“If software developers were doing their jobs properly, an extension ecosystem wouldn’t be necessary”
I have to disagree with this statement, It is because the FF team has done such a great job that there is an extension ecosystem.
You should have a look at:
a href="http://wiki.mozilla.org/Spectator"http://wiki.mozilla.org/Spectator/a
“Spectator is a Firefox extension that collects data about how Firefox is used. This information is used to better understand what people value in the browser.”
The problem with extensions is that if get a large enough sample size, almost everyone has three different “must haves”.
Down that path lies featuritis and madness…
(Although I would like to see Greasemonkey as default, and AdBlock as default, and…)
and I’m so disappointed to find out what Ruby XHTML support really is.
Firebug. A web developer’s wet dream.
The only extensions I have in FireFox are:
AdBlock,
NoScript,
Download Them All,
IE Tab
I don’t need / want any others. Don’t bake in features I don’t need. I understand your philosophy that most users would never bother with extensions (perhaps they don’t know they exist), but don’t burden power users in a futile attempt to serve ignorant users.
I have to disagree on one point. You say:
blockquotebIf software developers were doing their jobs properly, an extension ecosystem wouldn’t be necessary/b/div
But doing a development job properly does emnot/em mean that extensions are unnecessary. Especially since you point out that doing their job “properly” means folding in uthe most popular extensions/u. That leaves the other 99% of extensions out of the mix, ones that certains users will find invaluable.
A better statement would be that if developers were doing their job, then the most popular extensions (which equates to being the most desired features) are in the code while the users can get the individual or smaller set customizations they need elsewhere.
To say extensions wouldn’t be necessary is to make the mistake of assuming one set of features fits everybody’s needs, or that one development team can make everybody happy. We all know that’s impossible. 
I have to agree with the halfway solution others have suggested: having some extensions by default and being able to uninstall them.
But I’d like to take the idea further. Why only browsers? Word processors, for example, are now ultra-bloated with features most of us seldom or never need (and I don’t mean the optional features of MSWord, for example: each one of them would be the equivalent to 8 or 10 extensions).
Right now, AbiWord is the one that almost gets the idea, but its extensibility is mostly directed for file conversion, filters, and programming needs, when I mean things like adding/removing live word count, auto-text entries, mail merge capabilities, table auto-formatting, equation editing (perhaps with several levels of complexity, from basic to advanced math), bibliography management (a la EndNote or Zotero), corkboard (a la Scrivener) etc.
That, and a proper page preview (by default, please!) and I’d ditch MSWord or OpenOffice for a lighter word processor in the flick of a switch.
The funniest thing about this discussion is that it’s pretty clear who has tried the Firefox 3 betas. Also, asking for an extension that makes HTML rendering faster is like asking for luggage that makes your flight go faster.
I’m pretty sure firefox has incorporated some extension into it, most notably tabbing extensions and session-restoring ones, and that was in FF2.0! While you may complain that they don’t incorporate them fast enough, you can’t really claim that they aren’t being responsive.
FEBE lets you back up portions of your profile, including extensions. CLEO lets you create an XPI that packages together several of your installed extensions; the package can then be installed into another profile, onto another computer, etc.
The commentors who are describing distributing groups of extensions might want to have a look at these. However, I don’t know what the less straight forward behaviors are, e.g. when the destination already has an extension installed (perhaps at a different version level).
FEBE
a href="https://addons.mozilla.org/en-US/firefox/addon/2109"https://addons.mozilla.org/en-US/firefox/addon/2109/a
CLEO
a href="https://addons.mozilla.org/en-US/firefox/addon/2942"https://addons.mozilla.org/en-US/firefox/addon/2942/a
Recently I’ve been happily using Opera.
It has a lot more functionality by default than Firefox does, but it seems that many of the Firefox add-ons would be really good to have.
So now what… ?
- Switch to Firefox because of all the add-ons?
But Firefox is so… clunky compared to Opera, that maybe I’d rather not!
Then again… and so on.
I can’t help but think of the parallels between the bake-in/leave-as-addon balance and the similar one in programming languages- “Language Feature” or “Library”. For instance, in Icon you can get the length of list foo with “*foo”. Other languages, Length is a method/function in a library somewhere. How do you decide when something is necessary enough to bake in?
I tend to lean towards the more conservative approach to baking in- I see firefox as a “starting point” for my browser, upon which I pick and choose functionality based on how directly their features affect me and how heavyweight I want my browser to be. After all, wasn’t firefox developed to be a fast, lightweight version of the Mozilla browser?
The thing is, while baking in would obviously allow speed optimizations to be made for certain features (where those optimizations might not necessarily be possible via the plugin interface), the number of them that would have to be installed replace popular plugins would eventually bog the browser as a whole down, and there’s no guarantee that most of us would even use them all.
A better solution might be to simply have the firefox installer automatically include the 10 top favorite add-ons, but LEAVE them as add-ons, so that they can be removed for performance purposes, or upgraded on a more regular schedule (an add-on author can pump out updates a lot faster than firefox will push new release versions, after all).
As a sidenote, that “cmo” thing needs to be an extension, or at least an option with a checkbox next to it. Otherwise when .cmo or .nte domains come out, there’ll be hell to pay.
You bring up some good points that some of the most popular extensions should be baked right into the browser, but looking though my list of extensions (most which I couldn’t live without) are specific to me and what I’m doing with the web.
The only extensions that I have that I see as being beneficial to the entire community as a whole is the Download Status Bar.
I personally love Mouse Gestures as well and feel everyone should install them (they are in the default install of Opera by the way), but I’m not sure people would love having that forced on them. The behavior might confuse some people causing frustration at the UI.
Hm, I think, you are right in a certain way, but I disagree with your conclusion.
However, a warning, take my post with a grain of salt. I am a unix-user and I love to be able to make a program do exactly what I want.
I agree that it is annoying to be told ‘install this extension’, ‘install that extension’ and so on - and especially, usual users just wont do it. However, I disagree that you should fatten the core engine of the browser (or, the software) if you already have a good extension system. In fact, Id rather remove functionality from the core system and move it into extensions to get a slimmer core engine.
Why? Dont I disagree with you now?
Well, no.
I think you have to differ between the core engine and the (let me phrase it this way) dumb user distribution (less offensive: the non-power-users).
Imo, the core engine should be damn slim, because there exist powerusers that want to use the core engine the fastest possible way, with just the extensions they need (who needs a thing that tries to correct address if you just dont mistype?).
However, as you said, you should not force the less experienced users to do too much about installing extensions and such. Thus, you should track the extensions most users have installed (or, enough users have installed) in some way an just put all of these extensions in a single distribution for the lazy user.
That way, you are having a browser (or, a software, as this is pretty general) that can be used by powerusers in a very efficient way, however, lazy users will get a browser that is comfortable to use. The features of that browser will even expand pretty nice.
And this is pretty much what you want.
Lazy users? They just download the package with the popular extensions preinstalled. They wont care that much about a bit of bloat and maybe even some slowness, so thats ok.
The bit-more-experienced users? Well, they might grab the package with the popular extensions preinstalled, uninstall some, install some others and be happy.
And those freaky powerusers that need some engine to retrieve websites from a shell or want a slim speed demon? well, they grab the bare-bones version and install whatever they need and are happy as well.
OF course, you have to look out that a lazy users wont grab the bare-bones version by accident 
At least, thats my opinion, tainted by MICAHI (Make it complex and hde it)
Greetings
i totally disagree. 80% of the people use 20% of the browser. But other 20% than the others. So we have people which want adblock, but we do have people who don’t care.
The best solution would be a minimal browser (xul/webrunner based) with a lot of extensions, and preconfigures profiles with the most important extensions (history, bookmarks, tabs)
Dude,
I think you should use Opera once in a while. Considering the geek that you are, I’m surprised you’re not using Opera.
I can’t breathe in front of a computer which doesn’t have Opera installed! In the last gazillion years every machine I’ve sat on has to have Opera installed.
Give it a serious shot. It’s a bit like vi. Once you get the hang of it, you wonder how you managed earlier.
- abhijit.
Two things:
One application where extensions (mods) is constantly integrated in the core app would be World of Warcraft. Several mods have over the years been deemed so good that they became integrated as blizzard mods.
Second, wtf is up with http://feeds.feedburner.com/codinghorror ? FILE_NOT_FOUND
@Tom:
“Just because you’re nervous about downloading unknown quantities from the web or don’t feel confident about making these decisions, doesn’t mean that you won’t appreciate a URL spellcheck.”
-
For clarity’s sake: my comments are not about downloading potentially harmful mystery packets, they’re about making choices beyond ‘default’. Fear of the Big Bad Web and how to ease it for our grandparents is another discussion altogether.
-
My point is exactly that if one doesn’t know about the availability of choice and/or isn’t able to deal with making the choices involved, one is much more than likely not going to appreciate the improvements that would come from it.
“The kind of people who don’t use the web much are exactly the kind of people who misspell URL’s. I follow the “My Grandad” principle on this thing.”
Your example of URL spellcheck is a good one in this case: it breaks a very important principle of messing with expectations. I like to think that software that corrects my typing mistakes makes me a worse typist, because I never get a trigger to correct myself. For our grandparents it causes them to have an even bigger divide between what they think they entered and what the obvious result is. Hence, they’ll take even longer to scale the learning curve.
Wanna help out for this use case? Present a page with a message saying “Couldn’t find that page. Possibly you have misspelled that. Did you mean X, Y or Z?” (with X, Y and Z clickable links, off course)."
(oh, and be nice and try to find out if you have connectivity before assuming user error)
i think the really crazy thing is, they say: install this or that extension to get this or that features. But when your browser crashes, they will say: our core is great, blame the extensions!
Wtf?!