one of the slowest add-ons? any chance of improvement?

[sabret00the]

According to this, this is one of the slowest, but also one of my most used add-ons. Is there any way we can improve the impact it has on the browser?

[Giorgio Maone]

I'm trying to contact fligtar for a confirmation (he's not showing in IRC since 2 days ago), but I suspect the measurement is not entirely fair.
AFAIK it's an average over 20 runs on 3 platforms, but I believe they include the very first run after installation.

Now, FlashGot currently performs quite expensive post-install tasks (it extracts native executables for each platform and use them for external download manager autodetection) which may impact startup time not just by 50%, but even up to 3 times.

On the other hand, in my tests FlashGot 1.2.8.5 has never added more than 10% to the startup time after first run.

Nevertheless, latest development build should be well under 5% even under those measurement constraints.

[sabret00the]

Thanks for being so on top of this. I thought I was on the development channel, but I wasn't. Have added this latest dev build.

[therube]

While "times" can be important (like http://forums.informaction.com/viewtopi ... =10&t=4774 & so long as you're not running into issues like those), I would want to be more concerned with the time to actually load web pages & how extensions may affect that, rather then the time to load the browser from start.

[sabret00the]

While "times" can be important (like http://forums.informaction.com/viewtopi ... =10&t=4774 & so long as you're not running into issues like those), I would want to be more concerned with the time to actually load web pages & how extensions may affect that, rather then the time to load the browser from start.

I'm not sure what you're adding here therube. You appear to be posting for postings sake and patronisingly at that. If the extension in itself is adding 50% to start up time it's a problem. The fact that it's now down to 1/10th of that proves that and I congratulate Giorgio on performing a fantastic feat in doing so. Thanks to the work performed from the current dev version from the previous, my startup time impact is down from 96% to 51%. I can only at this point state that I'm happy that he doesn't share your attitude and be hope that more add-on authors are as responsible and proactive as Giorgio.

[therube]

I'm saying in most cases start up time will be relatively immaterial (except perhaps for bragging rights).
So if FlashGot (enter-your-extension-name-here) causes a slowdown - only at start up, I'm not really that concerned.
However, if there were a continued drag on browser performance by keeping FlashGot (enter-your-extension-name-here) installed, I would be greatly concerned.

Faster is better, sure. but if you only look at start up, you are not looking at the whole picture, only one aspect.

[GµårÐïåñ]

@sabret00the,

As for what therube is saying, whatever you take it as being, patronizing or posting for posting sake, is a legitimate statement. Regardless of whether it applies to what you are saying or not, it is a legitimate addition that while startup time and cost are certainly a concern, especially if they CONTINUE to drag the browser AFTER startup functions are completed, in itself and on the face of it, are not a resource concern unless you are on a specially slow machine, or specially limited resource machine (swap file, memory, hd speed, etc) and it affects prolonged or ACTUAL use of the browser. I took nothing in what was said to be either patronizing or foobar, neither should you have.

Every developer strives to minimize the resource consumption, execution time and efficiency, that's a given - unless you are a completely clueless or poor developer in which case you probably wouldn't be putting out anything useful to begin with. Someone of Giorgio's experience and skill will ALWAYS, regardless of external review or concern, strive for the fastest, most resource efficient program for the sake of all aforementioned reasons. Do memory leaks and resource consumption issues arise, of course, but not always due to something the developer is doing but most often because of an underlying bug or deficiency by the platform. Sometimes extra resource for the sake of stability is far more valuable than sacrificing function and quality for the sake of making it smaller, faster, this and that; so consider that and that's where what therube said becomes a legitimately valid statement.

I have known therube for a LONG time and despite any differences in opinion or approach we might have had over time (or differences in our experience and skill set), I have always found to be fair and supporting for the sake of support, not just posting for the sake of posting. So please keep in mind that although you are entitled to your opinion and perspective, do not be abusive or rude as it is the job of the mods and admins to do the best they can to provide the most useful and expansive perspective on a subject than a user might see or realize. There are things, as should be - as part of continually supporting a project, that we might see or remember or find possibly useful than the user might consider. At the very least, it does no harm and you can read it, use it, or ignore it but the response was neither fair, nor appreciative and uncalled for to someone who took the time to give you the respect of a response to help you and give you perspective.

@therube,
As stated above, you are well within your understanding of what's going on to make that statement and its completely valid for more reasons than we will need to get into here, but simply put, your response was thoughtful and at the very least provided perspective that might have been otherwise lacking. Keep up the good work.

[therube]

Some more details on Mozilla's add-on performance measurements

[therube]

On fluctuations in performance testing results


PS: NoScript v 2.1.0.2rc5:

Code: Select all

v 2.1.0.2rc5
==========================================================================
x Halved startup time

[Giorgio Maone]

@therube:
You forgot http://adblockplus.org/blog/how-reliabl ... asurements -- which is the most interesting for FlashGot, BTW.

[sabret00the]

Since this thread seems to have become a one stop shop for the conversation regarding addon startup speed, I'll add a little bit.

Is startup speed important? Yes. It's not paramount and I don't think no user is running around claiming such a thing. However the case with Firefox is that it's got a reputation of a slow starter compared to Chrome (as it's biggest threat).

At that point it falls onto the various Mozilla teams to take responsibility. You have the engineers that are working hard to reduce startup time. Some of those efforts saw the session save dialog removed. However there's only so much that can be done. Between end-users expecting Firefox to load 50 tabs in the same speed is takes Chrome to load one. And people expecting Firefox with 100 extensions to load up in the same speed it takes Chrome to load up with none. There's been issues that exist outside of the realm of what can be coded for from within the team.

Based on the fact that some Plugins and extensions do in fact have such a negative affect, it was right that the authors were given the responsibility to aspire to being the best. As a long term Firefox user, there's a quality that I expect all extensions to aspire for and that standard is the bar that's been set by ABP, NoScript and Stylish to name a few. Anything else irks me in that there are extensions that I can't live without but it bothers me the authors don't take such a level of pride in their work and update them (little things like authors updating to mark Firefox 4 compatible but still relying on the statusbar shim).

FlashGot has in fact taken the slur that's been put against it's name and no matter how rightly or wrongly it was put there, the pride and the professionalism in the extension has lead it to be improve and by doing that, it will hopefully lead more authors to improve. Even if it's not concious; An addon author downloads FlashGot simply as a reference and ends up doing things in a more efficient way.

Personally I only use FlashGot as a go-between for Firefox and JDownloader and to be honest, I wish there was a XULRunner alternative to JDownloader, but for now I stick with the most open download manager available. And I probably use less than 1/10th of the features available in FlashGot (I use download selection and download link). So it's unfair that my greatest concerns are how quickly Firefox loads when I need it and how quickly it can send the relevant links to JD. But ultimately I believe that the author has earned my trust and thus I believe he'll do the best job he can.

There are other concerns in terms of speed and being that FlashGot is so popular, I'm sure it'll appear on the shame in an unfavourable light again. But ultimately, this place is at least intended to be a friendly enough place that a user can come and highlight a concern and be welcomed with open arms while being told that plans to improve in whichever regard, are already in place or in fact as was my experience, being worked on.

[therube]

Jorge · 2011-04-08 23:02

We will be posting some rectifications in the Add-ons blog very soon and will be updating the page to clarify a few things for people who are arriving to it directly (like most of the media).

We don’t think that the list of top slow add-ons is incorrect. Those add-ons are slow, and the only exceptional case I know of is FlashGot and the skewed Mac OS results. This should be fixed soon. The list is incomplete due to testing failures that should be fixed soon.

http://adblockplus.org/blog/on-fluctuat ... ts#c003966