On Mon, 02 Jul 2001 21:33:14 -0500 esteemed Greg Miller did hold forth
thusly:
>
> But what parts do you compare? In which operations should performance
> hold up the 1.0 release?
First and foremost it is a browser. Go to some web sites and go thru a series
of pages. I peridiocally do this with Moz and IE and use the NT Task Manager
to see how much CPU time each uses. I go to some site on both of them. Click
thru the same series of pages in parallel and then see how much CPU time each
takes. I've done this on several sites and Moz usually takes about twice as
much CPU time.
I also go to a page on both browsers, click on a link using Moz, then click
on it using IE and then watch as IE finishes loading and displaying it first
even though Moz got the head start.
I also right click and open new windows and watch how long Moz and IE take to
open new windows. Well, IE is faster again.
Of course, Opera is faster than either of them.
>
> How do you set a requirement? If you set it relative to another browser,
> you get lots of cases where release would be held up because of
> differences that aren't even user-visible. Do you set some arbitrary
> line on jrgm's graph and say "it has to cross this"?
Time how long it takes to do things. I personally see very noticeable
differences subjectively. Then when I use clocks and CPU timings in NT Task
Manager I see hard data confirmation of this.
> It doesn't appear slower in page loading to me, and I'm on a pretty slow
> machine. It's certainly slower that Opera in many areas, but it's still
> quite fast. Of course, all the major browsers are pretty fast applications.
Open IE and Moz side by side on the same page and click either IE or Moz and
then the other one and see which loads the next same page first. IE always
wins when I do it.
> Why? I'd call it easily the best browser out there. It's quick (even if
> it's not as quick in many areas as some of the others) and the page
> display is a pleasure to work with (far better than the alternatives).
Its slow. Its still flaky.
The reluctance to compare seems unwise. When Moz is released I'm sure lots of
reviewers will have no problem doing all sorts of comparisons and some of
those comparisons will be quite quantitative.
