http://qs321.pair.com?node_id=492955


in reply to Re^4: Deprecate target attribute in <a> tag
in thread Deprecate target attribute in <a> tag

I didn't say "broken"; I said "b0rked".

Actually, you said, "b0rken." Twice. And I said, "b0rked." Not that this is particularly relevant...

It is unfortunate that browser-level control of this behavior is not widespread or standard.

I wholeheartedly agree.

The choice of a browser involves tradeoffs. You cannot infer acceptance of the consequences of abusive web coding from the choice of browsers.

It isn't "abusive web coding." There arguably is no such thing so long as the code you write adheres to the standard. The onus is on the browser to provide acceptable behavior. In the case of target="_blank" the HTML 4.01 spec states only that the browser "should load the designated document in a new, unnamed window."¹ It is not a requirement that the browser do so to maintain compliance. So, yes, when you can choose between more than one compliant browsers, some which behave as you wish and some which don't, then your is choice is at fault.

But, all of that said, I'm not inferring your acceptance at all. I'm only suggesting that you are barking up the wrong tree. In other words, go complain to your friendly neighborhood browser developers. Or write a browser yourself that behaves the way you want. Or join the W3C and work to get the standard changed. Complaining because PM is allowing perfectly acceptable HTML is just misplaced irritation.

I really dislike web sites that think they should control my browser by opening windows when I click links.

Well, you can dislike a site for whatever reasons you choose. I dislike animated gifs. If I were to say that your way, I'd complain, "I really dislike web sites that think they should control my browser by displaying crappy animations when I open a page." But, the point continues to be that, if the website is controlling our browsers, it's because our browsers have let themselves be controlled. In reality, websites can't really do more than provide hints.

Fortunately for me, extensions for FireFox, my favored browser, allow me to turn off image animations as well as handle target="_blank" the way I want to handle it. So, I've got no complaints. Well... that's not true. I have other complaints now. For instance, it drives me batty that FireFox makes a separate request when I want to view source. That's a real pain when debugging issues on pages generated from POST requests.

1. The use of the word "should" here is as defined in RFC2119 Key words for use in RFCs to Indicate Requirement Levels.

3. SHOULD This word, or the adjective "RECOMMENDED", mean that there may exist valid reasons in particular circumstances to ignore a particular item, but the full implications must be understood and carefully weighed before choosing a different course.

-sauoq
"My two cents aren't worth a dime.";

Replies are listed 'Best First'.
Re^6: Deprecate target attribute in <a> tag (Firefox cache)
by kelan (Deacon) on Sep 20, 2005 at 17:40 UTC

    Off-topic FYI:

    For instance, it drives me batty that FireFox makes a separate request when I want to view source.

    I've read that the next version of Firefox will not make a separate request for viewing source, as it tries to rely on the cache much more than it does now (also producing a speed-up in using the Back and Forward buttons).

      Thanks for the info.

      I really believe FireFox's current behavior (current as of 1.0.6 anyway) should be considered broken. If I want to view the source, then it would seem obvious I'm interested in the source of the page I'm looking at rather than the source of a page I haven't requested yet!

      -sauoq
      "My two cents aren't worth a dime.";
      
Re^6: Deprecate target attribute in <a> tag
by herveus (Prior) on Sep 18, 2005 at 16:29 UTC
    Howdy!

    *sigh*

    My attempts at subtle nuances of connotation were clearly too subtle. No, this is not meant to be taken as "questioning your ability to read for comprehension" or any of those wonderful (?) phrases that serve to keep quality flame wars warm and toasty.

    "broken" makes a clear statment that something is, well, broken, and (implicitly) should be fixed. "b0rken" was meant to convey the sense that, while I think there is something very wrong with the matter, I recognize that my view is (almost certainly) not universally held, and may, in fact, be a distinctly minority view.

    On "abusive web coding"...

    Just because something is legal by the relevant standards has no necessary bearing on whether or not it is abusive. I suspect that the hell of popup windows makes an instructive example. In the instant matter, I have yet to see, here or elsewhere, reasoned and convincing argument for *why* setting the target attribute to _blank is valuable. I have seen explanation of how it is a useful attribute in a framed setting.

    When a practice has the recommended behavior of a construct is to cause the appearance of a new window, that has consequences on my computer that I wish to be able to control. To compare it to animated GIFs is fatuous. As a reducto ad absurdum argument, it doesn't wash. That I have been able to use feedback from this discussion to adjust the configuration of my browser to mitigate this behavior does not make the attempt to exercise it less egregious.

    From the early days of the WWW, there has always been tension between people who felt it necessary to show that they knew how to work at the edges by putting up web pages and sites that presented truly obnoxious visions and viewers who rightly felt that they were the ones who should have considerable control over just what appeared on their screens. Some of us try to choose our browsers to maximize our control over the things web so-called designers try to inflict on us.

    You cast the choice of browser as between picking one that behaves as I wish versus one that doesn't. I wish it were that simple. It isn't, and to then assert that my choice (involving tradeoffs) makes me at fault for the effects of the shortcomings of the choice I did make is so much baloney.

    I have no expectation at this point that the target attribute will be removed from the list of acceptable attributes for the a tag. So be it. I floated the question, and it did not garner the sort of support I'd expect to see to cause it to be considered for implementation.

    yours,
    Michael

      I'm responding to portions of this out-of-order. Hope you don't mind...

      "b0rken" was meant to convey the sense that, while I think there is something very wrong with the matter, I recognize that my view is (almost certainly) not universally held, and may, in fact, be a distinctly minority view.

      Yes, well, as far as I am concerned, you needn't try to convey that. I give everyone the benefit of the doubt and assume that they know their views aren't universally held right up until they show they are so naive as to think otherwise. You don't strike me as naive, Michael.

      You do strike me as sort of a crotchety luddite. That's a compliment. I strike myself the same way.

      From the early days of the WWW, there has always been tension between people who felt it necessary to show that they knew how to work at the edges by putting up web pages and sites that presented truly obnoxious visions and viewers who rightly felt that they were the ones who should have considerable control over just what appeared on their screens.

      I'm very aware of the tension that built up as a result of people wishing to control presentation instead of relying on browsers to do it acceptably for them. (And the people on the other side of the fence that wished their browser would give them a way to turn off blink tags.) I started my career using Mosaic (v1.0 on SunOS) right around the time when NCSA's HTTPd v1.0 (with CGI!) came out. Fortunately, these days, that tension is mostly gone thanks to the slow but steady removal of presentation-only elements from HTML and the adoption of CSS. (The one big hump that has yet to be gotten over is table-based layouts; everything else is relative cake.)

      In the instant matter, I have yet to see, here or elsewhere, reasoned and convincing argument for *why* setting the target attribute to _blank is valuable.

      Here's one: I like it. I wish all external links on perlmonks had target="_blank". That would save me from needing to control click on external links. That said, I don't really like that the only convenient way to provide that functionality now is to co-op the HTML spec's target attribute. I'd rather that there were a convenient way to indicate that a link is "external" to the current site. As a "site" might span domains, an explicit hint is really needed for this. I also think a link's "externality" is an inherent property of the link which should be able to affect more than presentation, so homebrewed CSS solutions don't seem to be the right approach. An external attribute which could assume values yes or no would probably be sufficient. Unfortunately, the spec doesn't provide that so we've got to make do with what we have.

      When a practice has the recommended behavior of a construct is to cause the appearance of a new window, that has consequences on my computer that I wish to be able to control. To compare it to animated GIFs is fatuous. As a reducto ad absurdum argument, it doesn't wash.

      The only reason it "doesn't wash" in your opinion is because you don't feel the same way about animated GIFs as you do about new windows. I, on the other hand, find them severely irritating. If you felt the way I do about animated GIFs, you might decide that comparing them to new windows was fatuous for the same reasons you find the converse fatuous now. I'm pointing out here that it's all a matter of opinion. I have yet to see you even try to provide an argument that would convince me that new windows are somehow inherently worse than animated GIFs. Why should I accept that a priori?

      It isn't, and to then assert that my choice (involving tradeoffs) makes me at fault for the effects of the shortcomings of the choice I did make is so much baloney.

      It isn't baloney. It's just the way it is. What you fail to appreciate about my stance on this is that it is purely pragmatic. Petitioning one website to change isn't a scalable solution. Choosing a browser that provides the behavior you want is. Requesting your chosen browser's developers to add that feature is. Getting the spec changed is. When there are tons of people that feel the same way you do, and they all start boycotting sites because of this issue, well, then socially driven change might happen. But, I don't think you'll ever see that kind of across-the-board support on this issue. Like I said, I like target="_blank" and there are others like me. This isn't the blink tag we are talking about. Almost everyone hated that. ;-)

      -sauoq
      "My two cents aren't worth a dime.";
      
        I wish all external links on perlmonks had target="_blank". That would save me from needing to control click on external links.

        I'm glad they don't. I haven't made the switch from Mozilla to FireFox because in Mozilla (and Netscape), the difference between opening a link in the same window or in a new window is a slight adjustment of my index finger: Button-1 vs Button-3 (left button vs. middle button). In FireFox, Button-3 opens a new tab. I prefer windows over tabs, as I run a well tuned window manager, and not a tab manager.