Slashdot Mirror
Mozilla Rising ... As A Platform
ceswiedler writes "Salon is running a story about Mozilla's potential dominance as a platform for application development. They discuss the community development centering around Mozilla, and point out that its cross-plaform GUI environment is 'exactly the kind of thing Microsoft was trying to prevent when it launched its war against Netscape. It didn't want Netscape around, because Netscape was becoming a platform.' In what might be a Salon first, they even include a reference to a Slashdot comment by SkyShadow."
it seems that mozilla, as a whole, will evolve into a framework of reusable components that will transcend the browser application itself.
this will pose to be a problem for microsoft; why bother using microsoft components, which are bound to windows, when i can program across multiple platforms using mozilla components?
My vote is for SVG, even though the current support for it in Mozilla is pretty fragile [YMMV, I'm on 1.1 Linux].
With full support for SVG, Web applications could really take off in a big way (graphical and not just text interaction) that is unhindered by platform specific nonsense.
One big hitch though seems to be in rendering quality outline fonts. Everyone would love to have the precision of PostScript for determining exactly where text is located, how far it extends, etc, but there seems to be big players that are nervous about releasing outlines of their fonts and have punted about precise layout of fonts inside SVG, deferring to upper level CSS specifications and what not that permit layout decisions to change when we really need a web layout engine that doesn't change from platform to platform (and is free and open).
"Provided by the management for your protection."
I think that in order for it to really drive the nail in the coffin, it's going to need a niche market. Incredibly good functionality really isn't enough to make the average user go out of their way to get it. The future is likely in the ability to discover the niche application that makes it undeniably more useful -- then all it has to do is hang on for a couple of years (which is harder than it sounds...)
Sure, there is initial appeal to having your application look the same on all platforms. Who really wants to write the same application N times? However, cross-platform consistency isn't necessarily a good thing.
Each platform has its own quirks with how it should behave. For example, menus in Windows are expected to be static (that is, they stay visible after the user releases the mouse button), while Macintosh menus tend to be rubber-band (menu disappears when user releases mouse button). In Windows, a menu action simply happens while on Macintosh, the selected menu item flashes several times.
I could go on and on with the differences between the Windows and Macintosh platforms (to say nothing of UNIX!). The point is that an application that acts differently from every other program is an application that is harder to learn. Users are forced to keep two sets of expectations, which completely defeats the purpose of using a cross-platform GUI!
Yes, you can tweak the UI so that it looks more like the host operating system. This is a thin veneer, however, as the emperor's proverbial clothes come into view when the OS theme is changed.
It makes sense that the UI should be abstracted from the rest of the application, but XUL is not the answer.
Nathan
I like XUL. I think it's a great idea and the implementation rocks. But most of all, it's simple. There are no DLLs, no IUnknown pointers or registry issues to deal with. Mozilla is a great browser, in many respects superior to IE, and in some inferior (my dream browser would be a combination of IE, Mozilla and Konqueror which runs on Windows, OSX and Linux. Oh well). But the difference it was designed from the sart to *be* a platform, where with IE platformitis was an afterthought.
But I disgress. The key here is going to be Mozilla's ability to gain critical mass with average developers in Windows for it to take off. I'm not talking about XPCOM hackers, I'm talking about the ones quoted in the Salon article. It will do Mozilla no good if it takes off in Linux, because Linux has no desktop presence to speak of, and it has a far greater variety of browsers that, while good for competition, also cause fragmentation.
I think Microsoft's response to this (if they do get to the point where they consider the Mozilla *platform* a threat) will be to essentially take IE and turn it into a .NET platform. If they can offer a platform to people writing C# and VB.NET and JScript.NET, they'll be all set - assuming the .NET thing does take off like they want to. Of course, one of the catalysts to .NET acceptance will be how many computers it happens to be installed in - imagine if anyone who wants to use the next version of IE has to download the .NET runtime?
Still, Mozilla has the upper hand because it's off on the race and Microsoft is standing in the starting line wondering what the futz is going on and why are all these geeks cheering?
Cross platform is irrelevant when you have 95% of the market. That's like saying that a business isn't viable unless it gets 100% of potential customers. That's ridiculous. Ever notice that there isn't just one gas station in a city? How about restaurants?
Hell, I own my own retail store, and I get about 60% of all potential customers in my area, and that's more than enough. Why would I bother watering down my product line to pull in everybody else? It'd ruin my business. One size doesn't fit all in ANY product.
Non-commercial developers? What does that have to do with anything? I write apps all the time for myself using IE. What's your point? There aren't any restrictions if you write a shrink-wrapped app that grabs a few IE objects. You don't have to license IE. You just specify that IE is required to be able to use your product. And with a nearly 100% saturation on the Windows platforms, which have a 95% desktop saturation, that's not a problem.
But I want real examples where people were fully unable to continue or cope because of an "inconsistent user interface".
Since when was "no users give up" the only criterion for evaluating a program's usability?
The shareholder is always right.