Slashdot Mirror
The Forgotten Macro Language of HTML: XBL 2.0
tvlinux writes "The web is becoming more than just a media display; there is more interaction and more special things that need to be done. Right now, jQuery is the preferred method of a very dynamic user interface. There is a W3 standard called XBL2.0. It is the macro language of HTML. To me it seems like a great idea — reusable HTML widgets, where each one is a separate object contained with in itself. You can define properties, methods, and events, each of which is self-contained. If the browsers supported XBL2, I can envision a whole ecosystem of new widgets, charts, grids and inputs that people could add to web pages just like any other HTML element. I see less experienced developers being able to create fancy websites by just using DOM and not having to learn jquery. My question: why is XBL dead? I think a macro-language for HTML is a good idea."
XBL is alive and well, but only for XUL. Looks like another casualty of HTML5's rejection of XML.
A Mozilla-only technology that no other engine supports doesn't really qualify as forgotten, even if someone submits it to W3C.
Did you even read any of the documents you linked in this post? XBL was never a macro language for HTML, though it could be used for that purpose. It was created by Mozilla specifically for use with XUL. They submitted it to the W3C as a Technical Note (http://www.w3.org/TR/2001/NOTE-xbl-20010223/), and their implementation didn't match the specification they submitted, so naturally it didn't go anywhere. XBL 2.0 was done properly, and is at the Candidate Recommendation stage. However, it will likely never go beyond that: not just because it hasn't for 6 years now, but because the W3C requires two complete and interoperable implementations to exist. Since all the other browsers already have their own ways to mangle CSS into non-standardness, there's not much interest in adopting it.
XUL: Do you want this <body/>?
Venkman: Is this a trick question or what?
Plain old HTML:
<li onmouseover="doSomething();">...</li>
Same thing with XBL:
<xbl xmlns="http://www.w3.org/ns/xbl">
<binding element="#nav li">
<handlers>
<handler event="mouseover">
doSomething();
</handler>
</handlers>
</binding>
</xbl>
HTML, XML, and really the whole SGML family kind of suck-- ugly syntax, annoying to hand-edit, lots of boilerplate, and the list of faults go on. The idea of writing actual programs in such a language is terrifying.
It's been superceded by Web Components: https://dvcs.w3.org/hg/webcomponents/raw-file/tip/explainer/index.html
That's why it's dead.
Those figures are for IIS 8 specifically, not IIS in general. That just means a new version was released, and while only one guy way using it 6 months ago, there's now a few hundred active instances, so the statistics fly all to hell...
Netcraft says all Microsoft web servers, combined, total just 12% of active sites, as of Feb 2013: Google's own custom, in-house web server may soon surpass IIS in market share.
Meanwhile, Apache and nginx total 67% of the market.
http://news.netcraft.com/archives/2013/02/01/february-2013-web-server-survey.html
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
If you only know Apache thats all you see acctually according to some websites IIS is the fastest growing. http://trends.builtwith.com/Web%20Server/growth#!sixMonths With nearly 10x more users than 6 months before, you are also hard pressed to find a graph which shows less than 40% IIS usage.
Step outside your bubble before you make huge sweeping statements like this.
Unfortunately the stats you refer to are for IIS8, and are linked to a corresponding decline in IIS6 as people upgrade. If you look at the stats for all versions of IIS on the site you link you will see that far from being the fastest growing, IIS usage as a whole is slowly declining.
I have tried very hard by 2 times to use Dojo to develop a web application (1st try in the 1.2 era and the 2nd try with 1.9 since a week).
I get it somehow working with the 1.2 version with some workarounds (eg: I can't get it to insanciate a message box on demand. I have to have one hidden in my page, replace its content and display it when needed).
With 1.9, I can't even get it to load some HTML to put an horizontal menu in a content panel. And all the examples I find on internet are from older versions of dojo.
If I could just find some kind of template application with a login screen and a master/detail screen with the basic ADD/UPDATE/DELETE sub screens all done within one html page. I could start from that. But for now the examples are not enough to start using dojo has it should be used. And the available examples are diluted in the google results in a way that they are nearly impossible to find :/
I contrast jQuery is simpler to use, get a lot of plugins to add what you need it to do and 'lighter' that you can split your application on multiple page without having a loading time betwwen user intercation too long making the application feel unresponsive...
I like really the design of Dojo. But the step is too high for me to use it :( I should stick to jQuery :/
Yeah, that terrible Microsoft. They have one technology, ASP.NET, which over time loses favor. As does all the similar Web technologies like JSP etc. It turns out that the whole thing with pages and code-behind is a bad idea. Along comes Ruby and Rails, and things change. Microsoft sees that this is a Good Idea(TM) and they make their own similar stuff. As does everybody else. Lots of cross-copying takes place. Microsoft makes the (arguably) best template engine (Razor) and the Java world copies that (Play! Framework).
This is the natural way of things. Only when Microsoft does it it must be evil.
Seems like the mental faculties of the average /. reader is dropping faster than one would expect something sitting in a standard 1G well.
This is how you do decent shilling. Everything he said is true, his problem was his omission. The MS stack does provide a reasonable MVC programming model for web applications. So do a number of other competing frameworks, many of which are more scalable, free (to deploy and to get the dev tools), cross-platform (do you really want to be forced to run Windows on your server? Even if you like the platform, you'll pay more if you want to deploy it in something like EC2 because of the HVM overhead), and which have been around longer.
I am TheRaven on Soylent News
No, the problem is that everything he said is off-topic
The absolute minimum a developer needs to know in order to create a web application these days is: HTML, CSS, Javascript, some programming language on the server (e.g. PHP, Java, Python) and something to store stuff on the server (e.g. MySQL, PostgreSQL, CouchDB, MongoDB). It's also nice if the developer knows how his webserver works. At the very least he should know how .htaccess files work so he can configure his web application to work the way he wants.
Then there is not really necessary but certainly useful stuff to add like an Ajax library (jQuery, YUI, Dojo etc.) and a CSS preprocessor (SASS/SCSS). And there is a bunch of other useful stuff that doesn't really require any training, but they are beneficial to the development of your project like normalize.css, modernizr and html5shiv. These things help to make your web app cross browser compatible and make sure they sorta work with old (but not too old) IE's as well.
And because this is already a lot of stuff and you dont need to invent the wheel for the millionth time, it may be nice to wrap both your clientside and serverside code into a framework. This also helps to prevent things from getting too messed up as the project grows.
For the clientside the choice is relatively limited as Javascript is the only language available. A selection of different frameworks for review is available here: http://javascriptmvc.com/
For the serverside you pick a framework based on your language. Or you pick a language based on the framework you choose. Its up to you.
So no, we dont need yet another language on top of all this, thank you very much.
Pretty good is actually pretty bad.
In other words, Netcraft confirms it: Windows is dying.
I am officially gone from
I have tried very hard by 2 times to use Dojo to develop a web application (1st try in the 1.2 era and the 2nd try with 1.9 since a week).
I get it somehow working with the 1.2 version with some workarounds (eg: I can't get it to insanciate a message box on demand. I have to have one hidden in my page, replace its content and display it when needed).
With 1.9, I can't even get it to load some HTML to put an horizontal menu in a content panel. And all the examples I find on internet are from older versions of dojo.
If I could just find some kind of template application with a login screen and a master/detail screen with the basic ADD/UPDATE/DELETE sub screens all done within one html page. I could start from that. But for now the examples are not enough to start using dojo has it should be used. And the available examples are diluted in the google results in a way that they are nearly impossible to find :/
I contrast jQuery is simpler to use, get a lot of plugins to add what you need it to do and 'lighter' that you can split your application on multiple page without having a loading time betwwen user intercation too long making the application feel unresponsive...
I like really the design of Dojo. But the step is too high for me to use it :( I should stick to jQuery :/
Consider yourself lucky. I had to refund 2 months worth of billing because my attempts to use dojo/dijit were such an abject failure that the client rejected it outright.
Dojo was really hot when it was new, but I had the misfortune to try it when it was at a low spot. It also didn't help that I was attempting to plug it into someone else's DIY JS framework and they had mis-redefined core JS functions.
jQuery has been a lot friendlier to me.
Xbl is dead because it's got a steep learning curve and is painfully abstract. Having written a fair amount of it, it took quite a while to get used to. I used while doing a bunch of xforms work with the Orbeon engine; but even they have dropped support for it as their component model. It was pretty cool, you could nest a number of XBL components together and have them render based on the data type of your XML element. An example would be an XBL phone number editor. Every time your schema used that type in your form you always got that editor for it; but debugging was impossible. It all happened in the dark on a cloudy night through three layers of fog snow rain and ice.
Just remember that ISS is mostly used on internal servers. Netcraft doesn't and can't access that information.
Not true.
IIS may be mostly used on internal servers. The ISS is as external as you can get.
>less experience developers be able to create fancy websites by just using DOM and not having to learn jquery
and you expect more experienced developers to make this happen? Look at lawyers and the law since the 1600s - when has it ever gotten simpler or easier for newbies?
He didn't mention MVC, he was talking about WebForms so I don't think what he said is true, as all attempts to make web development like forms development to date has been complete and utter crap. WebForms forces you into a paradigm that often suffers impedance mismatch relative to the way the web and HTTP actually works in practice, and often can result in unnecessary transport overheard to support the faux event-based model.
In contrast, I'd say their MVC implementation is actually quite excellent, and it scales as well as the competition- it's only really a handful of the Java alternatives that may scale slightly better (or a custom C++ solution, but that has it's costs in terms of increased development time and higher potential for security nightmares and serious bugs). Just about all of the other alternatives scale much more poorly - Ruby, PHP, and yes, I'm afraid even Python right now (even though it's a great language in it's own right). As such I wouldn't really say ASP.NET MVC has scalability as a concern, it's not one of it's weaknesses relative to the market in general, in fact, it's one of it's strengths in that it's the only true mainstream alternative to Java in most cases. The other things you say are however largely valid:
- Cost is certainly true, though if scalability is an issue you will surely be working on something with a large enough budget that the cost of IDE is a negligible fraction of the total cost of development. This is also a cost that may even be paid for in increased developer productivity through having decent tools. Similarly the increased cost of Windows Server hosting isn't likely to be an issue if you've got a product that requires some reasonable degree of scalability, the additional cost is once again going to be a tiny fraction of your total costs, though if you're really trying to maximise profit margins this will have some impact. The cost thing is of course an issue if you're a lone developer or a small start-up, that's certainly true.
- Cross platform, obviously a bit of a problem here and I agree. If you don't want to use Windows I wouldn't even bother looking at a Mono based solution, by the time you're done fucking around with Mono you may as well have just used Java or whatever in the first place.