JTK wrote:
> Ian Hickson wrote:
>
>>On Fri, 22 Jun 2001, JTK wrote:
>>
>>>As for the "4 times", it was stated by somebody in .performance only
>>>a week or so ago to be ~8 times slower [than IE].
>>>
>>Yes. Things are moving fast.
>>
>>
>
> At the speed of light, I'd say! As in three light-YEARS and counting.
Do you know that a light year is not a measure of time but a measure of
the distance that light will travel in 1 year?
>
>
>>>I have no way to directly compare myself, so I can only repeat what
>>>I've heard from others.
>>>
>>On local network stress testing of page loading on an ~800 MHz Windows
>>2000 machine with 512MB of RAM (IIRC), IE scores in the 200ms per page
>>load range and recent Mozilla builds score in the 800ms per page load
>>range. (These are the results for jrgm's tests on hyatt's machine.)
>>
>>
>
> Are you referring to your post,
> <[EMAIL PROTECTED]>? That's
> actually the one I was thinking of too! I love this part:
>
> "
>
>>>Here it is in an easily digestible nutshell.
>>>
>>>SeaMonkey - 830[ms]
>>>MFCEmbed - 700[ms]
>>>
>>>Difference - 130[ms]
>>>
>>All in all I think this is a very acceptable differance, though that's not a
>>reason to [not] try to decrease it...
>>
>
> To put this in perspective... 130ms is about 40% of the TOTAL time for
> IE6
> on the same machine.
> "
>
> OUCH! Sounds to me like we're on the same side of the issue, Hixie!
> But I have to hand it to you again, I did in fact miss the "40% of the
> TOTAL time" part, you got me! So to sum up, this was the sarcastic
> comment I should have posted, and which I would have had this data been
> easily accessible:
Funny. CNet.com takes about 5 seconds to render in IE6. Netscape 6.1
PR1 (don't have a mozilla build installed currently here) renders the
page in 1.953 seconds, as per the timer in the browser. Go figure, I'd
have different results
>
> [whoever said:] "With the preloader for Windows, the difference [between
> IE and Mathuzilla] is nothing."
> [my sarcastic comment should have been:] "Except 10MB and 4 times slower
> rendering."
>
> Are we friends again? You're not going to sue me, are ya Hixie?
>
>
>>Naturaly, it does vary based on the configuration.
>>
>>
>
> Where's the charts or better yet graphs comparing the two, Mr. Hickson?
> I can't seem to find them anywhere. I can find a few scattered examples
> of Maozilla *only* tests, but what good do they do anybody when they
> lack the critical context that *you* *yourself* point out as necessary?
Actually, if you'll look at the startup tests they include NS6, NC4.7
and Mozilla.
>
>
>>>>I refer people who want to know the accurate numbers to the
>>>>n.p.mozilla.performance newsgroup.
>>>>
>>>Well that's the rub, ain't it Hixie? There really are no official,
>>>reproducable numbers in .performance or anywhere else as to
>>>Maozilla's performance in relation to either its predecessor or IE,
>>>is there?
>>>
>>Well, actually there are. Daily and weekly numbers.
>>
>>
>
> Daily numbers? Sorry Mr. Hickson, .performance often goes *days*
> without ANY posts whatsoever! You must be reading off Netscape's "For
> AOL Eyes Only" feed.
I'm not affiliated with Netscape in any way, and I see at least 5 - 10
posts per day in .performance. Do you need some assistance with your
newsreader?
>
>
>>>Where's the charts and graphs showing Mathuzilla's progress on
>>>rendering time, memory hoggage, etc, compared to these other
>>>browsers?
>>>
>>Well, oddly enough, charts and graphs are regularly posted to the
>>newsgroup I mentioned. Look for posts by jrgm, PaulW, and Sairuh.
>>
>>
>
> Regularly? No. The only thing I can find that's even an attempt at
> that is the page by Curt Patrick, and again, those lack any context by
> which to judge them by.
>
>
>>>Let me guess: behind a firewall.
>>>
>>Only one of the tests (to my knowledge) is behind a firewall, and that
>>is purely for technical reasons (the content comes from real web sites
>>and so copyright issues prevent Netscape from making the tests
>>publically available).
>>
>
> "Reasons of National Security", I see.
>
>
>>The results are published regularly. I once
>>again refer the interested reader to the performance newsgroup, where
>>one will find charts and graphs for many things such as page
>>rendering, page loading over a network, startup time, preferences
>>display, bookmarks and so on.
>>
>>Should the Netscape (AOL) numbers be doubted by anyone, I strongly
>>encourage these people to create and run their own tests and publish
>>the results and graphs regularly. Past experience has shown that
>>easily obtainable hard numbers are an amazing motivator.
>>
>>
>
> Hard to doubt numbers you can't find.
I can find and post article numbers to each of these tests, if you'd
like me to?
