Author Topic: DLS: Mozilla publishes name and shame list of slow Firefox add-ons (Read 32,227 times)

lanux128 · « **on:** April 04, 2011, 12:01 PM »

via Download Squad, "Mozilla, continuing its year-long crusade to speed up Firefox startup and shutdown times, has published a name and shame list of the Firefox's slowest add-ons." how many of the addons do you use?

[attach=#]
• http://downloadsquad.switched.com/2011/04/04/mozilla-publishes-name-and-shame-list-of-slow-firefox-add-ons/

PS. incidentally i use 8 from the 50 addons that were listed.

mouser · « **Reply #1 on:** April 04, 2011, 12:02 PM »

excellent find

f0dder · « **Reply #2 on:** April 04, 2011, 01:20 PM »

Might seem a bit harsh to provide such a list, but I think it's a really good idea - helps showing that, while FireFox isn't superduper nimble, it's extensions that are responsible for the worst performance degradations. And might even work as an incentive for those extension writers to fix up their code

epopuI · « **Reply #3 on:** April 04, 2011, 03:11 PM »

Made for interesting reading - thank you

Winkie · « **Reply #4 on:** April 04, 2011, 07:06 PM »

Nice find Lanux!

Quote from: lanux128 on April 04, 2011, 12:01 PM

how many of the addons do you use?

With my still-default Firefox 3.6: 53
With my soon-to-be-default Firefox 4.0: 47

So a count of add-ons on the list used with both leads to... 9

, no surprise they are on the list for most of these.

Surprises on the list, IMHO:
#17 Personas Plus, a Mozilla Labs add-on...
#18 Flagfox
#27 Web Developer
#28 Colorzilla
#42 Stylish, is relatively low. But I'll guess that also depends on the amount of styles that has to loaded and whether they are XUL-styles or not...

app103 · « **Reply #5 on:** April 04, 2011, 07:29 PM »

I have found it best to use different profiles for different purposes. You don't need Firebug or other web development related add-ons for normal surfing, so putting them in their own profile can go far to speeding up your profile used for normal surfing.

I currently have 7 that are on the list installed into my regular profile. I would have had more but I got rid of them awhile back to speed things up. Will I remove any of those 7? No, not likely.

housetier · « **Reply #6 on:** April 04, 2011, 07:30 PM »

I take the same approach as app103. Sometimes I install but disable extensions for later use.

wraith808 · « **Reply #7 on:** April 04, 2011, 09:15 PM »

Quote from: housetier on April 04, 2011, 07:30 PM

I take the same approach as app103. Sometimes I install but disable extensions for later use.

I just wish that FF approached extensions in the same way as chrome in regards to restarting the browser. I would use disable more for extensions that I don't use that much, but I hate restarting the browser to do so.

Tuxman · « **Reply #8 on:** April 04, 2011, 09:30 PM »

Hah, so glad that my extension isn't listed there.

However, it takes no wonder that deeply inspecting add-ons like Firebug cause slowness, right=

Curt · « **Reply #9 on:** April 04, 2011, 09:56 PM »

excellent find!
-but isn't the list from ff3?

Mine:

#07 (33%) Video DownloadHelper,
#18 (16%) Flagfox,
#26 (14%) Download Statusbar, and
#42 (05%) Stylish.

wraith808 · « **Reply #10 on:** April 05, 2011, 03:59 PM »

Mine:
#04 (50%) Flashgot (*)
#09 (30%) Xmarks
#12 (21%) Adblock Plus
#17 (17%) Personas Rotator (*)
#19 (16%) All-in-One Sidebar
#26 (14%) Download Statusbar
#43 (04%) Read It later

This actually informed me of a cool extension (integrated gmail) though I probably won't install it.

As far as whether I'll get rid of any of them, looking at cost versus usefulness, the ones with the (*) beside them are in danger of being removed, I think. The other ones, I'm willing to take the hit for.

Gothi[c] · « **Reply #11 on:** April 06, 2011, 12:03 AM »

Might be worth noting that they only benchmarked ff startup/shutdown time, not the actual impact on page rendering!

f0dder · « **Reply #12 on:** April 06, 2011, 12:06 AM »

Quote from: Gothi[c

link=topic=26306.msg244427#msg244427 date=1302048197]Might be worth noting that they only benchmarked ff startup/shutdown time, not the actual impact on page rendering!

Most plugins really shouldn't affect page rendering, and start/shutdown speed is what firefox is notoriously slow at? Makes perfect sense to me

Deozaan · « **Reply #13 on:** April 06, 2011, 12:08 AM »

Ironic that FastestFox is on that list.

lanux128 · « **Reply #14 on:** April 06, 2011, 03:06 AM »

Quote from: Deozaan on April 06, 2011, 12:08 AM

Ironic that FastestFox is on that list.

FastestFox tweaks an internal setting called 'network pipelining' which has been discussed here before. since these settings are at best experimental in nature the probability of breaking something is higher.

also to be noted is that FastestFox has a serious bug where it fails to restore the original (default) settings when it is removed, hence the users is stuck with a buggy profile.

mahesh2k · « **Reply #15 on:** April 06, 2011, 03:42 AM »

#2 Firebug
#5 FoxClocks
#9 Xmarks Sync
#12 Adblock Plus
#21 DownThemAll!
#50 Firefox Sync

These are essential add-ons for me. I don't know what can i do without them.

steeladept · « **Reply #16 on:** April 06, 2011, 04:27 AM »

I only have 2, and no, they won't be coming off. AdBlock Plus and NoScript are too essential. If they are really issue causing, perhaps Mozilla will write the browser to address the pains these two address. Happily, my must have, TabMixPlus, and the 4 other niceties are NOT on the list.

I can say, however, I have used many of them, and I do agree - many of them slow it A LOT! Read it Later seemed particularly onerous given what it does. That one got removed shortly after it was tried.

cyberdiva · « **Reply #17 on:** April 07, 2011, 10:45 PM »

I was surprised to find that I had only three from the list: AdBlock, ColorZilla, and Better Privacy (though Better Privacy had only a 1% impact). I disabled ColorZilla, since I realized that I almost never use it. I was relieved to find that the add-on I value most--Lazarus--is not on the list, though I'd keep it even if it were #1.

rgdot · « **Reply #18 on:** April 07, 2011, 11:16 PM »

In the top 10 I only use Firebug.

FF with 10+ extensions is near unusable, wish I knew how anybody has 50 or so. I am not talking about slow old machines either. Having said that Flash is the biggest culprit, one easy way that it's obvious to me is using Flashblock

AndyM · « **Reply #19 on:** April 08, 2011, 12:00 AM »

Quote from: rgdot on April 07, 2011, 11:16 PM

FF with 10+ extensions is near unusable, wish I knew how anybody has 50 or so

40 (according to one of my extensions, Mr Tech Toolkit

), and things seem to work pretty well.

Tuxman · « **Reply #20 on:** April 08, 2011, 08:33 AM »

I had 80 on fx3.. Waiting for updates..

Renegade · « **Reply #21 on:** April 08, 2011, 01:57 PM »

It makes me wish that they'd have an interface for other languages, and not just JavaScript.

It would be SOOOOO nice to be able to write plugins in C#. At the moment I have to write custom browsers when I need to do things. I would much rather write plugins that I could distribute.

lanux128 · « **Reply #22 on:** April 08, 2011, 03:14 PM »

Adblock Plus developer, Wladimir Palant gets back at the tests by Mozilla with some tests of his own.

• http://downloadsquad.switched.com/2011/04/08/adblock-plus-developer-pokes-holes-in-mozillas-new-add-on-perfo/

rxantos · « **Reply #23 on:** April 08, 2011, 11:40 PM »

It would be interesting doing the same thing but for windows applications. If the application is slow and bloated, users should know before buying.

JavaJones · « **Reply #24 on:** April 08, 2011, 11:55 PM »

Quote from: rxantos on April 08, 2011, 11:40 PM

It would be interesting doing the same thing but for windows applications. If the application is slow and bloated, users should know before buying.

I'd like to see that sort of thing as a regular part of software reviews. An additional metric. It would be particularly useful/interesting in comparison/"shootout" type reviews comparing multiple similar products. SoftMaker Office starts faster than MS Office 2010 and has more features and is cheaper? Then why bother with MS Office? etc.

But, as Wladimir's tests in the link above show, if you do tests it's important to be comprehensive and to be crystal clear about your test environment and process.

- Oshyan