the day, made a huge buck out of that, and if so, whether or not he's flaunting it.

     

    I'm sure that information is available. Let me Google that for y--- ooooo, nice try. Nice try.

     



  • @Zecc said:

    @LoremIpsumDolorSitAmet said:

    Hmm... Perhaps. But then you could just click the link and repeat the above.
    But what if the link you want to send is to goatse? Have you thought of that?

    @toon said:

    Not being familiar with proper QA (please don't ask sniff): why? Is it
    because then you need to test everything with the checkbox turned off,
    and everything with the checkbox turned on? The math seems solid but
    surely there are better ways to go about testing than try every possible
    inputs?"
    No, it doubles QA time for that particular functionality. So, instead of simply checking if the menu item is there, you'd also need to check for the menu item being there when set to visible on, and not when set to not visible. Hence, double QA time for that menu item.

     

    Or if you weren't checking for the menu item being there in the first place, because, you know, no one cares, and now you'd have to check both cases: that's infinitely more QA time.

    Seriously though, you'd be going from:

    - check if the menu item is there

    - check if it's actually working

    to:

    - check if the checkbox is there and correctly setting the corresponding preference

    - check if the option is there, or not, depending on the preference

    - check if the menu item is working

    That's approximately double the QA time, arguably more, allotted for that menu item.

    Don't you do such stuff with automated tests anyway? Or is there some trained monkey forced to click each and every option to see whether it works or not?



  • The fun part about OS X. Everything has a "Share" option. You selected text in your IDE and right clicked? Oh, you can share that by Email or Twitter! You opened a picture of your dick and want to send it to your GF/Wife. Right click -> Share -> BOOM!


    It's like magic. Kinda dumbfounding too that Windows doesn't have useful shit like that.



  • @Rhywden said:

    Don't you do such stuff with automated tests anyway?
    I've seen both the trained monkey and the automated approaches. And of course, the automated approach still needs a trained monkey to write the tests in the first place (probably the developer who wrote the darn thing in the first place, but not necessarily so).

    Anyway, I'm not defending that giving users choices is a bad thing. Just trying to justify the whole "one more checkbox doubles QA time" technicality*, even if it's ridiculous in such a simple example.

    I am very in favour** of giving users as much choice as possible, within reasonable expectations of what they may be able to customize.

    Good software should be flexible. But good software is hard to write.

    *Not that blakeyrat asked for help

    **Is this broken English? 



  • @gu3st said:

    It's like magic. Kinda dumbfounding too that Windows doesn't have useful shit like that.

    Because Windows developers are assholes, and there'd be like 57 of them all reading things like, "Share this item using Adobe AwesomeProduct 3.5 Email", "Share this item using Adobe AwesomeProduct 3.5 IM", etc.

    For some reason, probably because there's less money in the whole ecosystem, Apple developers aren't assholes.



  • @Rhywden said:

    Don't you do such stuff with automated tests anyway?

    a) Automated tests don't fucking work for a shitload of QA tasks, so no you usually can't save much time using automated tests. (They're pretty good for finding regressions though.) I dare you to write an automated test that can detect "Spanish localization uses the word 'drive' as in operate a car, not as in computer disk drive."

    b) The saying isn't LITERALLY true; the point is it is to remind you that adding an option to a product isn't "free". There's time and testing involved in it, no matter how trivial it looks. A lot of developers are like, "just put an option for it!" which is a sure recipe for shitty unusable software.


  • Considered Harmful

    Also, Adobe AwesomeProduct has nothing to do with email or IM but it has them as bundled bloat which it cannot be installed without. Also, they display a splash screen and take 20+ seconds to load, and nag you to update them every day or so even if you're not using them.



  • @blakeyrat said:

    @gu3st said:
    It's like magic. Kinda dumbfounding too that Windows doesn't have useful shit like that.

    Because Windows developers are assholes, and there'd be like 57 of them all reading things like, "Share this item using Adobe AwesomeProduct 3.5 Email", "Share this item using Adobe AwesomeProduct 3.5 IM", etc.

    For some reason, probably because there's less money in the whole ecosystem, Apple developers aren't assholes.

    This can always be dumbed down. Do you not set default applications for a reason? It's not like every link from mIRC asks "WHAT BROWSER YOU WANT FOO!". It opens in your default. IM, Twitter, Facebook and stuff would be the things that lack, as windows has no built in faculties for any of that... but a built in "Share with Mail" menu on selections would be really easy, and really useful.

    Is it too much to think that Microsoft might add useful features to an OS. Metro certainly isn't this (even though it does start some of these sharing functions.. but you haffto use Metro. It's like shooting your leg off to save your hand.)



  • @gu3st said:

    The fun part about OS X. Everything has a "Share" option. You selected text in your IDE and right clicked? Oh, you can share that by Email or Twitter! You opened a picture of your dick and want to send it to your GF/Wife. Right click -> Share -> BOOM!

    It's like magic. Kinda dumbfounding too that Windows doesn't have useful shit like that.
    It does, via the "Send to" menu. Which is even better, because AFAIK third-party programs aren't allowed to put their shit in it.



  • @Zylon said:

    @gu3st said:

    The fun part about OS X. Everything has a "Share" option. You selected text in your IDE and right clicked? Oh, you can share that by Email or Twitter! You opened a picture of your dick and want to send it to your GF/Wife. Right click -> Share -> BOOM!


    It's like magic. Kinda dumbfounding too that Windows doesn't have useful shit like that.
    It does, via the "Send to" menu. Which is even better, because AFAIK third-party programs aren't allowed to put their shit in it.

    That works on files.. I can't select a block of text here in this compose area and say "Send to Email" can I?



    Not in Windows.



  • @gu3st said:

    That works on files.. I can't select a block of text here in this compose area and say "Send to Email" can I?
    That sounds hideously invasive on the OS's part. Is copy/paste really such a cognitive burden now?

     



  • @Mcoder said:

     I'm now tempted to troll the Firefox bugzilla, reporting something like:

    Remove unnecessary DHCP requests

    importance: enhancement

    description:

    The requests the browser makes every time the user visits a page.

    There is no loss of functionality, since resolving the page URL can be done by the following alternative procedure:

    1 - Open a terminal

    2 - Type the command "host pagename.example.com"

    3 - Select the resulting IP address and copy it with CTRL + SHIFT + C

    4 - Paste the address into the address bar of the broser and press ENTER

    To be fair, if Firefox uses DHCP for domain name resolution, you should definitely report it to Bugzilla.



  • @Zylon said:

    @gu3st said:

    That works on files.. I can't select a block of text here in this compose area and say "Send to Email" can I?
    That sounds hideously invasive on the OS's part. Is copy/paste really such a cognitive burden now?

     

    It's not that it's hard.. It's just obvious. Why make someone copy/paste into twitter/email, when you could open a new email for them, or bring up a tweet sheet everywhere.

    Also, it's not really that invasive, considering OSX has had services for the longest time on blocks of text and files. Way back in 10.4, if one was inclined, could have created a service for Twitter that emulated what 10.8 does.

    Microsoft has no such mechanics at all... which is ridiculous in this day.



  • @The Dark Lord said:

    To be fair, if Firefox uses DHCP for domain name resolution, you should definitely report it to Bugzilla.
     

    DHCP! And I didn't notice it when writting. At least I only made the mistake once, I'm quite able to repeat this kind of mistake all over a document and not notice any ocurrence.

    But ok, how do I report a bug in my bug report?



  • @blakeyrat said:

    @Rhywden said:
    Don't you do such stuff with automated tests anyway?

    a) Automated tests don't fucking work for a shitload of QA tasks, so no you usually can't save much time using automated tests. (They're pretty good for finding regressions though.) I dare you to write an automated test that can detect "Spanish localization uses the word 'drive' as in operate a car, not as in computer disk drive."

    I'm reasonably sure that an automated test consisting of "Context menu on hyperlink has item 'xyz' when invoked over a hyperlink when Option 'MNO' is set to 'True'" and "Clicking on item 'xyz' when invoked over a hyperlink yields result 'abc'" can easily be written.

    I mean, I'm using a test suite which is able to cope with AJAX request results changing the DOM.
    This, in comparison, should be child's play.



  • @Rhywden said:

    I'm reasonably sure that an automated test consisting of "Context menu on hyperlink has item 'xyz' when invoked over a hyperlink when Option 'MNO' is set to 'True'" and "Clicking on item 'xyz' when invoked over a hyperlink yields result 'abc'" can easily be written.

    Why don't you read the second half of my post.



  • @blakeyrat said:

    @Rhywden said:
    I'm reasonably sure that an automated test consisting of "Context menu on hyperlink has item 'xyz' when invoked over a hyperlink when Option 'MNO' is set to 'True'" and "Clicking on item 'xyz' when invoked over a hyperlink yields result 'abc'" can easily be written.

    Why don't you read the second half of my post.


    If that one thing you forgot to test for causes your UI to break, running the program and looking at it is generally the better of the QA methods.



  • @blakeyrat said:

    @Rhywden said:
    I'm reasonably sure that an automated test consisting of "Context menu on hyperlink has item 'xyz' when invoked over a hyperlink when Option 'MNO' is set to 'True'" and "Clicking on item 'xyz' when invoked over a hyperlink yields result 'abc'" can easily be written.

    Why don't you read the second half of my post.

    Why don't you reply with an example which fits? I mean, I'm asking why you can't test for, say, the existance of something and you go off on a completely irrelevant tangent where you prove that the banana prices in Tobago are ridiculously high. Seriously, reread your example and tell me why that should make any sense. I know that automated tests don't work for everything - but yours has to be the stupidest answer I've ever heard.

    Plus, with that kind of reasoning, why would be ever put any additional feature into a software tool? Yes, maybe only 0.5% percent of people use that feature. With a user base in the millions, that's still a lot of people!

    @Ben L. said:
    @blakeyrat said:
    @Rhywden said:
    I'm reasonably sure that an automated test consisting of "Context menu on hyperlink has item 'xyz' when invoked over a hyperlink when Option 'MNO' is set to 'True'" and "Clicking on item 'xyz' when invoked over a hyperlink yields result 'abc'" can easily be written.

    Why don't you read the second half of my post.


    If that one thing you forgot to test for causes your UI to break, running the program and looking at it is generally the better of the QA methods.

    What is there to test? "If option set to 'False' menu item should not exist" - "If option set to 'True' menu item should exist".



  • @Rhywden said:

    @Ben L. said:
    @blakeyrat said:
    @Rhywden said:
    I'm reasonably sure that an automated test consisting of "Context menu on hyperlink has item 'xyz' when invoked over a hyperlink when Option 'MNO' is set to 'True'" and "Clicking on item 'xyz' when invoked over a hyperlink yields result 'abc'" can easily be written.

    Why don't you read the second half of my post.


    If that one thing you forgot to test for causes your UI to break, running the program and looking at it is generally the better of the QA methods.

    What is there to test? "If option set to 'False' menu item should not exist" - "If option set to 'True' menu item should exist".

    How would you write a test for a bug you haven't encountered yet?

    Edit: I think the problem here is that you're talking about this specific bug and I'm talking about QA in general.



  • @Rhywden said:

    Why don't you reply with an example which fits?

    Example of what? What are you talking about?

    @Rhywden said:

    I mean, I'm asking why you can't test for, say, the existance of something and you go off on a completely irrelevant tangent where you prove that the banana prices in Tobago are ridiculously high.

    The part of my post you didn't (and still haven't) read states that the "doubles the testing time" statement isn't literally true.

    @Rhywden said:

    Seriously, reread your example and tell me why that should make any sense.

    What example? Example of what?

    @Rhywden said:

    I know that automated tests don't work for everything - but yours has to be the stupidest answer I've ever heard.

    Are you talking about the example with the wrong localization string? Ok; you're the expert: how would you write an automated test to check that?

    @Rhywden said:

    Plus, with that kind of reasoning, why would be ever put any additional feature into a software tool? Yes, maybe only 0.5% percent of people use that feature. With a user base in the millions, that's still a lot of people!

    The point is you have to trade-off the additional work-effort and potential product bugginess of the feature which the usefulness of the feature. If the feature is useful to enough people, then of course the trade-off is worth it-- but before you write the code you need to be aware of the total cost of the feature. Sure you wrote it in 5 lines, but then you had to have it localized, you had to build the UI, you had to QA it, it turns out your code had 2 bugs in it you had to fix and didn't notice, it turns out your code interacts negatively with someone else's feature they added three years ago you had to fix, etc.

    Writing the code is the LEAST part of software development. The sooner developers get that lesson in their heads, the sooner we'll have high quality usable software.



  • I'm talking about this specific feature. One which you compared to "I dare you to write an automated test that can detect "Spanish localization uses the word 'drive' as in operate a car, not as in computer disk drive." which would be a completely different and non-deterministic task.

    Where in the seven hells do you get this nightmare of complexity from a simple on/off-toggle? By your line of reasoning, even a simple "hello world" is certain to be a bugfest which costs years of QA manpower.

    Learn to read, you fool.



  • @Rhywden said:

    I'm talking about this specific feature.

    Ok; that's not what I was talking about.

    @Rhywden said:

    Where in the seven hells do you get this nightmare of complexity from a simple on/off-toggle? By your line of reasoning, even a simple "hello world" is certain to be a bugfest which costs years of QA manpower.

    The point I'm making is that writing the code is the LEAST important part of developing software.



  • Yes, I got that ages ago. Your problem is going off irrelevant tangents which have nothing to do with what I wrote or asked and then accusing me of not reading stuff. Maybe you should take a step back and take your own advice for once, hmmh?



  •  @Zylon said:

    @gu3st said:

    The fun part about OS X. Everything has a "Share" option. You selected text in your IDE and right clicked? Oh, you can share that by Email or Twitter! You opened a picture of your dick and want to send it to your GF/Wife. Right click -> Share -> BOOM!

    It's like magic. Kinda dumbfounding too that Windows doesn't have useful shit like that.
    It does, via the "Send to" menu. Which is even better, because AFAIK third-party programs aren't allowed to put their shit in it.

    I think they can via going to the path 'shell:sento' and creating a shortcut there. I assume.

     


Log in to reply