LaurensVH - Slashdot User

Microsoft Complains About Google's Monopoly Abuse 384

Posted by CmdrTaco on Wednesday December 26, 2007 @01:18PM from the play-fair-or-i'll-tell-mom dept.

I Don't Believe in Imaginary Property writes "Frustrated at the FTC's blessing of the Google/Doubleclick merger, Microsoft is complaining to the EU. Its latest filings detail how the merger would give Google a stranglehold on the advertising industry. While these complaints aren't new, the diagram [PDF] Microsoft created gives you an interesting look at the sort of competition Microsoft fears from Google."

Submission + - Splitting dynamic and static parts of webpages?

Submitted by

LaurensVH

on Monday June 11, 2007 @02:22PM

LaurensVH writes: "While recently philosophising on how cool mod_parrot would be, I suddenly realised it would be even cooler if we took the idea one step further.

If anything, I'd say some of the most exciting stuff to happen in web development recently is all based on splitting up different tasks. First, markup was separated almost entirely from design through the wonders of CSS, used in combination with (X)HTML. I think we can safely say we're all glad we've managed to get rid of tablular page layouts. CSS implementations in some browsers (and especially non-browsers) are still a bit lacking (most notably ACID2 gets royally messed up). At any rate, they differ. Enough to force web designers (or developers, in cases where they overlap) into restorting to ugly kludges to get their shiny stylesheet to render a webpage correctly in most common browsers. For now, it's the best we have.

The second big advancement came with the dawn of the MVC model. MVC stands for "model, view, controller", a design paradigm applicable to dynamic web pages and their development. While CSS and (X)HTML separated markup from design, MVC separates the data model, the code that operates on the data model, and the way the content gets displayed (or, more accurately, gets passed to the HTML markup, where it eventually gets prettied up by CSS).

However, I feel there's room for even more improvement. Or, at the very least, there's plenty of interesting development still left to be done.

There are a number of really cool lightweight web servers out there, such as lighttpd (pronounced: "lighty"), fnord, and gatling. Those last two, besides both living at fefe.de, focus on delivering static content. A lot of static content, blazingly fast, while keeping server load minimal.

Most of you are probably thinking: "Well, obviously... But static content is so boring. Sure, we can put our images and even dynamically generated front pages made static (think lazy caching), but what's new?" right now. Well, if this is all you're going to be doing with it, yes, you're right, it ends here.

Imagine, however, that you combine this with the already existing technology of AJAX (well, more or less AJAX). Imagine all relevant static data, including information on how to get the dynamic data, beings erved by an extremely fast and efficient static web server.

The client then executes the javascript code that gets the dynamic data from a specialized "web" server. I'm not sure that's the correct term, because, in theory, it should never display any web pages. It simply returns JSON (or whatever object format is handiest in your particular setup) objects that the Javascript code uses to fill in the website content.

The most obvious downside is the Javascript requirement. It's the only feasible technology I'm aware of. It would be really cool if we could do this in arbitrary scripting languages. You could do this using XUL, but that isn't nearly as availible or commonplace as Javascript is, unfortunately.

The first person I told this said that superfluous AJAX usage has dramatic influences on website responsiveness, but that's in cases were you're fetching pieces of web page somewhere in the middle of user interaction with it. What I'm suggesting is doing it in on_load, meaning more or less the same amount of data has to be transferred.

Another issue is that it's hard to develop pages like this here and now, mainly because there are no specialized servers that only return objects at the time of writing. I'm not here to impress people with cool existing technology, I'm just trying to see what could be done with it in the future. Regarding the "hard to develop" argument, I can only imagine that very, very good toolkits will emerge if this idea is good enough.

So, in closing, dear Slashdotters, I'd like to ask you: do you think this would work? Am I on to something? Is the end of Apache nigh? Or should I go and get started on mod_parrot? ;-)

Thanks for listening.
Laurens Van Houtven (lvh \at\ laurensvh \dot\ be"

Slashdot Top Deals