MozillaZine

New Favorites Plug-in for K-Meleon

Tuesday January 21st, 2003

Joe Elwell writes: "The original light weight Mozilla based browser K-Meleon released a new Favorites plugin late last week. It boasts faster loading speed and new sorting options. Here's hoping that means a K-Meleon .8 is right around the corner."


#12 Re: Re: k-meleon/chimera

by asa <asa@mozilla.org>

Saturday January 25th, 2003 2:06 PM

You are replying to this message

Oops, I didn't notice the switch to talking about KHTML. I thought we were still talking about K-Meleon. I misread ""KHTML still appears to value performance and footprint over standards compliance which I think is a better weighting since that's more what users care about. " and my response was talking about K-Meleon developers, not KHTML developers. My response was trying to state that K-Meleon developers aren't doing anything about the rendering performance and footprint since they're simply using the existing Gecko which does happen to have a large number of Mozilla developers working on footprint and performance every day. Gecko is Mozilla's largest component and the area that needs the most performance and footprint work.

If K-meleon really, really cares about performance and footprint above all else then why aren't they embedding mshtml since they're windows only and mshtml would give them some more performance and smaller footprint? The answer is probably that they think Gecko is better. I agree with them and applaud their choice. I think it's much better. How is it better? It supports a lot more than mshtml and it supports that much more _correctly_ and across a dozen platforms identically. Doing that takes a larger and in some places slower engine. Doing more work simply takes more time, more memory and more cpu. We could decided that images aren't worth supporting in Gecko just like MS has decided certain w3c standards aren't worth supporting in mshtml. If we didn't support images I'll bet our rendering would be considerably faster than that of IE. But peformance isn't the only goal of this project and standards support plus support for the millions of non-standard pages is a major goal that in a lot of cases takes precedent over performance. If that wasn't the case then we could stop rendering images and <H1>s and fixed-position divs or whatever might buy us some performance.

The k-meleon developers clearly care about more than speed and performance. They could have gone with mshtml as a rendering component and they didn't. I think that's great and have always wished them nothing but the best of luck in developing a successful gecko-based product (just like I have for OEOne, ActiveState, Galeon, Beonex, and any of the other not-mozilla.org sponsored projects). If they make something that appeals to a lot of users they will have helped put more Gecko user agents on the web and that's good for standards, good for the web, and good for mozilla.org.

--Asa