But the thing is, Mozilla is not monolithic. It is is a discrete set of components running in the same address space. Chrome overlays make them look like a single app but most of the code each part of the suite appears in its own distinct DLLs and.jar files under the surface. And if you don't want the 'clutter' of multiple tools, the answer is not to install them in the first place. IIRC a mailto: link in the browser will open your default email client if there is no mail/news installed for example.
The perception that running stuff seperately is going to be some magic panacea is wrong. At the end of the day you will have subsets of Mozilla running in their own process space instead of the whole lot in one. Aside from some potential stability improvements (not that Moz is bad now), the effort is more to facilitate a UI rewrite than to fix any fundamental bustage. There is also a downside that you might lose integration that some people appreciate such as a unified pref dialog, a single profile, being able to open a tab in the browser from a link in an email and so on.
The UI would definitely benefit from simplification it has to be said, but the suite has to come out the other side of this process as functional as it went in, and that also includes ensuring stuff like the editor and other less visible parts (e.g. JS debugger, DOM inspector) are not left behind in the process.
What the hell are you talking about? If you didn't want the mail/news component why did you install it?
Yes, that checkbox in the installer does indeed control whether you get the mailnews component. If you're using a.zip or.dmg that gave you no choice, then a little surgery such as removing the mail/news chrome does the same trick.
Cluster sizes are very important if you create lots of small files, for example if you develop source code, or even if you have hundreds of bookmarks. On any modern hard drive you're looking at 32k clusters.
So a moderate size of 100 bookmarks, eats up a whopping 3200k of space. By comparison, a single file containing all the bookmarks would be less than 100k. Develop something like Mozilla with 1000s of source files and the wastage would be grotesque as would the fragmentation.
Unless you're upgrading from Win98 and don't want to lose your files, it's miles better to move to NTFS, not least for all the other things it offers.
Personally I'd be a little concerned about WinFS until it proves itself to be reliable, which is not the case for the system database people already use in NT / XP - the registry. The thing can be a real bastard to fix when it fails.
I own a Palm and I love it but what the hell is the point of this? It's too small to be a proper PDA, too chunky to be a normal watch and you need to recharge it every four days! Excuse me, but I prefer my slim conventional watch which has a battery life measuring in years.
It's not like you can't get organiser watches from the likes of Casio for much if you want that kind of thing, and I bet they last considerably longer than this. Come to think of it, why not just buy a Palm Zire for $120 and buy a really nice watch with the change?
This is not to knock the demo which is cool, but I'm surprised that something similar hasn't been attempted with Linux. While the QNX microkernel is small (which helps) after throwing in the various disk, network, mouse, drivers it's probably no different than what you could achieve in Linux. We have already seen the likes of Toms RTBT which pack a single disk with an amazing amount of command line stuff.
I wonder if some of that could be jettisoned for some kind of microwindows based GUI and perhaps a browser of some kind. Having tried Knoppix and LNX BBC also, I think these things are inherently cool and more importantly useful.
It gives a warm fuzzy feeling to know that you can carry around a disk or CD which boots into a full blown emergency repair kit or demos what can be done with Linux. The best that has so far been mustered for Windows is the excrable recovery console - something while better than nothing is still incapable of solving all but the most basic problems.
What are you talking about? IE for Mac has not improved much since Mac OS 9 days. It might have been ported it to OS X (thanks to Carbon), but it's still the same slow ass, single threaded browser it always was.
Mozilla had the advantage that it has been designed from scratch for multi-threaded and particularly Unix systems in mind. The Mach-o release exploits this to present an Aqua front end and a multi-threaded sockets based backend and is subsequently very fast.
The core might be older than that, but it hasn't been on the Mac before Safari turned up. Therefore it is silly for MS to claim their lack of development on IE for two years or abandonment now had anything to do with it. It might have been the writing on the wall, but I suspect for a long time that MS hasn't been dominating the market quite as much as it likes.
Given Mozilla / Netscape's continual popularity on the OS X download site, it is a fair bet that it owns a major slice of OS X users. Whether that remains the case when Safari appears has yet to be seen, but from what I've seen of Safari it is aiming at the simple browser market, so power users will probably stick with Moz.
IE for the Mac OS X hasn't undergone any significant upgrade for two years. Ironically it was once lauded as the most standards compliant browser around. Any talk of Safari killing it makes no sense at all as it's only been out for a few months and in beta form at that. Microsoft could have improved their browser significantly in that time (e.g. making it multithreaded instead of a straight carbon port), or making it more compliant with the look and feel of the OS.
Anyway, Mozilla played as much part in its demise. I've used Moz since it's been available for OS X and aside from being slightly sluggish in early versions, it has always been a better, more stable, more compatible browser.
SMS! Sign up and you can message peoples phones for free. This is an incredibly useful feature.
I can't be bothered with the interface however. It's like someone said - let's design a non-standard user interface which buries our bastard users up to their eyeballs in configurable options! There is add a simple and expert mode, but some of the frequently used settings don't appear in the simple mode.
I'd prefer AIM anyday. My understanding is that under the surface they use the same IM technology anyway so whatever reason for keeping them seperate must have more to do with politics than technical difficulty.
The problem was the scenes were forced and too long. I'm sure a few second shot of smiling Arwen as Arargorn lay dazed holding his amulet, plus a sentence or two such as "I love her more than life itself" or whatever would have conveyed the same message.
Portraying an entire race of peoples as evil Islamic terrorists is true? Of course it isn't. It is absolutely untrue and a gross distortion and slander. Certainly there are those who right or wrongly attribute their squalor and oppression to the US & Israel but I suggest they are a tiny minority. The vast majority or people of any race or religion are polite, friendly, and hospitable who just want to get on with living. But are they portrayed that way? Of course not.
So I suggest before you complain of the speck in someone elses eye you remove the twig from your own.
I actually thought the cinematic release of TTT could have done with less scenes. All that flashback crap containing Arwen completely dragged down the middle segment. A five minute trim would have made it a better picture.
Now the extended edition does appear to introduce some interesting new stuff, but I wonder if yet more flashbacks concerning Boromir and Faramir is just compounding the error. Enough with the flashbacks!
Cool! I recall a game aeons ago called CRobots which was a similar thing where you wrote robots in C and watched them fight it out in an arena. I had always been tempted to do the same in Java but sadly laziness kicked in and I never bothered. I'm glad to see someone else did.
My understanding is that these services cannot bill you for receiving messages and I do not know any mechanism that they could (except if they're operated by your own mobile phone provider).
They work by charging you a premium for that initial text message. Premium text services are usually very obvious, being strange numbers with no dialling code and they'll cost you some rate such as 1.50 for the message. That premium buys you whatever you were expecting back - alerts, ring tones or whatever. They could also string you along by requiring you send multiple messages, back and forth thus increasing their revenue. By comparison, it would be cheap for them to text you in reply so there would be no need to bill you in the other direction.
Unsolicited phone spam makes me furious, clogging up my phone's puny memory and requiring manual deletion. The only time I think it is justifiable is for roaming (when you get a welcome message telling you various things about the network) or for system outage information. Otherwise it should definitely be opt-in.
I suspect people on pay as you go plans are likely to be abused by their telcos however. O2 spammed me a couple of times about competitions, so perhaps they got a clue on how angry it makes people. There is also the possibility in Europe that such spamming quickly runs afoul of various data protection and telephone junk marketing laws.
Hmm, not in the UK and probably the rest of Europe as far as I know. The sender pays. I don't know what happens with roaming so I guess you might get charged to receive messages in that situation.
It Moz calls fork() (and I haven't checked), it is either safely contained within some ifdefs, or behind an abstraction. For the most part, the browser is as portable as you might hope considering how many platforms it builds on.
GRE end product is a lightweight runtime, so you are correct in that regard, but it is not built that way. Perhaps the emergence of various standalone / distinct clients will cause the build system to be redone, but I believe at the moment, that even for Firebird you must still build Mozilla first before you can build Firebird.
I have no experience of webcore, but I would guess that it would suffer many of the same issues. It might look lighter, but then you still have to port zlib, png, jpeg, expat, QT and various other bits and pieces to make it work. So the work involved might amount to the same and of course you don't have a browser at the end of it, just a browser engine.
The GRE is a stripped down version of Mozilla, but the stripping down is done after Mozilla has been built. In other words, you must build Mozilla for the stuff that the GRE cherry picks to be there in the output dir. There is no "make gre" that only builds the gre stuff. It is a post-build step that you will see in mozilla/embedding/config.
So if you did want to build only the GRE bits, you would have to make a list of what things those are compile around all the other stuff (leaving a mess of ifdefs in your wake). As the other stuff is mostly some extra C++ components, plus of course all the chrome, and other settings that make the GRE useful for something, it kind of makes no sense to not build Mozilla. Besides Mozilla is the most convenient way to test if the port actually works or not.
Of course, you could break down the port into small steps (always a good idea). I would suggest that it would follow this general model.
Makefile / configure system. Detect Amiga from the script and add some #defines to config generated headers for conditional compilation. Fix bustage in perl scripts etc.
Port NSPR. Since NSPR is already portable this might be easy.
Port XPCOM, strings. Sort out the directory provider and component manager, inter-thread marshalling, local file object, memory manager.
Port widgets & gfx. Code for creating windows, drawing graphics etc.
From there upwards it should become plain sailing as stuff becomes increasingly abstract and generic. Mozilla does have some chrome overlays for various platforms but you could probably get away with using the Unix one for a while. It obviously also helps to disable stuff like psm, mail / news, mathml, xslt and other bits and pieces that can be enabled later once the core is working.
There is no such thing as the GRE at the build level anyway, since you must build the entire Mozilla first before running the script that cherry picks the subset of files that represent the GRE. If Mozilla doesn't build, you have no GRE.
So it is better to just forget about there being a GRE at all to begin with. Just port Mozilla and the GRE will fall out of it, but the reverse is not true.
I would guess the most straightforward way of porting to the Amiga would be via some X11 port on that platform and reuse most of the X11/GTK widgets. Treat the Amiga as a weird Unix variant, use gcc, gmake and as many GNU tools as are required to make things easy on yourself.
You still have the NSPR and Makefile system, and some assembly used by XPCOM to contend with, but the length task of writing widgets and gfx classes from scratch would go away.
I had a PC emulator for my Amiga 500. It clocked at a breathtaking 0.1 Mhz!
You could literally type something on the command prompt and wait a minute for the characters to agonizingly appear one at a time on the screen. Other than its excrutiating speed it was useful for transferring files and whatnot but it fast it was not.
MS.cab files use the same principles as tar.gz, compressing the concatenation of all files rather than individually. This results in better compression, but makes the format pretty lousy for random access.
Personally I'd like to see a zip specification which allowed bzip2 compression. You would still have random access to files but with better compression.
Unfortunately renaming is only a short term solution. It would be trivial for the next version of the driver to scan memory for certain bytes unique to the software or hashing some bytes from the executable file, or see what other DLLs it's loading or any manner of things and come to same conclusion.
You'd end up in an arms race, the likes of which anti-virus and anti-spyware find themselves in today. You put in one measure and the opposition counters with something different.
The only solution is for the driver writers to stop cheating in the first place. Perhaps there should be an industry sanctioned penalty for cheats, such as a product review blackout for a month, deducted points, 'cheating bastards' awards etc.
Slashdot Mirror
The perception that running stuff seperately is going to be some magic panacea is wrong. At the end of the day you will have subsets of Mozilla running in their own process space instead of the whole lot in one. Aside from some potential stability improvements (not that Moz is bad now), the effort is more to facilitate a UI rewrite than to fix any fundamental bustage. There is also a downside that you might lose integration that some people appreciate such as a unified pref dialog, a single profile, being able to open a tab in the browser from a link in an email and so on.
The UI would definitely benefit from simplification it has to be said, but the suite has to come out the other side of this process as functional as it went in, and that also includes ensuring stuff like the editor and other less visible parts (e.g. JS debugger, DOM inspector) are not left behind in the process.
Yes, that checkbox in the installer does indeed control whether you get the mailnews component. If you're using a
So a moderate size of 100 bookmarks, eats up a whopping 3200k of space. By comparison, a single file containing all the bookmarks would be less than 100k. Develop something like Mozilla with 1000s of source files and the wastage would be grotesque as would the fragmentation.
Unless you're upgrading from Win98 and don't want to lose your files, it's miles better to move to NTFS, not least for all the other things it offers.
Personally I'd be a little concerned about WinFS until it proves itself to be reliable, which is not the case for the system database people already use in NT / XP - the registry. The thing can be a real bastard to fix when it fails.
It's not like you can't get organiser watches from the likes of Casio for much if you want that kind of thing, and I bet they last considerably longer than this. Come to think of it, why not just buy a Palm Zire for $120 and buy a really nice watch with the change?
I wonder if some of that could be jettisoned for some kind of microwindows based GUI and perhaps a browser of some kind. Having tried Knoppix and LNX BBC also, I think these things are inherently cool and more importantly useful.
It gives a warm fuzzy feeling to know that you can carry around a disk or CD which boots into a full blown emergency repair kit or demos what can be done with Linux. The best that has so far been mustered for Windows is the excrable recovery console - something while better than nothing is still incapable of solving all but the most basic problems.
Mozilla had the advantage that it has been designed from scratch for multi-threaded and particularly Unix systems in mind. The Mach-o release exploits this to present an Aqua front end and a multi-threaded sockets based backend and is subsequently very fast.
Given Mozilla / Netscape's continual popularity on the OS X download site, it is a fair bet that it owns a major slice of OS X users. Whether that remains the case when Safari appears has yet to be seen, but from what I've seen of Safari it is aiming at the simple browser market, so power users will probably stick with Moz.
Anyway, Mozilla played as much part in its demise. I've used Moz since it's been available for OS X and aside from being slightly sluggish in early versions, it has always been a better, more stable, more compatible browser.
I can't be bothered with the interface however. It's like someone said - let's design a non-standard user interface which buries our bastard users up to their eyeballs in configurable options! There is add a simple and expert mode, but some of the frequently used settings don't appear in the simple mode.
I'd prefer AIM anyday. My understanding is that under the surface they use the same IM technology anyway so whatever reason for keeping them seperate must have more to do with politics than technical difficulty.
The problem was the scenes were forced and too long. I'm sure a few second shot of smiling Arwen as Arargorn lay dazed holding his amulet, plus a sentence or two such as "I love her more than life itself" or whatever would have conveyed the same message.
So I suggest before you complain of the speck in someone elses eye you remove the twig from your own.
Now the extended edition does appear to introduce some interesting new stuff, but I wonder if yet more flashbacks concerning Boromir and Faramir is just compounding the error. Enough with the flashbacks!
How is that any worse that the usual Hollywood trash of arabs always being portrayed as evil terrorists out to get the US?
Cool! I recall a game aeons ago called CRobots which was a similar thing where you wrote robots in C and watched them fight it out in an arena. I had always been tempted to do the same in Java but sadly laziness kicked in and I never bothered. I'm glad to see someone else did.
They work by charging you a premium for that initial text message. Premium text services are usually very obvious, being strange numbers with no dialling code and they'll cost you some rate such as 1.50 for the message. That premium buys you whatever you were expecting back - alerts, ring tones or whatever. They could also string you along by requiring you send multiple messages, back and forth thus increasing their revenue. By comparison, it would be cheap for them to text you in reply so there would be no need to bill you in the other direction.
I suspect people on pay as you go plans are likely to be abused by their telcos however. O2 spammed me a couple of times about competitions, so perhaps they got a clue on how angry it makes people. There is also the possibility in Europe that such spamming quickly runs afoul of various data protection and telephone junk marketing laws.
Hmm, not in the UK and probably the rest of Europe as far as I know. The sender pays. I don't know what happens with roaming so I guess you might get charged to receive messages in that situation.
It Moz calls fork() (and I haven't checked), it is either safely contained within some ifdefs, or behind an abstraction. For the most part, the browser is as portable as you might hope considering how many platforms it builds on.
I have no experience of webcore, but I would guess that it would suffer many of the same issues. It might look lighter, but then you still have to port zlib, png, jpeg, expat, QT and various other bits and pieces to make it work. So the work involved might amount to the same and of course you don't have a browser at the end of it, just a browser engine.
So if you did want to build only the GRE bits, you would have to make a list of what things those are compile around all the other stuff (leaving a mess of ifdefs in your wake). As the other stuff is mostly some extra C++ components, plus of course all the chrome, and other settings that make the GRE useful for something, it kind of makes no sense to not build Mozilla. Besides Mozilla is the most convenient way to test if the port actually works or not.
Of course, you could break down the port into small steps (always a good idea). I would suggest that it would follow this general model.
From there upwards it should become plain sailing as stuff becomes increasingly abstract and generic. Mozilla does have some chrome overlays for various platforms but you could probably get away with using the Unix one for a while.
It obviously also helps to disable stuff like psm, mail / news, mathml, xslt and other bits and pieces that can be enabled later once the core is working.
So it is better to just forget about there being a GRE at all to begin with. Just port Mozilla and the GRE will fall out of it, but the reverse is not true.
You still have the NSPR and Makefile system, and some assembly used by XPCOM to contend with, but the length task of writing widgets and gfx classes from scratch would go away.
You could literally type something on the command prompt and wait a minute for the characters to agonizingly appear one at a time on the screen. Other than its excrutiating speed it was useful for transferring files and whatnot but it fast it was not.
Personally I'd like to see a zip specification which allowed bzip2 compression. You would still have random access to files but with better compression.
You'd end up in an arms race, the likes of which anti-virus and anti-spyware find themselves in today. You put in one measure and the opposition counters with something different.
The only solution is for the driver writers to stop cheating in the first place. Perhaps there should be an industry sanctioned penalty for cheats, such as a product review blackout for a month, deducted points, 'cheating bastards' awards etc.