Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
×
The Internet Programming Software Upgrades Technology

No More Version Numbers For HTML 336

An anonymous reader writes "HTML5 will be the last version of HTML that carries a version number. Ian Hickson, a Google engineer and editor of the HTML5 standard, announced that the language will be transitioned to a 'living standard' without version numbers. A bit like Chrome, if you will."
This discussion has been archived. No new comments can be posted.

No More Version Numbers For HTML

Comments Filter:
  • Not a Standard. (Score:5, Insightful)

    by Anonymous Coward on Thursday January 20, 2011 @03:13PM (#34943742)

    If you never finalize it's not a standard. This sounds like a Microsoft move to me.

    • Re:Not a Standard. (Score:4, Informative)

      by Anonymous Coward on Thursday January 20, 2011 @03:53PM (#34944262)

      The real reason is because the committee which desides whats in the standard couldn't get a consensus, so HTML 5 took forever and its still not a W3C recormendation.

      Look at http://www.w3schools.com/w3c/w3c_html.asp

      HTML 4.01 became a W3C Recommendation 24. December 1999
      XHTML 1.0 became a W3C Recommendation 20. January 2000
      On January 22nd, 2008, W3C published a working draft for HTML 5.

      So 10 years after xhtml 1.0 we still dont have a W3C recormendation for HTML 5, if HTML 6 carried on business as usual then we would probably be looking at 2025 for it...

      • Re: (Score:3, Interesting)

        by sdiz ( 224607 )

        You missed all those XHTML 1.1 Modules. That's how W3C wasted all its time.

        Original HTML5 draft came from WHATWG, not W3C.

      • by bradley13 ( 1118935 ) on Friday January 21, 2011 @05:57AM (#34950380) Homepage

        HTML 2/3/4, XHTML/1, and CSS/1 were all small, simple, understandable standards. Then the web got popular - in part because web standards and technology were so simple. Once the web had exploded, every damn company wanted to stick its oar in. CSS 2 took years, is overly complicated, but still just barely manageable. Look at CSS 3 - everybody's special wishes are in there - the thing is immensely complex and as a standard, frankly, it is therefore nearly useless. HTML 5 is much the same - too many special wishes and fancy features. One needs to take a weed-whacker to it and to CSS, to restore some degree of simplicity.

        Think of it this way: why is there a competition to see how well browsers score on the ACID tests? The standards ought to be simple enough that any decent browser scores 100%. The fact that this is not the case is proof that the standards are far, far too complex.

    • Sure it is. The standard is what is currently on their page. If you are using deprecated methods, you are non-standard.

      You could be standard compliant one day and be out of compliance the next.

      Maybe it's better if I put it this way. There's a new standard every day. It's just the nobody puts a version on it.

      • That will truly help the industry, when contracts calling for levels of compliance become impossible and designers can never get paid because their work is never compliant.

    • Now browsers only have to support moving targets.
  • by shadowrat ( 1069614 ) on Thursday January 20, 2011 @03:13PM (#34943744)
    POST!
  • terrible idea (Score:5, Insightful)

    by godrik ( 1287354 ) on Thursday January 20, 2011 @03:15PM (#34943760)

    You'll get pages that becomes invalid with time despite they were valid before. That sounds like a very stupid idea.

    Until you name the revision by dates, which is basically the same thing as giving version numbers...

    • Re:terrible idea (Score:5, Interesting)

      by hedwards ( 940851 ) on Thursday January 20, 2011 @03:33PM (#34944010)
      That was my thought, it's tough enough to get browsers in compliance with a specific revision of HTML, now they're wanting to do away with numbering them?

      I have to assume that this is an early April Fool's joke or the person suggesting it is full of it. But then again he works for Google and is probably just the sort of arrogant git that doesn't understand the implications of it for people that aren't constantly upgrading their browsers.
    • Re:terrible idea (Score:4, Interesting)

      by Anonymous Coward on Thursday January 20, 2011 @03:47PM (#34944196)

      Until you name the revision by dates, which is basically the same thing as giving version numbers...

      Or names! Firefox supports HTML Insomnia. IE is now on HTML Narcoleptic. Chrome upgrades to HTML Sonambulia. Opera is still on HTML Purgatory, but they will go to HTML Heaven next month. Oh, the possibilities!

    • by DragonWriter ( 970822 ) on Thursday January 20, 2011 @03:55PM (#34944288)

      You'll get pages that becomes invalid with time despite they were valid before.

      That is a result of backward-incompatible changes, not the absence of version numbers.

      • by EdIII ( 1114411 ) on Thursday January 20, 2011 @05:08PM (#34945404)

        You'll get pages that becomes invalid with time despite they were valid before.

        That is a result of backward-incompatible changes, not the absence of version numbers.

        Quite true, but what I think the poster was saying is that without version numbers it would be impossible to claim they were "standards" compliant at any one time. So even if you wrote very good code that was compatible across 99% of all browsers out there, a few years go by and you look like lazy morons that just don't care.

        As for the backwards-incompatible changes, without version numbers you would really have no way to tell what you are doing anyways. Since you can't reference it by version number you would be forced to reference by a specific instance of a problem. The newest Firefox blah blah blah tends to have a problem with this, this, and this, and Opera v.x tends to have a problem with that, that, and that.

        Next thing you know the browsers will go versionless too and then at that point all you can do is drink heavily.

        • So even if you wrote very good code that was compatible across 99% of all browsers out there, a few years go by and you look like lazy morons that just don't care.

          That doesn't happen unless the standard is accepting backwards-incompatible changes to widely-established features, which they've committed not to.

          As for the backwards-incompatible changes, without version numbers you would really have no way to tell what you are doing anyways. Since you can't reference it by version number you would be forced to reference by a specific instance of a problem. The newest Firefox blah blah blah tends to have a problem with this, this, and this, and Opera v.x tends to have a problem with that, that, and that.

          Which is what you have to do with features in the real world anyway.

          • by Haeleth ( 414428 )

            That doesn't happen unless the standard is accepting backwards-incompatible changes to widely-established features, which they've committed not to.

            Provided you only use "widely-established" features. Which ones are those, specifically? Because they certainly have not made any commitment to reject backwards-incompatible features in general. Quite the opposite: they make it very clear that if they decide something is "broken", they will change it without warning. Hope you weren't relying on that "broken"

    • Numbers are so square. Introducing HTML version "Fred"!

    • This is the same thing that happens with Operating Systems' need for versioning. What does the board plan, "reflexion" API's where each dev must ask the browser what it can do and assume the missing features in the ethereal standard are enough for the page to render 3 years from now? 10 years from now?

      Just like an OS, the standard can drop features at any time; the point of numbers is to tell the dev from an easy test what stylesheets to junk and what error messages to give.

    • by Dracos ( 107777 )

      Validity rot isn't the issue here, it's that compliance is now a constantly moving target.

      Browser vendors (some more than others) have always had a tough enough time getting their products to comply when the spec was finished and became static. With a dynamic spec chaos will surely ensue and developers will pay the price, because the lowest common denominator won't be as much of a known value as it is now.

      Also, didn't Daniel Glazman lament a few years ago about how one of CSS's biggest flaws was the lack o

  • by tokul ( 682258 ) on Thursday January 20, 2011 @03:15PM (#34943762)
    Now we'll have beta quality software and beta quality standards. Another engineer brainwashed.
    • by Anonymous Coward on Thursday January 20, 2011 @03:17PM (#34943790)

      That's The Google Way, eternal beta.

    • by Nadaka ( 224565 )

      I know. At this point I would seriously consider junking the whole steaming pile and writing my own. Not that it would ever be used by anyone. The steaming pile has already reached critical mass.

  • by MrEricSir ( 398214 ) on Thursday January 20, 2011 @03:15PM (#34943768) Homepage

    ...I can always render the latest HTML in Netscape Navigator. Right?

  • Um... (Score:5, Insightful)

    by FatSean ( 18753 ) on Thursday January 20, 2011 @03:15PM (#34943770) Homepage Journal

    People will still need to differentiate between implementations of HTML that have different features...do they expect us all to just use the latest and hope nothing breaks?!

    • Re:Um... (Score:5, Insightful)

      by I8TheWorm ( 645702 ) * on Thursday January 20, 2011 @03:19PM (#34943834) Journal

      Yeah, I look forward to the "this site is compliant with some of HTML standards and not others because they're too new. We can't really define that for you because there is no version, so best of luck to you" badges.

    • It is possible that individual features of HTML will have versions instead of the entire standard. Maybe each tag will have a version? :P

      • by hitmark ( 640295 )

        Or as time moves on, various tags, or sub-features of a tag, gets defined as deprecated.

        As long as the basic definition of a tag, and its layout have some kind of default (hello, xml) then this can be done.

    • People will still need to differentiate between implementations of HTML that have different features...

      This presumes that future versions of HTML from the point of adoption forward have backward-incompatible changes. The solution to that is to minimize backward-incompatible changes.

  • Slow Browsers (Score:4, Insightful)

    by Anonymous Coward on Thursday January 20, 2011 @03:15PM (#34943772)

    Wow, so now my browser has to interrogate every single element on a page to determine what's supported BEFORE going to plugins etc.

    Yikes...

    • As a practical matter, that's pretty much what you have to do now. Modernizr and the like make it easier for you, but ultimately that's what they're doing. You can't really trust what each browser claims to support anyway.

    • Just have the browser "not render" the element that's not supported anymore. That's pretty much what they do now.

  • Translation (Score:5, Insightful)

    by dgatwood ( 11270 ) on Thursday January 20, 2011 @03:18PM (#34943810) Homepage Journal

    Microsoft got tired of people asking when they were going to fully support HTML 4....

    Now everyone will be able to say "We support HTML" even though nobody fully supports all aspects of the spec. Just like today, only nobody will be able to point their finger at any sort of milestone that they missed, so companies that drag their heels in standards compliance end up looking better.

    How is this a benefit again? It seems to me that we need smaller, more frequent milestones, not elimination of those milestones.

    • by 0racle ( 667029 )
      This sounds more like a standards body getting sick of people and developers asking when HTML5 standardization would finally be finished.
    • by jc42 ( 318812 )

      Now everyone will be able to say "We support HTML" even though nobody fully supports all aspects of the spec. Just like today, ...

      Yeah, and the major benefit will be that developers of web sites will no longer waste their time trying to figure out what numbered version to declare in their DOCTYPE line, and pointing their fingers accusingly at browsers that don't support exactly that standard. They'll go right to testing against a flock of more-or-less current browsers (plus IE6 ;-), and making sure their HTML works somewhat sensibly in all of them.

      Fact is, it has never worked very well to study any particular HTML standard and code s

  • Living Standard? (Score:5, Insightful)

    by ultranova ( 717540 ) on Thursday January 20, 2011 @03:19PM (#34943824)

    So, in the future it's impossible to figure out what browser supports what? Because, after all, browser support is dragging behind years even now. Or is that the very goal of Google? Make Chrome the de facto standard, and force everyone else to play the catch-up game?

    Seriously, don't do this "living standard" crap. At the very least use minor version numbers to identify a given set of standards. Don't force me to guestimate how a web page I write today is going to behave in browsers 5 years from now; let me specify what behaviour I want.

    • by jc42 ( 318812 )

      So, in the future it's impossible to figure out what browser supports what?

      Why would you expect the future to be different from the past?

  • Problem (Score:5, Insightful)

    by Improv ( 2467 ) <pgunn01@gmail.com> on Thursday January 20, 2011 @03:19PM (#34943832) Homepage Journal

    There will be no way to pressure browser developers to be compliant with "NGHTML 4.7" if we can't even talk about it because it lacks a name. It'll also be hard to enumerate features of releases, to decide what version of the standard we're talking about and have programmatic support for that, etc.

    This eliminates most of the benefits of having standards to begin with.

    • "Don't you see that the whole aim of Newspeak is to narrow the range of thought? In the end we shall make thoughtcrime literally impossible, because there will be no words in which to express it."

      http://www.george-orwell.org/1984/4.html [george-orwell.org]

    • if we can't even talk about it because it lacks a name.

      Hey, we've seen this before - no numbers, but we can have HTML Pro, HTML Extreme, HTML on Acid, HTML JC, etc.

      Seriously though, if there is a written standard, no matter what they don't call it, people will label it. HTML 2012, or whatever, will be what was in effect as of January 1st 2012.

      Maybe what they're trying to do here isn't to keep browser writers from having excuses for not keeping up, but for keeping the standards body from feeling like if th

  • So instead... (Score:5, Interesting)

    by DoofusOfDeath ( 636671 ) on Thursday January 20, 2011 @03:20PM (#34943838)

    So instead of versions, we'll have a big vector of flags, where each flag indicates whether or not a particular HTML feature is required, supported, etc.? And a given web page will work with a given browser only if their two flag vectors are compatible?

    This is stupid. Standards exist for a reason.

    • by mini me ( 132455 )

      We'll have exactly what we have now. Browser vendors were already adding draft features into their product before the specification was finalized. Just look at how many browsers support HTML5 features, even though HTML5 does not exist as a standard yet.

      • I was going to mod, but I cannot let this pass. Browsers have always added features that's true. And in some ways it is required, you cannot make progress by mindlessly adding to the standard, you will have to try out new features first.

        But any *mainstream* site should be able to choose a HTML version to support, maybe taking into account some features that are badly supported by the browsers. This way you can be reasonably sure that most browsers (and with the internet appliances market soaring, there are

    • we'll have a big vector of flags

      No. There will be a /htmlcaps.xml in the root of every website, enumerating the features necessary to render it.

    • I think the "vector of flags" idea has merit, but it introduces worse issues than those it solves. Consider privacy and user-tracking issues; this vector would make it trivial to uniquely identify users because it contains that much more information (see also the EFF's Panopticlick [eff.org]).

      We still need "milestones" which can be marked, even if they are years, quarters, or months instead of versions. In this manner, we can still determine compatibility without introducing millions of different combinations of

  • by Anonymous Coward on Thursday January 20, 2011 @03:21PM (#34943858)
    Go straight to the source [whatwg.org] instead.
    • The subheading on that link seems particularly appropriate:

      > Please leave your sense of logic at the door, thanks!

      Sigh.

    • by Anonymous Coward on Thursday January 20, 2011 @05:06PM (#34945378)

      Indeed, and as that article points out, this change in naming applies ONLY to what the WhatWG was calling "HTML5", not to be confused with what W3C calls "HTML 5." For anyone that's been following this, or has read Zeldman's HTML5 book, knows, "HTML5" and "HTML 5" can refer to entirely different sets of standards.

      The W3C, as far as I can tell, is still taking "snapshots" of WhatWG's "HTML" spec and numbering them, and the W3C is still the primary authority when it comes to official web specifications.

      This change really isn't as big of a deal as people here seem to think, and the original article does confuse the issue.

  • Just like Chrome? (Score:4, Insightful)

    by Anonymous Coward on Thursday January 20, 2011 @03:21PM (#34943862)

    Do they mean the browser Chrome? As in Google Chrome 8.0.552.237?
    Is 8.0.552.237 not the version?

  • HTML5 is lacking in adoption, incompatible, slow etc.

    And there is no need to further develop a bi-plane when you fly a jet already.
    • Really? That's weird because I thought browsers were waring over who could get HTML 5 features out first, who had the most, and showing them off.

      But I might be totally crazy.

      • ...The various Javascript APIs, web sockets, IndexedDB, SVG, etc, are not HTML 5.

      • It's a typical problem, the browsers battle it out first, then applications and sites tend to pop up. Which makes it a bit awkward at times.

        From what I can tell, they're opting to keep the aspect of HTML which is more or less the most broken under the justification that they've always done it that way, regardless of the fact that HTML5 was mostly supposed to be changing that. A new set of standards that both modernized and theoretically was implemented by all the browsers.

        It's more or less inevitable
    • by MightyMartian ( 840721 ) on Thursday January 20, 2011 @03:25PM (#34943898) Journal

      And what is this jet you speak of? Javascript, CSS, DOM? If these are jets, then someone put the turbines in backwards, pasted the wings on with glue sticks, and is using banana peels as fuel.

      • Well, yeah. Those features came first, then standards came behind to try to reduce the amount of extra work necessary to work around browswer weirdness.
  • Huh? (Score:4, Interesting)

    by gstoddart ( 321705 ) on Thursday January 20, 2011 @03:28PM (#34943940) Homepage

    Since when did Google become the keepers of the HTML spec?

    I think a randomly changing feature-set sounds like a bad idea. HTML is supposed to be a standard, not something which just changes without any real control behind that.

    This is like agile programming run amok -- let's expect the customer to have to upgrade to the latest nightly build. That'll work!

    • Since when did Google become the keepers of the HTML spec?

      Since Ian Hickson moved to Google?

    • It's Stockholm syndrome. They're so used to the unreliable renderings that they're needing to create something which keeps that aspect of cruising the web.
    • by mini me ( 132455 )

      Browsers were already adding support for draft features as they happened. I believe his point is that there is no use in waiting until the spec is finalized; add the features they become available and let people start using them now.

    • Re:Huh? (Score:4, Informative)

      by DragonWriter ( 970822 ) on Thursday January 20, 2011 @04:27PM (#34944776)

      Since when did Google become the keepers of the HTML spec?

      Google is not "the keepers of the HTML spec". Ian Hickson, who happens to work for Google, is the editor of the HTML5 spec. Usually, spec maintainers work for a firm involved in the area the spec addresses.

      I think a randomly changing feature-set sounds like a bad idea.

      In none of the discussion of this change has there been any indication that the WHATWG process for HTML will involve random changes.

      HTML is supposed to be a standard, not something which just changes without any real control behind that.

      There is a process, which is discussed in the WHATWG FAQ [whatwg.org]. The process just doesn't involve version numbers anymore.

  • Comment removed based on user account deletion
    • by cjcela ( 1539859 )
      Maybe HTML will die on its own, on the hands of incompetency. Then hopefully a reasonable standard will arise. One that is not designed by committee.
  • by Anonymous Coward on Thursday January 20, 2011 @03:37PM (#34944064)

    Their justifications for the decision are here:

    http://wiki.whatwg.org/wiki/FAQ#What_does_.22Living_Standard.22_mean.3F

  • +1 to everyone who thinks this is stupid. In particular given how significant HTML is to the web-as-we-know-it surely there must've been some consultation before making a call like this? With a cacaphony of "NO" coming through here (and very little, if any, support) one has to wonder....

    • by Dracos ( 107777 )

      If this is Hickson's idea, it certainly tops all the other terrible or poorly reasoned ideas he's come up with.

  • Bad engineering (Score:3, Insightful)

    by cjcela ( 1539859 ) on Thursday January 20, 2011 @03:44PM (#34944162)
    We are in the hands of morons. A constantly evolving standard is bad news for everybody except maybe a few companies that sell HTML development tools. Fast forward 10 years, and we will not be able to read half of the web, and will need 10 different browsers to see our usual choice of sites. There is a reason for versioning. Keeping up with a website will be a pain. I guess I should not complain, many people will have a job thanks to this.
    • A constantly evolving standard is bad news for everybody

      The standard was, in fact, constantly evolving anyway, and all the browser vendors (and other heavily interested parties) were engaged in the process and doing pretty much exactly what they would do under a version-number-less process.

      The only difference is that people that weren't deeply involved were dealing with snapshots that the major players were treating as outdated.

  • by swsuehr ( 612400 ) on Thursday January 20, 2011 @03:50PM (#34944230) Homepage

    I broke the cardinal rule and read TFA. From TFA:

    "Hickson mentions that the group will be dropping the HTML5 name immediately, but it we have not received a confirmation that this will happen over at the W3C as well."

    So WHATWG will no longer be using numbers? WHATWG can call it "Hullapuhjelpus" as far as I'm concerned as long as W3C still continues using version numbers. Version numbers provide excellent reference points to featuresets and are useful to implementers, developers, and end users alike.
    From the WHATWG Blog:

    "However, shortly after that we realised that the demand for new features in HTML remained high, and so we would have to continue maintaining HTML and adding features to it before we could call "HTML5" complete, and as a result we moved to a new development model, where the technology is not versioned and instead we just have a living document that defines the technology as it evolves."

    Because there's demand for new features you no longer want to use a numbering scheme? Many standards are evolving. Why not just increment the minor version when new features are added? HTML version 5.1 added this cool thing, 5.2 this cool thing, etc.

    If we're dumping version numbers then why bother calling it Internet Explorer 6, 7, 8, and 9? Why not just call it "Internet Explorer"? We all know that each of those versions render pages the same, right? Hmm. I just realized that I invoked Internet Explorer in a discussion about standards. Mea Culpa.

    How does removing the version number help the people who need to implement and work with the standard?

    • Say hello to HTML 2011.

    • by Bogtha ( 906264 ) on Thursday January 20, 2011 @04:38PM (#34944936)

      How does removing the version number help the people who need to implement and work with the standard?

      It doesn't, it's a fucking disaster. I'll give a concrete example. I used HTML 5 audio on a site with a Flash fallback for browsers that didn't support it. All is good and well. One day, I start getting complaints that the audio is broken. Turns out that a) the HTML 5 spec had changed and b) Firefox had changed to match in a minor point release. Firefox 3.51 worked, Firefox 3.5.2 didn't, as I recall. The new API was indistinguishable from the old API in as much as all the same objects and functions were there, but a return value had changed. So, even with the best practice method of feature detection, anybody writing to the old API was screwed.

      So I fixed it up by removing the HTML 5 audio and made the decision to wait until HTML 5 was published in its final form. Something that I should have done to begin with really, it's madness to use HTML 5 at the moment as it's just not finished yet. You don't know what is going to change.

      And now they want to do away with a "final" version altogether? Gee thanks, guys! How am I going to be able to trust it to be stable enough to rely on ever again? What's going to stop the same thing from happening over and over again?

  • Funny distribution of approval on the linked blog's comments there (I hope not to have violated the rule of not RTFA I just read the comments there I swear).
    First 10 comments, 2 negative. Last 10 (around n.50 as I write) 9 negative.

  • Color me "not surprise". Engineers, much like artists, have a hard time knowing when something is done and want to "tweak and tweak" everything to death.

    Solution? Rather than *finish* something, just remove the versions! It'll be in development for perpetuity - an engineer's dream come true.

  • HTML ME, closly followed by HTML XP

  • Bad interpretation (Score:5, Informative)

    by robmv ( 855035 ) on Thursday January 20, 2011 @04:09PM (#34944512)

    What was said is that the moving spec in development is now called HTML, when a snapshot is taken it will be called HTML5, next HTMLX.X.X or any other name. The WHATWG spec is not a finalized document, HTML5 will be snapshoted sometime

  • Forever in beta. (Score:5, Interesting)

    by westlake ( 615356 ) on Thursday January 20, 2011 @04:18PM (#34944664)

    Ian Hickson, a Google engineer and editor of the HTML5 standard announced that the language will be transitioned to a 'living standard' without version numbers. A bit like like Chrome, if you will."

    The HTML standards committee takes eternity and a day to finalize anything.

    Which is how and why workable solutions - like Flash - that evolve outside the committee gain traction.

    20% of peak hour Internet traffic in the states was a content-protected Netflix stream before Netflix offered a streaming-only service. HVEC - aka H.265 - will be ready in about two years. High Efficiency Video Coding / HEVC / H.265 : Beyond H.264 [vcodex.com]

    Half the bit rate of H.264 for content of the same quality...

    • by Art3x ( 973401 ) on Thursday January 20, 2011 @10:29PM (#34948346)

      The HTML standards committee takes eternity and a day to finalize anything.

      Exactly. Ten years passed between HTML 3 and 4. Another ten have now passed from 4 to 5, and 5 is still not an official standard.

      Meantime, requests for features and API tweaks flow in, and all browsers are going ahead and building them (even IE!). If you froze the spec after so many features, you would be drafting HTML 8 before HTML 5 became standard.

      Second, I don't know about you, but I write web apps for a living, and I've used HTML since 1997, and never once did version numbers help me. By the time I got serious, it was HTML 4. But none of the browsers posted anything like "we are an HTML 4 browser," and if they did, they lied, especially Internet Explorer. To know what worked, you tested and read about tests other people did on each the browsers.

      Finally, the term "HTML 5" has already been stretched so much to be meaningless. I'm not even talking about the journalists who use it to mean things that you could do with 4. HTML 5 is huge: Canvas, video, audio, three persistent storage APIs and one session storage API, various APIs to do with the web address, geolocation, and others. Does a browser need to call itself HTML 4 until fully implements all of these? How would that help me?

      The only thing I have ever really looked for are those comparison tables with the red and green squares, like we've always done, to figure out what to use in my web page next.

"I got everybody to pay up front...then I blew up their planet." "Now why didn't I think of that?" -- Post Bros. Comics

Working...