Google Chrome



  • Don't know how many of you have heard of this.  Looks like it's pretty recent news: http://googleblog.blogspot.com/2008/09/fresh-take-on-browser.html



  • thx4thelink ... haven't heard about it ...

    looks like reinventing the wheel, but who knows, they might pull it off ... 



  • Assuming it's not a complete flop, I am looking forward to this.  The web is in dire need of someone to step up to the plate and fix things (not just write white-papers filled with gibberish about "fixing" things and then bitch when the half-assed solutions aren't implemented to spec).

     



  • @NullAndVoid said:

    Assuming it's not a complete flop, I am looking forward to this.  The web is in dire need of someone to step up to the plate and fix things (not just write white-papers filled with gibberish about "fixing" things and then bitch when the half-assed solutions aren't implemented to spec).

    My thoughts exactly.  And thanks Tster for the heads-up, I've not heard of this until now (been a bit Ostrich on current developments recently), looking forward to trying it out.



  •  Its pretty obvious where they are heading. They want the browser to be just as rock-stable and reliable as the operating system that runs it. Because as far as Google is concerned, the browser _is_ the operating system. 



  • @NullAndVoid said:

    Assuming it's not a complete flop, I am looking forward to this.  The web is in dire need of someone to step up to the plate and fix things (not just write white-papers filled with gibberish about "fixing" things and then bitch when the half-assed solutions aren't implemented to spec).
     

    IMO, the browser is only half the battle.  The web is sorely in need of a complete rethinking of the underlying technologies.  HTML/JavaScript/CSS were not designed to deliver "rich" applications - it seems that building anything remotely complicated with them is an exercise of how many hacks can you come up with.  If delivering full-featured applications via the web is what we want, we need to develop technologies designed to do so; not just mash crap together.

    It's just shamefull that HTML doesn't even specify a numeric only input field.  You know how much time we all could save if we had something like <input type="numeric" min="0" max="10" float="no" />.

    In any case, I hope Google's browser provides some innovation and forces the other browser makers to either copy/one-up them.



  • @lpope187 said:

    In any case, I hope Google's browser provides some innovation and forces the other browser makers to either copy/one-up them.
     

    This seems to be the idea, as far as I can tell from the comic.



  • @lpope187 said:

    IMO, the browser is only half the battle.  The web is sorely in need of a complete rethinking of the underlying technologies.  HTML/JavaScript/CSS were not designed to deliver "rich" applications - it seems that building anything remotely complicated with them is an exercise of how many hacks can you come up with.  If delivering full-featured applications via the web is what we want, we need to develop technologies designed to do so; not just mash crap together.
     

    I think XForms is/was an attempt to build a new foundation for web applications, but as it seems, progress is incredibly slow, so I wonder if it will be fully supported by major browsers. Currently, I'd guess that AJAX frameworks have already delivered everything that XForms promises and a lot more.



  • @ammoQ said:

    Currently, I'd guess that AJAX frameworks have already delivered everything that XForms promises and a lot more.

    Flex as well ...



  • @lpope187 said:

    It's just shamefull that HTML doesn't even specify a numeric only input field.  You know how much time we all could save if we had something like <input type="numeric" min="0" max="10" float="no" />.

    What a horrible mess.  You will still have to validate the data on the server-side and instead of using a rich programming language to construct arbitrary business logic you are stuck cramming attributes into your tags.  This sounds like a big WTF.



  • @morbiuswilters said:

    @lpope187 said:

    It's just shamefull that HTML doesn't even specify a numeric only input field.  You know how much time we all could save if we had something like <input type="numeric" min="0" max="10" float="no" />.

    What a horrible mess.  You will still have to validate the data on the server-side and instead of using a rich programming language to construct arbitrary business logic you are stuck cramming attributes into your tags.  This sounds like a big WTF.

     

    But it could allow the user to customize how the browser enforces (or even if it does enforce) the tags.  I think there are a lot of possible additions to HTML that could be added.  And it's not like it's a security thing or anything.  It would just allow for a better user experience.  The "numeric" is kind of like a hint to the browser.  .NET controls have text boxes where you enter in the format of the data that is expected and it makes well written .NET applications very user friendly.  I wouldn't mind web pages being more like that.



  • @morbiuswilters said:

    What a horrible mess.  You will still have to validate the data on the server-side and instead of using a rich programming language to construct arbitrary business logic you are stuck cramming attributes into your tags.  This sounds like a big WTF.
     

    Any specification for how to ask the user for input could be considered "business logic".  A drop-down list of values, for example, is often generated server-side via business logic.  None of these input specifications replace server-side validation, they are just instructions for the web browser on how to prompt the user for the input. 



  • @Jeff S said:

    @morbiuswilters said:

    What a horrible mess.  You will still have to validate the data on the server-side and instead of using a rich programming language to construct arbitrary business logic you are stuck cramming attributes into your tags.  This sounds like a big WTF.
     

    Any specification for how to ask the user for input could be considered "business logic".  A drop-down list of values, for example, is often generated server-side via business logic.  None of these input specifications replace server-side validation, they are just instructions for the web browser on how to prompt the user for the input. 

    And a drop-down is a fairly simple idiom that can be handled quite well in HTML.  Whatever options are acceptable are just there.  Comparing this to the arbitrary logic allowed by a real programming langauge is absurd.  Imagine for a moment if your numeric field was only numeric (or only required) under certain circumstances.  This would have to be represented in the HTML add-ons.  HTML could just ignore these cases, but a lot of developers will opt for Javascript over the new HTML form elements simply because they won't want to scrap the whole thing when the requirements change and HTML is no longer capable of working for them.  So either you end up with a very simple addition to HTML that only works in a few cases, requires a change from existing practices to utilize and a switch back to existing practices if the requirements change enough to justify it and ends up being about as verbose as the JS required to do the same validation OR you end up with a ridiculously complex spec that essentially emulates a programming language in XML just to give developers the functionality they need.  Sounds like a massive WTF to me.  Finally, by using JS to do your own validation you also gain the ability to have errors handled within your GUI instead of having goofy "standard" errors built into the browser such as "Field xyz is required to be a number that is between the values of 1 and 10 and does not have a decimal."  The browser obviously has no way of knowing why the restriction is placed on the field and thus cannot provide useful assistance to the end-user.



  •  I've found hints that <input type="int"> already exists, though firefox, opera and konqueror seem to ignore it...(IE not tested)



  • @morbiuswilters said:

    And a drop-down is a fairly simple idiom that can be handled quite well in HTML.  Whatever options are acceptable are just there.  Comparing this to the arbitrary logic allowed by a real programming langauge is absurd. 

    Luckily, no one here is suggesting adding "arbitrary logic allowed by a real programming langauge" to HTML.



  • @Jeff S said:

    Any specification for how to ask the user for input could be considered "business logic".  A drop-down list of values, for example, is often generated server-side via business logic.  None of these input specifications replace server-side validation, they are just instructions for the web browser on how to prompt the user for the input. 

    Um... nope. That would be the "presentation" layer, which handles how the interface is presented to the user. Business logic means stuff that actually runs the business, like "bill order 666 to customer 101", "create order 493 with items 4,2,1 and 9"; which in a proper MVC environment will run server-side, and will do it properly no matter what its fed from the presentation layer.

    Client-side validation is OK, but it isn't something you can rely on; anyone could just build up a copycat POST form and send unvalidated data like Feb 66, 9999 and such.



  • @danixdefcon5 said:

    Um... nope. That would be the "presentation" layer, which handles how the interface is presented to the user. Business logic means stuff that actually runs the business, like "bill order 666 to customer 101", "create order 493 with items 4,2,1 and 9"; which in a proper MVC environment will run server-side, and will do it properly no matter what its fed from the presentation layer.
     

    You quoted and interpreted what I wrote completely out of context. 

    I completely 100% agree with what you wrote. 

     



  • @Jeff S said:

    Luckily, no one here is suggesting adding "arbitrary logic allowed by a real programming langauge" to HTML.

    I wouldn't suggest that because it would probably end up looking like the WTF known as XSL. I can understand why we ought to separate display elements, presentation of the elements, client-side behavior, server-side behavior, server-side data, etc. However, I do not understand why these things must be split amongst five or more different languages. I can see how it got that way (they were all developed independently of each other), but I don't think it should stay that way. Is there a super-language that can handle it all? I have asked this elsewhere and the best answer I got was Lisp, but sadly, I don't think that's gonna fly.



  • @morbiuswilters said:

    @lpope187 said:

    It's just shamefull that HTML doesn't even specify a numeric only input field.  You know how much time we all could save if we had something like <input type="numeric" min="0" max="10" float="no" />.

    What a horrible mess.  You will still have to validate the data on the server-side and instead of using a rich programming language to construct arbitrary business logic you are stuck cramming attributes into your tags.  This sounds like a big WTF.

     

    I was thinking more along the lines of providing a better way to provide limited client-side restrictions along the lines of "don't even allow anything but numeric characters to be entered".  Honestly, I try to put just basic validation on the client side (make sure it is a date, number, etc), and put the same basic rules as well as the complex business rules on the back-end.  Just like any other system, validation still needs to be done at the back-end and/or middle tiers. 



  • @NullAndVoid said:

    Is there a super-language that can handle it all? I have asked this elsewhere and the best answer I got was Lisp, but sadly, I don't think that's gonna fly.
    There is already one language that handles it all, and has done it since at least 1999.

    Java.

    Check out the J2EE 1.4 specs, for example: you put all the business logic on an application server on the server-side; then you build, say, a Swing client that connects (and authenticates) to the AppServer. You get a nice client-side GUI doing the presentation stuff, and all the BL on the server-side using EJB's.

    Too bad the AJAX craze came along; now everything seems to be centered around Web 2.0, ripping off the Aqua look-and-feel, and making the browser do things it was not designed for. So now J2EE (and the next EE's) are mostly being used in web-only apps, using things like Hibernate and Spring so the client/server implementation I mentioned earlier becomes impossible to implement if the app was made using these tools.



  • It took some time, but here's an article from The Register which basically explains the Web 2.0 evolution, and how the thing has grown into a complex mess. (The third UML-ish "diagram" is priceless.)



  • first 5 minutes:

    1. This WYSIWYG editor is totally fucked.

    2. middle mouse doesn't do autoscroll. Have no idea how to turn it on (assuming it is possible)

    3. Javascript alert still takes over the entire mother fucking browser.

    4. I find this kind of silly

    @google said:

    Going incognito doesn't affect the behavior of other people, servers, or software. Be wary of:

    1. Websites that collect or share information about you

    2. Internet service providers or employers that track the pages you visit

    3. Malicious software that tracks your keystrokes in exchange for free smileys

    4. Surveillance by secret agents

    5. People standing behind you

    • It seems to use less memory then it's overweight brother.

    • Movement and everything is fast and smooth.

    • Bookmarks aren't easy to get to unless you open a new tab.

    • doesn't use much real estate at all.

    • Not very many settings.

    Overall:

    Pretty good for a release that works and shows the major concepts. Not something that I would actually want to use in it's current state given other options. Firefox 3, IE7, and Opera are all better in their current state. But I will give this thing a shot later.



  • @danixdefcon5 said:

    @NullAndVoid said:

    Is there a super-language that can handle it all? I have asked this elsewhere and the best answer I got was Lisp, but sadly, I don't think that's gonna fly.
    There is already one language that handles it all, and has done it since at least 1999.

    Java.

    Check out the J2EE 1.4 specs, for example: you put all the business logic on an application server on the server-side; then you build, say, a Swing client that connects (and authenticates) to the AppServer. You get a nice client-side GUI doing the presentation stuff, and all the BL on the server-side using EJB's.

    Too bad the AJAX craze came along; now everything seems to be centered around Web 2.0, ripping off the Aqua look-and-feel, and making the browser do things it was not designed for. So now J2EE (and the next EE's) are mostly being used in web-only apps, using things like Hibernate and Spring so the client/server implementation I mentioned earlier becomes impossible to implement if the app was made using these tools.


    Turns out people don't like it when a web page loads up stupid Java webstart. It takes a lot longer than loading a regular site, and it reduces immersion. Furthermore, I don't know if you've noticed this or not, but most professional websites look a lot cooler and more customized that most profession Java applications. Furthermore, it allows you to save certain parts of the web site as a bookmark or send a link to a friend.



  • @tster said:

    1. This WYSIWYG editor is totally fucked.

    TRWTF is CS.

    Anyway, I like Google Chrome. Nice screen layout, fast. The concept looks good to me.



  • @tster said:

    1. This WYSIWYG editor is totally fucked.

    I've only used the editor here a few times, so it hasn't ruined any habits of mine, but I can confirm that the control is broken/missing.

    @tster said:

    2. middle mouse doesn't do autoscroll. Have no idea how to turn it on (assuming it is possible)

    Middle mouse button scrolling works here.

    @tster said:

    3. Javascript alert still takes over the entire mother fucking browser.

    I opened up the built-in JavaScript console with autocompletion and confirmed this fucking annoying behavior.

    @tster said:

    4. I find this kind of silly

    The missing WYSIWYG (which is apparently a fairly widespread plugin, meaning perhaps thousands of sites are broken) is very silly.

    @tster said:

    6. Movement and everything is fast and smooth.

    Indeed. Maybe even too fast in places. One quirk is that clicking a link immediately blanks the page before loading the new page. This is good in the sense that it may cut down on double-submit oopses, but the effect is jarring.

    @tster said:

    7. Bookmarks aren't easy to get to unless you open a new tab.

    Ctrl-B toggles a bookmark bar with drag-n-drop.

    @tster said:

    8. doesn't use much real estate at all.

    I am not sure if it is possible to make a less intrusive interface.

    @tster said:

    9. Not very many settings.

    It's worth noting that "send usage statistics to Google" is enabled by default. Not that they don't already know everything I do.

    There are several FireFox plugins that I have come to rely on, so I won't be able to "make this my default browser" until this void is filled. The JavaScript tool is nice, but the usefulness is limited without some of the other web tools. I hope the developers of those plugins are impressed enough with Chrome to port their stuff pronto.

    Overall, I was very impressed with this browser and I think it has a strong future.



  • @NullAndVoid said:

    It's worth noting that "send usage statistics to Google" is enabled by default. Not that they don't already know everything I do.
     

    not here ...

    i got a checkbox "Optional: Help make Chrome better by automatically sending usage statistics and crash reports to Google." when downloading it and it was off by default...



  • it seems that chrome has found the flash plugin on my computer or it comes with the preinstalled one, anyhow i did not have to install it separately ...



  • @NullAndVoid said:

    Middle mouse button scrolling works here.
    What OS are you running? I was not able to get middle mouse button scrolling to work, either, which is by far my biggest problem with Chrome since I have a habit of using this functionality a lot in everything I do. It might be that this is just a problem with Chrome in Vista. I'm running Vista HP 64.



  •  @lpope187 said:

    @NullAndVoid said:

    Assuming it's not a complete flop, I am looking forward to this.  The web is in dire need of someone to step up to the plate and fix things (not just write white-papers filled with gibberish about "fixing" things and then bitch when the half-assed solutions aren't implemented to spec).
     

    IMO, the browser is only half the battle.  The web is sorely in need of a complete rethinking of the underlying technologies.  HTML/JavaScript/CSS were not designed to deliver "rich" applications - it seems that building anything remotely complicated with them is an exercise of how many hacks can you come up with.  If delivering full-featured applications via the web is what we want, we need to develop technologies designed to do so; not just mash crap together.

    It's just shamefull that HTML doesn't even specify a numeric only input field.  You know how much time we all could save if we had something like <input type="numeric" min="0" max="10" float="no" />.

    In any case, I hope Google's browser provides some innovation and forces the other browser makers to either copy/one-up them.

    Don't forget HTML 5, the "counter project" to XHTML 2. They basically try to take the whole mess of abused HTML, tag soup, javascript and CSS and freeze it into a spec, while adding a number of new features that should ease the creation of "web applications".

    The whole activity is VERY practice-driven, with backwards compatibillity and and fullfillment of concrete real-world use cases two of their top priorities. 

    The downside of this is of course that the result is a patchwork monstrosity without a really consistent internal structure and without any clearly outlined "purpose". But maybe that's what our internet today needs, who knows... 



  • @Welbog said:

    @NullAndVoid said:

    Middle mouse button scrolling works here.
    What OS are you running? I was not able to get middle mouse button scrolling to work, either, which is by far my biggest problem with Chrome since I have a habit of using this functionality a lot in everything I do. It might be that this is just a problem with Chrome in Vista. I'm running Vista HP 64.

     

    We need to make sure we are refering to the same thing.  I am talking about "autoscroll" which is when you press the middle mouse button down and then you move your mouse up or down and the little arrows come up and the screen moves smoothly.  Middle mouse scroll works just fine for me, but I don't like reading blogs with stupid jaggy scroll.  (for the record, I am using Vista)



  • @tster said:

    I am talking about "autoscroll" which is when you press the middle mouse button down and then you move your mouse up or down and the little arrows come up and the screen moves smoothly.
    That's what I'm talking about, as well. You think maybe NullAndVoid is talking about just regular wheel scrolling?

    The broken autoscrolling and the fact that I can't disable tabs (or at least make middle-click open a new window) to me make Chrome not worth my time. I used it for my usual browsing for about two hours last night and liked a lot of what I was seeing, but since I have a preference for new windows rather than tabs and I couldn't scroll around in my usual way I concluded that Chrome is not for me.



  • @Welbog said:

    The broken autoscrolling and the fact that I can't disable tabs (or at least make middle-click open a new window) to me make Chrome not worth my time. I used it for my usual browsing for about two hours last night and liked a lot of what I was seeing, but since I have a preference for new windows rather than tabs and I couldn't scroll around in my usual way I concluded that Chrome is not for me.

     

     Well, what you CAN do is to drag the tab outside the window, then the tab will "morph" into a window of its own. I agree however thst this is pretty cumbersome if you don't want tabs at all.

    There seem to be a lot of hidden "about:" pages (last but not least "about:internets" ;)  so maybe there is a pendant to "about:config" too?



  • I'm checking out TDWTF from Chrome.

    The WYSIWYG editor is b0rked; I can't see the toolbar but somehow, it still works. Ok, at least better than the mysterious IE errors I used to get.



  • @Welbog said:

    @tster said:

    I am talking about "autoscroll" which is when you press the middle mouse button down and then you move your mouse up or down and the little arrows come up and the screen moves smoothly.
    That's what I'm talking about, as well. You think maybe NullAndVoid is talking about just regular wheel scrolling?

    The broken autoscrolling and the fact that I can't disable tabs (or at least make middle-click open a new window) to me make Chrome not worth my time. I used it for my usual browsing for about two hours last night and liked a lot of what I was seeing, but since I have a preference for new windows rather than tabs and I couldn't scroll around in my usual way I concluded that Chrome is not for me.

    My guess is that my driver isn't sending teh same codez. I'm using a ThinkPad (Vista) and for me, I hold down the middle button, a miniature scrollbar appears, and I can scroll with the nipple-thingy. No smooth scrolling option, but it does work. Maybe it's emulating a wheel.

    For autoscroll, I swipe the edge of the touchpad and it quickly scrolls through the whole document (usually by accident).



  •  @danixdefcon5 said:

    I'm checking out TDWTF from Chrome.

    The WYSIWYG editor is b0rked; I can't see the toolbar but somehow, it still works. Ok, at least better than the mysterious IE errors I used to get.

    You probably only think it's working.  Try posting something as a list with 1 item on each line.



  • The cool thing is that chrome (like IE8) creates a separate process for each tab ... If one crashes (or has anoying endless javascript alerts), the others are intact ...

    The problem is that if you have a lot of tabs, the tasklist tend to get really messy ...

     



  • @Nelle said:

    The cool thing is that chrome (like IE8) creates a separate process for each tab ... If one crashes (or has anoying endless javascript alerts), the others are intact ...
    And we've also heard that alert() holds down the whole browser. So which is it?



  • @wybl said:

    And we've also heard that alert() holds down the whole browser. So which is it?
     

    kind of both ... 

    it seems that the javascript engine runs in the main chrome process and the alert box is not bound to the originator ...

    i just  loaded a simple page showing an alertbox, and then I killed its process without confirming the alert ...

    [URL=http://img526.imageshack.us/my.php?image=chromealertqh4.jpg][IMG]http://img526.imageshack.us/img526/7928/chromealertqh4.th.jpg[/IMG][/URL]

    if one tab has endless alerts you can kill its process without disrupting other tabs, but its last alert remains alive until you click it ...

     



  • @Nelle said:

    if one tab has endless alerts you can kill its process without disrupting other tabs, but its last alert remains alive until you click it ...

    At least for me, it will show a "don't show more alerts from this page" check box after a script has popped up a certain number of alerts. So you can just get of an "alert loop" without killing any processes at all.

    Unrelated: Another nice feature: It allows the user to resize <textarea> elements, even if the page itself did nothing about that: try [code]data:text/html,<textarea>foo</textarea>[/code]



  • Typing in :% in to the nav bar nukes the whole of chrome. Nice bug that one, still most likely won't effect anyone.



  • @Jimmy Savile said:

    Typing in :% in to the nav bar nukes the whole of chrome

    Ah, nicely put.

    Downloadzored!

    I'm upset that it doesn't do middle-click-accelerated-scroll. It's good on long pages, such as TDWTF. Opera and IE have inconvenient completely unusable implementations of that, respectively: IE's acceleration curve approaches a sixth-order function that completely defeats the feature (i.e. slowslowslowslowWARPDRIVE), and Opera doesn't display the origin point and has a similarly poorly tweaked acceleration curve. Opera's scroller also doesn't release when you let go. FFX's is the only one that actually works.

    Textareas have a dragthing, just like Safari. Nice. No more Firebug hacking to increase tiny fields.

    Tabs touch the top of the screen = infinite tab size = fast mindless tab switching = good interface

    Its inspector borrows more from the default one in Safari than from Firebug. The Network tab equivalent is almost a copy/ripoff.


Log in to reply