Guest post by “Gran Paradiso Alpha 4 is now available for testing. New features in this development milestone of Mozilla Firefox 3 include the FUEL JavaScript library for extension developers, a redesigned Page Info window, improvements to offline application support and Gecko 1.9 bug fixes. Several Mac OS X additions have also been made, including support for Growl notifications, improvements to the Cocoa user interface and an initial version of the Breakpad crash reporting tool.”
at what point does FF become the cross-platform rich-media uber-client of justice, such that people port their legacy apps using whatever GUI kit to FF and just have done with it?
When FF, or any other Web platform, boasts superior development toolkits and graphics presentation protocols than native development toolkits (think kdelibs-4) and protocols (think NX).
This is the crux of the “Web vs. Desktop” dilemma. I see the Web as a band-aid for Microsoft’s and other platform vendors’ failure to bring applications to the desktop over the network in a compatible and compelling manner. Will browser platforms like FF ever leverage client-side graphics hardware? Will there be an easy and flexible way to share data between Web applications?
My view is that we should be much closer to delivering on the promise of network computing via something like NX or RDP than we are to getting there by way of the Web. But the browser made the Web an easier application delivery mechanism than remote desktop clients did. The desktop *nix world shunned the network transparency of X11 while it should have embraced it. We could have powerful desktop experiences on meager client hardware, but instead we have memory-hungry browsers on fat clients.
Imagine if your browser could be a container for X11 applications as well as Web applications. Better yet, imagine if this was tightly integrated into your desktop shell. “Google Apps” for whatever you want and you’re running the latest application everyone’s been talking about. No installation, no management hassles, less resource consumption. Sound good?
Edited 2007-04-29 03:32
Will browser platforms like FF ever leverage client-side graphics hardware?
Actually, Firefox 3 already does using Cairo.
Will there be an easy and flexible way to share data between Web applications?
I’m sure there will be. There’s already permanent storage in a database.
Imagine if your browser could be a container for X11 applications as well as Web applications. Better yet, imagine if this was tightly integrated into your desktop shell. “Google Apps” for whatever you want and you’re running the latest application everyone’s been talking about. No installation, no management hassles, less resource consumption. Sound good?
Yes. But not without problems. You’ve basically moved much of the hardware requirements from the client to the server, and not many companies would have the resources to host apps on the internet – most people would be limited to serving a couple hundred people at a time.
I don’t expect web apps to ever replace native desktop apps, but they do have their place.
You’ve basically moved much of the hardware requirements from the client to the server, and not many companies would have the resources to host apps on the internet
What do you think the Web is doing?
Consider the current model of software delivery and it’s various problems–piracy, compatibility, security, service, support, etc. The distance between the vendor and the user is a problem, and the business model of licensing bits is untenable in many situations. As the software industry moves to more of a service-oriented business model, application hosting makes much more sense. You’re paying for the vendor to host the application, provide the computing resources, manage the software updates, etc. Hosting the application isn’t a burden, it’s the reason the vendor has a revenue stream. Think of the implications for OSS economics…
The key is choice. The same app can be hosted remotely for a subscription fee (or provided by your employer/institution) or run locally either gratis or for a licensing fee. Some intensive applications might not translate well over the Internet, but even a game might run well on a LAN over a good compressed protocol. I’m not here to tell you that all applications should run remotely. I’m arguing that it should be a choice for most applications.
The Web started as a remote platform and is now dipping its toes into local (offline) applications. The desktop has predominantly been a platform for native applications to run locally, but it has potential for remote hosting that hasn’t been fully exploited.
What do you think the Web is doing?
There’s a big difference between streaming a text file and running an entire application for each person. At my job, we’ve got a server running that people can VNC into, but it is only to be used when you are sick or traveling because it gets extremely bogged down whenever more than 2 or 3 instances of AutoCAD are simultaneously running.
But yes, if you charge a fee for using the service it would be tenable. I just don’t think the traditional ad based revenue would be enough to pay for all the hardware you would need. You would also be giving up on dial-up users, although there are fewer of them each year.
I guess I’m just not entirely sold on the idea of software being a service yet.
Edited 2007-04-29 05:41
There’s a big difference between streaming a text file and running an entire application for each person.
I think that streaming a text file is a gross simplification of what the Web is today and especially of what it is becoming.
Here’s another argument: Consider the trends in hardware and OS development. How does increasing thread-level parallelism, virtualization, and clustering solve typical fat client problems? It doesn’t. But these advancements definitely further the viability of application servers. The market for computing is shifting to the datacenter while graphics processing is increasingly becoming the core element of a client computer. The movement of profit margins in the consumer space is a clear indication of these market forces.
XUL and friends are the hero of the moment, not necessarily Firefox. XUL is being used; Songbird is one example, but I think it is a solution trying to find a problem.
On the desktop there are a laundry list of APIs which can provide the same sort of multiplatformness as XUL, on the web side of things, there is Ajax and other technologies which can provide that ‘web experience’.
With that being said, I would love them to finally fix the memory leaks and bloat that exists within Firefox – that is the one reason why I use Internet Explorer; when you’ve got 5 tabs open, and its chewing through hundreds of megabytes worth of memory, it painful to say the least.
Sure, in my case, I have a Core 2 plus a gig of memory, but at the same time, the programmers should be aiming to reduce its footprint – Firefox developers seem to forget that with a web browser, like an operating system, its what runs onto that is important.
Like in the operating system, that the applications are important, in the case of web browser, the most important key is the web site. The web browser shouldn’t dominate, it is merely the delivery platform for content; the web browser doesn’t exist for web browser sake, but to allow the end user to access information via their internet connection.
With that being said, I would love them to finally fix the memory leaks and bloat that exists within Firefox – that is the one reason why I use Internet Explorer; when you’ve got 5 tabs open, and its chewing through hundreds of megabytes worth of memory, it painful to say the least.
The leaks may continue to exist because they are hard to track down. I’ve never had any memory problems with Firefox and as a test, I just opened three Firefox windows each with a dozen tabs (all pointed at different sites) and my total memory usage sat at around 50 MB. I don’t doubt that you have issues, but the problem is certainly not universal.
Try using it for a long time; for 30 minutes usage, it isn’t noticeable, but use it for a couple of hours, and you’ll find memory usage will rise.
There are tools out there to detected memory leaks, what I see the problem is that Mozilla developers are more concerned about adding features than trying to get the underlying technological foundation secure, stable and reliable.
Take Firefox and the lack of native widgets; this has been an issue known since the release of Mozilla 1.0 and each release, they’ve failed to address it; underlying problems with the Mozilla code base, again, known for quite some time, and failure to address them.
Try using it for a long time; for 30 minutes usage, it isn’t noticeable, but use it for a couple of hours, and you’ll find memory usage will rise.
This computer runs for days at a time with multiple copies of Firefox open. I still have never noticed any problems. I wonder if the issue may have to do with extensions or Java. I don’t use any extensions and I refuse to install Java on any of my machines.
For me, I don’t have extensions installed; as for Java, I don’t think the problem is the technology, but its over use of it and using it in the wrong place where a lighter technology would have done the same job.
The same thing occurs with Flash and Shockwave, there seems to be this infectious thing within web developers brains that seems to think that as soon as a new technology comes out, they must use it for every damn thing on their website, without justification or weighing up the consequences of that choice.
I use Firefox for days without restarting, and the memory usage does rise higher until I eventually have to restart it. The thing is, I’ve actually found that IE7 is even worse. I haven’t used it continuously to see if it leaks memory as much, but when I open 10-15 tabs in IE7 and the same in Firefox, IE7 is using considerably more memory.
There are tools out there to detected memory leaks, what I see the problem is that Mozilla developers are more concerned about adding features than trying to get the underlying technological foundation secure, stable and reliable.
I can’t say I completely agree or disagree with this – I don’t think they focus enough on those things, but I don’t think it’s fair to say they don’t care about them either. I think there have been a lot of changes in this area that will be showing up in Firefox 3, ironically because they didn’t want to make huge changes in 2 for fear of making things unstable.
Aye? What version are you running? right now 4 tabs, and it uses 72MB is total, which is a far sight less than what I experienced with Firefox.
Firefox isn’t just a memory bloater on Windows, use Konqueror on *NIX and compare it to Firefox, for instance; compare Epiphany to Firefox too.
But what I want aren’t big changes, but fixing up long standing bugs which piss end users off; these can be made right through the 2.x cycle with no impact on the over all stability of the application.
Yes, crawling through code and fixing up stupid coding mistakes can be as painful as being kicked in the goolies, but its better to fix them up now before they cause bigger problems later on.
I thought the browser wars were over?
I love the current FF and what the Mozilla team envision for it, but, aren’t they being celebrated for something Microsoft is quite often criticized for?
“I thought the browser wars were over?”
According to Microsoft they are. The fact remains that the users choice for any default application will always result in competing applications. A war the first time around, sure, but now it is more about choice and standards. So while I wouldn’t call it a war, you can bet that Microsoft will do whatever it takes to win.
“I love the current FF and what the Mozilla team envision for it”
Eh, while FF is my favorite browser and current choice, the fact that they have chosen to use their great plugin system for some of the expanded features, while forcefully bundling others makes me question where it is going. The phishing filter I had to disable and delete the sql list because it is too big to quickly sync to USB and back. Further, projects like Damn Small Linux and other low-resource projects will likely not be upgrading to newer Firefox’s because of the force-bundle. I think they should extension everything that is not explicitly core-browser. Including the awesome spell checking feature which I do enjoy.
Perhaps someone can fork the Firefox Suite and make a fast and feature lite browser that can be expanded by plugins? One can hope.
“but, aren’t they being celebrated for something Microsoft is quite often criticized for?”
They are leveraging their superior market in a given area to discourage choice and innovation in another? They are creating new communication systems that are non-standard that encourages development for only their software? They are breaking laws in several countries and simply buying their way out?
Besides being successful, where are they like Microsoft? And even in that regard, Mozilla’s success is by user choice, a far cry from Microsoft’s reasons for success.
I’m hoping you have an interesting point to discuss that I missed, but you didn’t explain what you meant in your post, so I’m guessing I’m not the only one curious as to what you meant.
Edited 2007-04-29 00:55
The browser war ended when Microsoft included IE in their operating systems. Just like the IM war ended when they included MSN in their operating systems.
Life goes on and software continues to be developed.
I thought the browser wars finished a week or two ago?
“I love the current FF and what the Mozilla team envision for it, but, aren’t they being celebrated for something Microsoft is quite often criticized for?”
They are, but I think the following points are significant:
A) The source code for Firefox is public, and there are no patents on their ideas, so any other browser vendor can implement the same thing if they choose; and
B) Firefox is truly cross platform, so implementing new features that web sites may depend on doesn’t prevent anyone from using them. All they have to do is install FF, not Windows.
I think it’s fantastic that Firefox has re-energised the browser space, and are leading the way towards a faster, richer web experience.
You mean people who use other browsers, like konqueror, opera, etc., can effectively go away? So how is this different from what IE’s been doing with activex et al?
Huh?
… Are you flame baiting or am I just failing to understand your core argument?
If you don’t like the extension system? Take the code, yank the parts that you do not like, and create your own, Mozilla based browser.
Can you do the same (read: take the code and do what-ever you want with it… Including full re-branding – E.g. IceWeasel)
What on earth are you talking about?
– Gilboa
“but, aren’t they being celebrated for something Microsoft is quite often criticized for?”
No. Microsoft’s extensions were HTML and other content-oriented changes that created lock-in to their browser.
Firefox’s extensions are for developing plugins for the browser. They enhance the functionality for the user independent of the content in question. They are not extensions geared towards the content format.
The release notes mention “Many Cocoa user interface improvements”, anyone noticed what they are? I was expecting native widgets, but not only are they non-native, they look even worse than those in the current Firefox version
Also, I really, *really* hope they consider making Pinstripe the default Mac OS X theme for Firefox:
http://kmgerich.com/2006/09/27/pinstripe-for-firefox-now-with-20-mo…
Unified titlebar/toolbar would be nice too.
I know, those are “shallow” issues, but they can really help in making Firefox feel “native” for OS X users.
(Note: I know about Camino, and I have it installed, but one of the main features of Firefox is support for extensions, which is missing in Camino)
Pinstripe was the previous default that they exchanged for the current default, so I kinda doubt they’ll switch back. Unfortunately.
Edited 2007-04-29 02:54
Yeah I agree with the unified look which is absent from Firefox.
I use Camino for all my browsing for the past 2 years and love it.
I use Firefox on my Linux wkstn and also on Windows at work and find Camino is ok.
To use Camino I recommend you install CamiTools, which comes close to some of the plug-ins that are missing.
http://www.nadamac.de/camitools/
Is a build tagged with “Gran Paradiso” only in an alpha/beta release? The nightlies still say “Minefield.”
Gran Paradiso is the code name for Firefox 3, while Minefield is the code name for the current trunk. I think those are still the same code at the moment, but eventually a separate branch will be made for Gran Paradiso.
Mozilla are working close to the trunk this time, instead of branching for the 3.x line (as was done with both FF1 and 1.5 on the 1.7 and 1.8 trunk lines respectively). This is done to reduce patch porting to many brances, and also keep FF3 close to the very latest developments on the trunk. Whether this decision proves to be a QA nightmare for feature locking the betas is yet to be seen.
kudos to the devs and community.
Can the FF development team please start concentrating on the quality of their product, instead of adding a zillion new features each release?
Firefox is very much lacking in performance, memory use and stability.
I would easily give up half the features if this would mean I would receive a truely polished product instead.
I just installed it on Fedora (linux), but what did they do with the fonts? The fonts look completely different then on FF 2.
I mean WTF still no native GTK+ widgets on web forms and pages, what are Mozilla playing at?
Open source browser, open source toolkit, hello Mozilla?
… Feel free to implement it yourself
Open Source, you eat your own dog fod.
– Gilboa
I’d been using earlier Alpha versions and they’d used native widgets on Mac OS X, except for the first version, I believe.
I was surprised to see that they’d been removed but I also noted that bugs were fixed. I think that it’s not merely coincidence. Whatever happens, XUL should be dead or at least, minimised, and most of the memory leaks should vanish.
I noticed that they’ve mentioned their intention to revive Places, which appeared in the original version 2 Alpha release, so I hope that bookmarking will finally be usable, reliable, and flexible.