It will (in theory) emulate <nolayer>, <layer>, <ilayer>, an the Layer() constructor from NS4 in Mozilla, saving some DHTML from a rewrite. I also am doing it to kind of prove a point about how powerful Mozilla is.
If you want to see some semi-working examples already (the project was just announced today at mozdev.org ), find me on AIM or MSN Messenger...
Why post it in this forum though? Good question, since there is no DHTML forum it appears, and it is not solely a JS-related topic, not to mention this *is* emulating some old-fashioned proprietary HTML tags .
I know exactly what you're saying Javaman. But with Mozilla finally > 1 (1.1 Alpha was just recently released, and branched a while ago ), and with AOL seemingly about ready to implement Gecko, more and more people will be using the awesome browser. But of course, that means more and more uneducated people complaining "oh, I'll never use Gecko because none of my poorly coded cut-n-paste, err, I mean my excellent DHTML works!"
By suppyling the legacy support, I hope to eliminate some of those complaints and turn more users on to Gecko.
I mean, what kind of browser allows a developers to fully emulate a pre-existing model without touching the source code of the browser? A very well-thought out one.
I think that supporting legacy browsers is beyond useful, it is absolutely essential. There's serious need for a project like this, because at this particular point in time, enough people still use netscape 4 that supporting it is necessary (personally I still support netscape 3 ...) But for someone who's only just starting to learn DHTML, it can seem, and indeed is, a daunting prospect to learn the ins and outs of programming for ns4.
I should know ... it took me long enough to grock it ... and some of it's more hardcore features are so bizarre to implement, it's hard to remain sane at times
As for the point you're proving .... well, I've made my views clear enough. It's just the rendering .. that's all ... I just don't like the way things look in it; there's a harsh quality to its rendering that I just can't get comfortable with.
Just a quick status update (current files have *not* been uploaded through CVS yet so it only exists on my hdd), but I have a *very* usable emulation going.
There are only two significant problems I need to tackle, both of which are easy:
1. Implement clip and child properties (that's by far the most important)
2. Fix my src and load() implementation - it kinda works but doesn't at the same time.
I think pretty soon I can take it to the test - against the often-ugly and just as often poorly coded DynamicDrive scripts .
Sorry to preach here, but I wish you would show a bit more respect. Some of the scripts on that site may be a bit hacked or sloppy, but they work, and that's the bottom line.
I wonder if you appreciate the day to day realities of maintaining a public web site. It is not practical to take such a high-minded, academic approach, however "correct" that might ultimately be.
You know, I don't see your name next to any of the scripts on that site
Sorry to go off topic ... apart from that minor irritation, full kudos to you mate
You made your point. True, they work, and that's what really matters to most people. But whenever I look at the code for one, I see so many ways it could have been written better, performance-wise, variable-wise, etc etc.
Of course, not all of them are like that, but a good many are, and when people start complaining about script conflicts with DD scripts, I just sigh.
And as for my name not appearing next to any scripts (), I write code all the time, but I never document it nor explain it well enough for other people to use it. I'm trying to change that - my LayerEmu code is becoming highly commented, and I *need* to write a detailed page describing what properties/methods it emulates 100%, kinda, and not at all.
But I will try to stop dissing Dynamic Drive so much, as you did make a few good points.
Originally posted by jkd
I write code all the time, but I never document it nor explain it well enough for other people to use it. I'm trying to change that - my LayerEmu code is becoming highly commented, and I *need* to write a detailed page describing what properties/methods it emulates 100%, kinda, and not at all.
Yeah man. Good documentation make any script twice as useful, even though it takes a long time. I find it helps with development - writing a README you might find yourself saying something like, "if you do x then y will hapen because such and such a behaviour is not supported .." and then thinking, but hang on, maybe that can be supported Simply writing the documentation helps in development because the process of explaining something to someone else helps you to step away from it, as it were, and see it from a user's perspective.
Like I said before, you've hit on something that a lot of people will want. I almost hate to mention $ ... but you could definitely make some
Are two other projects aiming to do the same thing, with different implementations. As a matter of fact, the owner of the latter project posted to the LayerEmu mailing list offering usage of any of his code. Nice guy...
Anyway, at the rate this is going, I could have 95%+ emulation done by Sunday .