Extension support

Extension support

Developers should add extension support to Opera (just like firefox, ie). I think it will be great for Opera, if it is happened. So much users are using firefox becouse of extension support.

Thanks!

this would be a huge boost i think. i have to admit the reason i switched to firefox years ago was cuz of extension support. however, i recently switched back to opera cuz it uses the Q Toolkit, only to find out they don't have a Qt4/x86_64 build!

Opera should have SDK and extensions, but it should'nt become Firefox.

it will be amazing specially with a program like opera which already has a plenty of features (almost 80% of features anybody can imagine; i guess). so when someone add extentions will not add more than 5 extentions (when i was using firefox i had around 30+ extentions).
but i don't think opera will move in that.

Of course, extensions would be a great feature. Chrome has understood that and is currently developing an extension framework. The first extensions have seen light already. I don't know why the Opera team doesn't want extensions. Sometimes, change is good. I would remove the widget framework, I would extract the IRC client, turn it into an extension, and provide an extension manager. People would love it and the extension web site would populate in no time (compared to the abandoned widget web page)

Originally posted by calande:

People would love it and the extension web site would populate in no time (compared to the abandoned widget web page)

Abandoned?? Are you completely mad? Sim Aquarium has had more than 16 THOUSAND downloads in the last 7 days!

Why are you spewing nonsense without checking the facts? I don't get it. People always do that.

Originally posted by GoJoeGo:

Why are you spewing nonsense without checking the facts? I don't get it

What a hypocritical thing to say.

+∞ for an extension api

+1
Lots of websites (or volunteers) develop extensions to add functionalities in Firefox, sometimes in IE (example: ebay, google). It's impossible with Opera because there is no extension system.

+1
It would make opera way more customizable than what it already is

Originally posted by calande:

Sometimes, change is good. I would remove the widget framework, I would extract the IRC client, turn it into an extension, and provide an extension manager. People would love it and the extension web site would populate in no time (compared to the abandoned widget web page)

+1 so far..

Originally posted by calande:

I would extract the IRC client, turn it into an extension, and provide an extension manager

I would do the same with every component that is not related to browsing (the rss, mail, torrent, irc clients for instance).

-1 There are other things to do.

+ 1/2
m2c: it would be sufficient if one could get javascript api to extend the ui through userJS. In future versions the widget-file-api (or some other functionality to store settings) could be ported to userjs and there you have it: a pretty neat and lightweight extension system (another nice effect: userJS would get some attention and not be pushed into a niche ^^)

So why +1/2? because imo it is not necessary to implement a full "extension api" if you can work on a basis of what you allready have.

This doesn't address UI changing elements but would allow you to define new apis. I believe this is how google gears are made... although I'm mostly speculating.

Opera should support extension but they must limit the amount of extension that users can have because most of the time when extension slows down the speed of browsing most people would blame the browser instead.

-1

Not important to me at all, since I would never use them......since I'm sure it would be another Firefox fiasco, where every time there is a new snapshot, the extensions would all stop working, and one extension would probably create compatibility issues with others. I would also be afraid Opera would stop adding new innovations into the core product, allowing extension authors to do it instead.

-1 .

-1

Originally posted by serious:

+ 1/2
m2c: it would be sufficient if one could get javascript api to extend the ui through userJS. In future versions the widget-file-api (or some other functionality to store settings) could be ported to userjs and there you have it: a pretty neat and lightweight extension system (another nice effect: userJS would get some attention and not be pushed into a niche ^^)

So why +1/2? because imo it is not necessary to implement a full "extension api" if you can work on a basis of what you allready have.

+1,000,000,000

Originally posted by poorrichard01:

-1

Not important to me at all, since I would never use them......since I'm sure it would be another Firefox fiasco, where every time there is a new snapshot, the extensions would all stop working, and one extension would probably create compatibility issues with others. I would also be afraid Opera would stop adding new innovations into the core product, allowing extension authors to do it instead.

Well I've had no problems with Firefox for many years. I use 5 extensions and they have alwys worked fine.

I've just started to use Opera and wouldn't have done so if it wasn't free. I do prefer the OpenSource background of Firefox and still use Thunderbird fo email.

Opera's widgets seem like mickey mouse add ons and I don't like them. However Opera10 seems to do all I want. I can't customise it quite as much but there's enough basics to tweek to get a good custom gui.

Originally posted by serious:

+ 1/2
m2c: it would be sufficient if one could get javascript api to extend the ui through userJS. In future versions the widget-file-api (or some other functionality to store settings) could be ported to userjs and there you have it: a pretty neat and lightweight extension system (another nice effect: userJS would get some attention and not be pushed into a niche ^^)

Yes, one of the major downsides of UserJS at the moment is that it cannot be really integrated into the UI, so, among other problems, every time you want to change a setting of the script you have to open the .js file in a text editor. This is not difficult of course, but it can be very off-putting for less experienced users.

Originally posted by GoJoeGo:

Abandoned?? Are you completely mad? Sim Aquarium has had more than 16 THOUSAND downloads in the last 7 days!

And over two milions downloads in total... That, is insane. I have to wonder who would want an aquarium widget.

- 1

Opera is a compact, all-in-one browser which no so slow (program startup, program running, interactions...etc) as Firefox. Opera shouldn't support extension.

I would also be afraid Opera would stop adding new innovations into the core product, allowing extension authors to do it instead.

I agree with that.

-1

Why not take a poll of yes no or maybe?

Because extensions are developed independantly. Each update to Opera has the chance to break an extenstion, which would then require an update by the dev. This means that:
1- People will be unhappy when their extensions stop working properly, and they will blame Opera.
2- Developers will eventually stop producing timely updates, this causes more unhappiness.
3- Extensions are a security risk. This problem has been raised with Firefox extensions many times already. If an extension is malicous or is hacked, Opera will recieve the blame.
4- Most of Firefox's extensions have an equivelant in Opera already.

Originally posted by serious:

+ 1/2
m2c: it would be sufficient if one could get javascript api to extend the ui through userJS. In future versions the widget-file-api (or some other functionality to store settings) could be ported to userjs and there you have it: a pretty neat and lightweight extension system (another nice effect: userJS would get some attention and not be pushed into a niche ^^)

So why +1/2? because imo it is not necessary to implement a full "extension api" if you can work on a basis of what you allready have.

I'd rather not have javascript touch the UI.
Those scripts are good to make a clock or get some light processing but not for a responsive environment.
I'd rather have opera take care of the UI.
In firefox there is an option to make XPCOM components programmed in C++ that are supposed to be faster but
it's no better than an integrated solution ( activex/coms = fat & slow).
Like the post above suggest it will also bring bugs and problems, I'd rather have opera develop the bulk of components
well like they do already than good fan ideas getting poorly executed and slowing the browser.
For non "invasive addons" there are already widgets (which could gain some integration in the browser).

Originally posted by ColdWhiskey:

I'd rather have opera develop the bulk of components
well like they do already than good fan ideas getting poorly executed and slowing the browser.

Nothing would force you to use the fan made extension though. Nor does the existence of an extension prevent Opera from including an improved version in the browser. Take the trash can, in FF it started out as an extension (inspired by Opera IIRC) now it's an integral part of the browser.
The great thing about extensions is that a fan might come up with an idea that the devs missed. With extensions Opera could react very quickly to new features from other browsers. FF users saw and like the speeddial in Opera. Rather than having to wait for the devs to decide whether they liked it, build it and test it, the FF community was able to have an extension with similar features ready quite rapidly.

As an example, it would be great if we had an extension for fine tuning search engines while we wait for the Opera developers to include this sorely wanted feature. (i know about Searchedit but it's just not the same)

There seems to be an assumption by some in this thread that supporting extensions will make the developers stop developing features for inclusion into the browser. I do not think that that is a correct assumption.

Originally posted by Zotlan:

There seems to be an assumption by some in this thread that supporting extensions will make the developers stop developing features for inclusion into the browser. I do not think that that is a correct assumption.

Discussions about an extension api is generally shrouded in FUD. Get used to it.

Nothing forces me , but it feels somewhat wise this way.
For an analogy , sometimes there are movies getting out in the theater in the US:
the fans in the US go to the cinema watch it immediately , in europe some clever will want to download a badly filmed version immediately,
some other wiser will wait for it to appear one month later in europe theater or even if they have not been watching the news be pleasantly surprised when discovering the film one year later on TV.
There are the forums for peoples coming up with new ideas and there is always the possibilty to use two browsers for those really impatient.
Though I understand what you mean not certain it's the better way of handling the matter,
while just looking at the results nowadays firefox is one of the slowest browser and doesn't look well updated beyond what is has to offer.
Google chrome aiming simplicity is fighting opera the fastest browser podium...

Originally posted by ColdWhiskey:

For an analogy , sometimes there are movies getting out in the theater in the US:
the fans in the US go to the cinema watch it immediately , in europe some clever will want to download a badly filmed version immediately,
some other wiser will wait for it to appear one month later in europe theater or even if they have not been watching the news be pleasantly surprised when discovering the film one year later on TV.

The difference is that the movie fans know that the movie will be out in a month and, if they really like it they might even watch the downloaded version and the cinematic version once it hits the cinema.
We, on the other hand, have very little idea of what features Opera will, or will not include in the future. Take inline spell check, there was a huge request thread which ran for years, never was there any word from Opera. Until, suddenly, the feature was there.
Besides which with extensions fans can make features that are for such a small niche audience that they would never be interesting for Opera to include. They can come up with features that Opera has not. The increased potential for tinkering will also attract a lot of tech-savvy users to Opera (though no doubt some will be turned off by the fact that it is closed source).

Originally posted by ColdWhiskey:

There are the forums for peoples coming up with new ideas and there is always the possibilty to use two browsers for those really impatient.

Depends on the feature. If FF has an extension for editing its search engines that does not help me on Opera does it? Besides, i prefer to stick to a single program, so do most users i think.

Originally posted by ColdWhiskey:

while just looking at the results nowadays firefox is one of the slowest browser and doesn't look well updated beyond what is has to offer.

But is it slow because of extensions, because of the Gecko engine or for another reason altogether? Just becausethere are problems with the way FF has done things does not mean the principle is flawed.

Originally posted by Zotlan:

But is it slow because of extensions, because of the Gecko engine or for another reason altogether? Just becausethere are problems with the way FF has done things does not mean the principle is flawed.

It's slow because of the XUL overhead. You're right about the principle being sound, but the implementation being flawed. I would love to see a "system of automated extensibility" built in to Opera, but I wonder whether this would be what users in this thread really want, or do they want pretty much an exact copy of Firefox's system.

serious's m2c above hits the nail right on the head imo. But his idea might not be in line with exactly what others want - fearphage for example blogged about it here but rubbishes the suggestion of a UserJS UI access api (in the second comment on that blog) saying it would not be sufficient.

So it's clear there is division in the ranks of Opera users on this issue, whether we all want automated extensibility or not.

My vote is "yes" to automated extensibility, ideally in the form of an extended UserJS api, "no" to anything resembling XUL.

Originally posted by lucideer:

automated extensibility

automated extensibilty?? What's automated about it? I don't want to install firefox extensions but I do want the power level of customization that extensions provide. Just like Opera's userjs implementation dwarfs greasemonkey's functionality, i would like opera's extension api to dwarf mozilla's (and now chrome's) api.

For me at least, one of major downsides using UserJs generally is the fact that JavaScript has to be activated globally in the browser (possible workaround I didn't tested yet).
Besides the power of UserJs is limited. For what it is indeed useful (content-, layout-, header-control/manipulation) I'm already using a third party application with a JavaScript engine of its own. So far at least for me UserJs has very little importance if any.

Originally posted by Zotlan:

But is it slow because of extensions, because of the Gecko engine or for another reason altogether?

The slowness of Firefox is mainly caused by XUL the user-interface language it uses as lucideer pointed out.

Originally posted by lucideer:

My vote is "yes" to automated extensibility, ideally in the form of an extended UserJS api, "no" to anything resembling XUL.

You are perfectly right about XUL but I don't understand why XUL should be considered as the only viable option.

Originally posted by fearphage:

automated extensibilty?? What's automated about it?

All I mean by "automated" is an installer - unlike the current method of "installing" userJS.

Originally posted by Krake:

I'm already using a third party application with a JavaScript engine of its own. So far at least for me UserJs has very little importance if any.

I'm not sure what exactly you mean by this. Could you elaborate? I'm curious.

Originally posted by Krake:

The slowness of Firefox is mainly caused by XUL the user-interface language it uses as lucideer pointed out.

The slowness of a bare Firefox install is mainly caused by XUL. The slowness of a heavily customised Firefox install is all too often caused by rogue extensions - there are many well written extensions that don't slow the browser down at all, but there are just as many that do.

There's many an argument claiming you can get all of Opera's functionality in Firefox by adding extensions - I wonder how much of Opera's functionality you can get by limiting yourself to efficient well written extensions that don't slow the browser to a crawl.

Originally posted by lucideer:

I'm not sure what exactly you mean by this. Could you elaborate? I'm curious.

Proxomitron, a local filtering proxy

Originally posted by Krake:

The slowness of Firefox is mainly caused by XUL the user-interface language it uses as lucideer pointed out.

Originally posted by lucideer:

The slowness of a bare Firefox install is mainly caused by XUL. The slowness of a heavily customised Firefox install is all too often caused by rogue extensions - there are many well written extensions that don't slow the browser down at all, but there are just as many that do.

XUL is slow as an interpreted form of language , but javascript is also interpreted (it's not platform compiled).
Or did I miss something?

Originally posted by Zotlan:

There seems to be an assumption by some in this thread that supporting extensions will make the developers stop developing features for inclusion into the browser. I do not think that that is a correct assumption.

Then how do you explain the fact that Firefox is the ONLY one of the main alternative browsers not including as part of the core product something that at least LOOKS like Speed Dial? It seems to me the only logical reason is that the developers figured, "Hey, it's already available as an extension, so why bother?"

Originally posted by poorrichard01:

Then how do you explain the fact that Firefox is the ONLY one of the main alternative browsers not including as part of the core product something that at least LOOKS like Speed Dial? It seems to me the only logical reason is that the developers figured, "Hey, it's already available as an extension, so why bother?"

Well, obviously i cannot look into the heads of the firefox developers so i cannot know for sure. Perhaps it is because of the extension, it could well be that the developers thought there was nothing to improve on it (i've never used the thing so i cannot judge it). Perhaps they simply felt that there was not enough demand for it to warrant inclusion in the browser. The firefox developers seem bent on including as few features as they can get away with (i.e. only things that 90% of the users would actually use) and leaving the rest to the extensions. Even for something as basic as editing the shortcuts you need an extension.
Now, once again, just because it works that way for Firefox does not mean it will work that way for Opera as well. Opera contains myriad features that only a limited amount of people use.

Originally posted by ColdWhiskey:

Originally posted by Krake:

The slowness of Firefox is mainly caused by XUL the user-interface language it uses as lucideer pointed out.

I'm afraid my statement was ambiguous.
The slowness (and resource consumption) of Firefox attributed to XUL, relates to its start up time.
XUL doesn't/shouldn't affect the rendering speed.
Nevertheless:
Since out of the box FF is very poorly featured users are almost damned to install extensions. Even supposing that all extensions are well done (wonder if this is the case) there is the major risk of negative interferences between extensions. Such negative interferences might very well slow down the rendering speed and also rise resource consumption.

Theoretically it would be possible to slow down even Opera only by using dozens of UserJs just in case some of them would bite one another.

UserJS can slow things down, especially if you have lots and they conflict with each other.
They can also cause rendering errors, which is why it's advisable to disable your UserJS when you have an issue (such as a lower score on Acid3). I can't really think of a case where UserJS has actually affected rendering time though, although it's possible in theory.

Originally posted by SteelRealm:

UserJS can slow things down, especially if you have lots and they conflict with each other.
They can also cause rendering errors, which is why it's advisable to disable your UserJS when you have an issue (such as a lower score on Acid3). I can't really think of a case where UserJS has actually affected rendering time though, although it's possible in theory.

It's sure javascripts cans slow things down , it's even now the main mesure of browser performances with SunSpider.
Though it discussable as it is clearly intended to advantages Chrome ( performance is normally mainly regular html/css rendering).

Originally posted by Krake:

I'm afraid my statement was ambiguous.
The slowness (and resource consumption) of Firefox attributed to XUL, relates to its start up time.
XUL doesn't/shouldn't affect the rendering speed.

Memory usage is another aspect of XUL. High memory usage in Opera can normally be mostly attributed to aggressive caching, maybe a little towards heavy M2 usage, with the actual process generally using very little - a clean fresh uncustomised Firefox however uses a relatively large amount in comparison.

Originally posted by ColdWhiskey:

It's sure javascripts cans slow things down , it's even now the main mesure of browser performances with SunSpider.
Though it discussable as it is clearly intended to advantages Chrome ( performance is normally mainly regular html/css rendering).

Performance issues are not generally as attributable to in-page javascript as they are to UserJS as the two perform very different functions.

UserJS is usually modifying the DOM structure of a page, something which can be inherently resource intensive. However, if you're the author of the page you can just modify the html directly, there's no need to do so with javascript, so in-page scripts tend to perform more benign, less resource intensive functions. The exceptions would be with javascript animation or <canvas>, but in the grand scale of things these make up a tiny portion of what the browser encounters on the web, which is why SunSpider et al are fairly irrelevant when it comes to performance.

Originally posted by poorrichard01:

Then how do you explain the fact that Firefox is the ONLY one of the main alternative browsers not including as part of the core product something that at least LOOKS like Speed Dial?

Because they don't want to perhaps? Did you ever consider that? Chrome has extensions too btw.

Originally posted by fearphage:

Chrome has extensions too btw.

Chrome is supposed to have a very efficient javascript engine , so extensions won't be such a ressource hug as on firefox.
But I suppose that once plugins for chrome will be avalaible , the one having chrome with all equivalent plugins to opera
(M2 mail content blocking etc), will certainly notice it to be slower than opera.
For the simple reason whatever task you do an interpreter is still slower than a hard compiled code.
Chrome shines for now in its light version: it is the fastest for its functionnalities.
For now , opera is the fastest for its functionnalities.
Anyway if opera team had the wish to include extensions I suppose they would be better at pushing up the performance of theier javascript engine before doing this.

Originally posted by ColdWhiskey:

Chrome is supposed to have a very efficient javascript engine , so extensions won't be such a ressource hug as on firefox.

Originally posted by ColdWhiskey:

Anyway if opera team had the wish to include extensions I suppose they would be better at pushing up the performance of theier javascript engine before doing this.

To my knowledge extensions, as they are now in FF at least, have nothing to do with javascript. How efficient the javascript engine is has nothing to do with the performance of the extensions.

Originally posted by Zotlan:

Originally posted by ColdWhiskey:

Chrome is supposed to have a very efficient javascript engine , so extensions won't be such a ressource hug as on firefox.

Originally posted by ColdWhiskey:

Anyway if opera team had the wish to include extensions I suppose they would be better at pushing up the performance of theier javascript engine before doing this.

To my knowledge extensions, as they are now in FF at least, have nothing to do with javascript. How efficient the javascript engine is has nothing to do with the performance of the extensions.

Well most are just javascript apps just like widgets, they just use XUL rather than HTML/XML for formating the UI.
Now it's no wonder why so many peoples trash widgets :
they don't know it's just about the same background.

Originally posted by Zotlan:

To my knowledge extensions, as they are now in FF at least, have nothing to do with javascript. How efficient the javascript engine is has nothing to do with the performance of the extensions.

Eh.. no. The vast majority of Firefox extensions operate almost exclusively on javascript. You can call compiled code with XPCOM, but most don't.

Originally posted by ColdWhiskey:

For the simple reason whatever task you do an interpreter is still slower than a hard compiled code.

The new breed of javascript engines (V8/Carakan/etc.) are using JIT compiler's on some or all of the code which is part of the reason some of them are seeing quite large performance improvements.

Originally posted by lucideer:

Eh.. no. The vast majority of Firefox extensions operate almost exclusively on javascript. You can call compiled code with XPCOM, but most don't.

I stand corrected then

Originally posted by lucideer:

The new breed of javascript engines (V8/Carakan/etc.) are using JIT compiler's on some or all of the code which is part of the reason some of them are seeing quite large performance improvements.

It would be good geting a just in time compiler for opera.EDIT: seems to be forecast in carakan.

Originally posted by Zotlan:

I stand corrected then

Extensions look well integrated to the browser that's why it cans bring confusion on what's behind.
That's a bit why I would have liked to have the options for widgets to insert better in the browser context.
I suppose it would be an easy way of fullfilling a good part of the needs of those wanting extensions .

A lot of firefox's functionality began as extensions. Here is another on its way into the core.

http://www.techradar.com/news/software/applications/revealed-the-world-s-best-browser-591144

Originally posted by amoun0:

Opera's widgets seem like mickey mouse add ons and I don't like them.

The only "Mickey Mouse" here is your extremely poor understanding of widgets. They are not supposed to be like extensions at all. They are "standalone" applications developed using web technology.

Originally posted by GoJoeGo:

They are "standalone" applications developed using web technology.

More like child windows of opera. They are as stand alone as the irc or mail clients in opera... basically not at all. Widgets are great in the right context. I don't believe the desktop is that context.