Out of curiosity, what is the point? Firewire presumably offers comparable performance doesn't it, or is there some compelling reason not to use it such as lower bandwidth or contention issues? And firewire seems to be a standard feature on an increasing number of PCs these days.
Now, I would welcome any replacement to conventional IDE / ATA which has been the bane of my life. I couldn't count the number of times I've had to screw around swapping cards and drives in order to accomodate that ribbon. I will be happy to see that particular technology go the way of the dodo.
The thing is, PDAs, phones, MP3 players and various other handheld thingies should never ever crash. My fridge doesn't crash, my camera doesn't crash and I don't expect these things to either. They are consumer gadgets not full blown computers with lots of drivers and bits that fail.
My Palm Pilot Vx has had an uptime that must measure in years now. And that's with day to day use. The only time I had to reboot it was when I updated it to PalmOS 4.
I would be most upset if I 'upgraded' to a Windows CE or Linux PDA and discovered the thing died for no apparant reason whenever it felt like it. Even once a month would be unacceptable. If this is a common thing in CE or the Zaurus (I don't know), then it is absolutely shocking that people put up with it.
The last time around it was the abysmal shiny screen, and now it's no headphone jack.
Space is a pretty weak reason for not including one at all. Lot's of MP3 / MiniDisc players are faced with the same size constraints yet manage to design a proper solution. A smaller jack and a couple of buttons could control the volume (no need for a wheel) or do what many minidisc players do and stick the volume controls onto the cord of the headphones itself.
As it stands, it sounds like Nintendo have an utterly clueless testing department to miss great big howlers like this, or more likely it is a cynical ploy to release sucessive Gameboy versions with major design flaws to drive sales for add-ons and the next one Gameboy which will supposedly fix the problem.
Well I don't think I'll be buying one. The GBA was the last straw for me - the screen is just atrocious and I'm not going to make the same mistake again.
At least you can see the Mozilla bug list. You can see how many are actual bugs, how many are meta bugs, how many are for minor quirks, how many cover future work or esoteric things, how many are duplicates, how many are platform specific, how many are mail / news / browser / embedding / editor / site specific, how many are for other projects sharing the bug database such as the webtools, who is working on what, what patches are available and so on. You can even raise new bugs and fix them if you like since the source and the bug system are open to all.
Contrast that with IE where none of these things are possible. Got a bug? Good luck trying to raise and track the issue with Microsoft.
Why does anyone care what jwz thinks about Mozilla? He hasn't been involved with the project in any way at all during the last four years. None whatsoever. Except griping from the sidelines of course.
I didn't mean that way. I mean if up2date says there is a new version e.g. 1.2.1 and you have 1.2.0, and there is an incremental diff available (i.e. 1.2.0-1.2.1) then it should fetch and apply that rather than fetching the whole 1.2.1 package which could be massive. After patching you now have 1.2.1 as if you had done rpm -Uvh on it.
There is no 'hotfixing' or piece patching here. The result of the incremental diff is the same as installing the whole new version, just considerably easier to download. As I mentioned, a kernel update is 35mb of patches. I would be surprised if an equivalent incremental patch were more than a megabyte.
It would be even more invaluable if Red Hat et al made RPM and their updates incremental. It is rather silly to expect people on 56k modems (of which there are still many) to download 30-50mb of patches to fix what probably amounts to 1mb at most of code changes. Kernel changes are particularly horrible - a one line patch means 35Mb download! How many users will bother with that? Now perhaps that's their own fault when they're rooted, but it's bad for everyone else too - a rooted box is a springboard for further attacks, not to mention dragging the reputation of Linux through the mud.
Why isn't it possible to produce incremental binary patches containing just the diffs? Not only would it vastly increase the chances of people downloading them (which is good for everyone), it is good for Red Hat too since their bandwidth for up2date is slashed.
Now obviously there are times when incremental diffs are not useful, but with the proper safeguards (e.g. checksums and backing up originals etc.) I don't see what the problem is. If anyone from Red Hat or RPM land is listening, please consider implementing this feature. Pretty please with a cherry on top.
MozillaQuest Magazine: So why are we are such a bad source for news?
Linus Torvalds: Because you're an utterly clueless one man operation that has nothing pleasant, informed or useful to say about anything. Having demonstrated woeful ignorance of your namesake Mozilla and after being shunned by that community you're now expanding your ignorance about other topics too. In fact I am surprised I am even talking to you - perhaps someone should have warned me.
But Mandrake is French. Let the assholes boycott that if they want to make an ineffectual protest at another nation that dares to have its own foreign policy and interests.
It certainly is superior, however from my own experience of moving up, you have to bite the bullet and rewrite your firewall script. While you could leave everything using ipchains, (2.4 has an ipchains module after all), distributions such as Red Hat don't like it very much and some of the network tools won't work when ipchains is running. IP Masquerading (NAT) and VPN also just works which is a compelling reason in itself. If you want VPN in 2.2 you have to arse around downloading kernel patches to enable it.
It's better to rewrite everything to use iptables, though this does require some effort since the syntax is not quite the same. The biggest hurdle is figuring out how to log and drop a packet. In ipchains it is one command, in iptables you must create a new chain that does both actions and redirect packets to that.
You assume the likes of Asherons & EQ are balanced. I only have experience of EQ (two years worth), but I can tell you this - it was horribly, disgustingly, offensively unbalanced.
Essentially the situation was that unless you were some no-life loser who spent 10 hours a day on the system you were doomed to slowly, very slowly slog, slog away watching that exp bar crawl up one pixel at a time. After an eternity you raise a level, learn a few new skills or spells and repeat. The process for the casual player (as in a few hours a night) was just an exercise in tedium. There was no 'balance' here - the game was tuned to make progress as slow and as painful as possible. Worse, it was tuned assuming folks were twinked with unfeasibly over powered armour and weapons. So unless you wasted a disproportionate amount of time raising funds to buy uber gear you stood no chance of progressing because the mobs would murder you in a second.
It wasn't just the game that was the problem. Patch after patch and expansion after expansion demonstrated beyond a doubt that Verant didn't give a shit about the casual player. Every single expansion without exception has been deliberately aimed at the high level player. Sure you might see some 'newbie' zones but by and large expansions were developed for 30+ players, i.e. those already 'hooked'.
So casual gamers could basically fuck off. If you weren't constantly running EQ fullscreen for 10 hours a day there was little chance of progress. After the abysmal Shadows of Luclin expansion followed by a price rise I dumped the damned thing and I'm glad I did. It was a wrench to be sure, no doubt from the Skinner box like reward model, but I'm happy to be rid of the bloody thing. I don't think short of some extraordinarily positive reviews I would ever touch an online game written by Sony or Verant again.
Now addressing your points of EQ / AC players moving back from version 2 to 1. I would not be surprised if Verant and Microsoft offered 'migration' paths that enable characters to upgrade from the old version to the new one. I would be extremely surprised if they offered a route in the other direction.
Dell & Compaq both sell higher specced laptops, complete with windows tax for $799. I don't see what the big deal was about the Lindows one.
It seems odd to me that it was getting any press at all. Underpowered and overpriced would be apt descriptions. The only thing going for it would be the weight, but that is a high price to pay for all the things your money didn't get you such as a CDROM even.
But it is also a tax on everyone else. Whether you think people are stupid or not, it is still a crime and still requires investigation as does any other. Not to mention that money flying out of the country never to be seen again also affects the economy in other detrimental ways.
I have no problem with it being a database, just with it being a critical database with little structure and static settings and dynamic settings and generally any and all crap being stuffed into it with little concern to its importance to the system or enforcement to stop one program peppering settings in all kinds of weird places.
Like any database (or configuration), settings should fall into logical categories and the underlying file structure should follow that. It's like designing an SQL database where every field is a blob and employee records are kept in the same table with the number of toilet rolls and the price of a cup of tea. Something crashes the system while the price of a cup tea is being modifed and the whole thing is screwed!
That alone would make the whole thing more palatable. Of course text files also mean that if you are ever unlucky to find yourself stuck on the command line that you can attempt manual repairs with a text editor. Not so with the tens of megabytes of registry.
Personally I think/etc is a bit of a mess in terms of consistency. Every app has its own file format and own parsing routines. It would be great if there were a common lib akin to PAM for security, which could read and write the common config file formats and that the apps could use instead. This alone would make writing graphical tools over the top of these files considerably easier than it is at present. Still, to my mind it beats the registry hands down.
Well, dolt (since you called me idiot). The issue is not FS corruption but recovering from that corruption. The registry is a couple of files, it is live and continuously changing. If the power goes during a write and that file gets trashed, you are screwed. And in my day to day use of NT/W2K/XP for the last six years I have seen the registry fuck up or the system become unbootable about a couple of times a year. When it happens you might be lucky and recover (even with strangeness such as every hardware device being redetected), but just as likely you're grovelling around in the recovery console or reaching for rescue disks.
In the same period of time, my Linux firewall has been screwed a couple of times, each time I was able to recover quite handily. Now as my Linux firewall is currently running on my old NT/W2K box (I have a new machine for XP), I know it has nothing to do with the hardware. Simply put, Linux is far more recoverable and part of that is the way it stores files.
I have yet to form an opinion of what OS X is like (another machine I have) since it throws out a lot of the Unix configuration files for its own thing, so possibly it is a half way house. So far though, I have seen two kernel panics and a dozen or so lockups in two years with full recovery from them all which isn't bad going. Compared to XP it is still twice as reliable.
The simple fact is that a system consisting of many config files, which are all static are less likely to be trashed and if they do, there is still a good chance that you can boot up and fix whatever corruption there is. Even kludging the registry with hives and transactional logs still doesn't approach that reliability.
In my day to day experience of Windows I have seen ten or so serious screw-ups which have been registry related. This isn't machine specific, but has been due to bluescreens and power failures from which the computer has been unable to reboot from. Now if you're lucky the computer will manage to recover, but if it doesn't... then you're hosed since you have no idea what bit of the registry is broken and you just have to pray a backup will fix things for you. This is in stark contrast to multiple config files where you can normally figure out where the problem is from the boot sequence or what service isn't running.
Just remember to never, ever encrypt your temp folder when you're an admin. While this might seem a Good Idea considering all the crap that apps write in there, it is actually suicidally dangerous since service packs and other installers also use the temp folder and you stand a good chance of overwriting system dlls with new versions which have been encrypted by being unpacked into the temp folder first. Other users can't open those dlls and the system screws up in all sorts of weird ways.
It took me hours to fix a system I broke this way. The sad part is, the XP online help actually recommended I do this in its best practices section on using encryption.
The registry is an awful thing for the simple reason it sticks all your eggs in one basket. Now I know technically there are various 'hives' but if the registry gets corrupted in any signifcant way you are completely screwed whether one hive is nobbled or another.
Your choices after that boil down to - restoring from a backup registry and praying that it works, or reinstalling. The recovery console is a joke and a last ditch effort. The only times I've required it are when I foolishly marked my temp folder as encrypted and a service pack used it before peppering my system32 dir with encrypted files and during recent filesystem data corruption. On neither occasion was it particularly useful and I was sorely pushed each time to recover to a working system.
At least Unix gives you a fighting chance since configuration files are all individually named and occupy different places on the disk. It is quite possible to identify the precise problem and fix it if necessary. Those files might be messier, but at least its easy to back them up (since they're not 'live') and *much* easier to restore them. It is my opinion that the registry is quite possibly the most awful things about Windows, even before considering the mess of registry keys it actually contains.
As has already been stated, there are ways to do this without disclosing information. Most Linux distributions manage it quite handily by fetching a list of packages from the server and comparing it to the local installation. The local software then fetches new versions of the packages it requires without disclosing what other packages or third party software might or might not be on the machine.
Why can't Microsoft do the same? The answer of course is they could, but then they wouldn't have the luxury of knowing what software is installed on machines, drivers, unpatched versions of IE, and so forth. Whether they use it in an underhanded way is debateable, but one could easily envisage that they could use it in various nefarious ways such as being able to monitor how successful their competitors are, how many people are using their software (to determine piracy rates etc.), geographic usage of their software and other nuggets of info.
Well if you want to sit down, you might as well buy a disabled cart. Do a Homer Simpson and become morbidly obese and you might even qualify to get one for free.
Well there is a solution for that - an on/off switch on the backlight.
As for ambient light, I find it hard to believe they paid any attention to that at all or they would have used a matt finish on the screen to diffuse the light rather than the shiniest reflective surface they could find to bounce most of it back into your eyes.
The GBA shipped with a shiny, glarey, barely visible screen with no backlight. Was this just gross incompetance on the part of Nintendo or part of some scheme to keep selling us new models that promise to fix the massive deficiencies in the last one? I find it hard to believe they didn't know, all the way through developing and testing the thing that the screen was so appalling. How much would the backlight circuitry have cost them? A few cents maybe?
Frankly, I might have bought more games for my GBA if I thought I could actually see them. Instead it sits in a drawer except for occasional forays in good lighting. Nintendo can go to hell if they think I'm going to make the same mistake twice.
Slashdot Mirror
Now, I would welcome any replacement to conventional IDE / ATA which has been the bane of my life. I couldn't count the number of times I've had to screw around swapping cards and drives in order to accomodate that ribbon. I will be happy to see that particular technology go the way of the dodo.
My Palm Pilot Vx has had an uptime that must measure in years now. And that's with day to day use. The only time I had to reboot it was when I updated it to PalmOS 4.
I would be most upset if I 'upgraded' to a Windows CE or Linux PDA and discovered the thing died for no apparant reason whenever it felt like it. Even once a month would be unacceptable. If this is a common thing in CE or the Zaurus (I don't know), then it is absolutely shocking that people put up with it.
Space is a pretty weak reason for not including one at all. Lot's of MP3 / MiniDisc players are faced with the same size constraints yet manage to design a proper solution. A smaller jack and a couple of buttons could control the volume (no need for a wheel) or do what many minidisc players do and stick the volume controls onto the cord of the headphones itself.
As it stands, it sounds like Nintendo have an utterly clueless testing department to miss great big howlers like this, or more likely it is a cynical ploy to release sucessive Gameboy versions with major design flaws to drive sales for add-ons and the next one Gameboy which will supposedly fix the problem.
Well I don't think I'll be buying one. The GBA was the last straw for me - the screen is just atrocious and I'm not going to make the same mistake again.
Contrast that with IE where none of these things are possible. Got a bug? Good luck trying to raise and track the issue with Microsoft.
Some security there.
I apologise for my outburst. I was wrong and should have checked first. Sorry.
I was too harsh and incorrect, I apologise completely.
Why does anyone care what jwz thinks about Mozilla? He hasn't been involved with the project in any way at all during the last four years. None whatsoever. Except griping from the sidelines of course.
There is no 'hotfixing' or piece patching here. The result of the incremental diff is the same as installing the whole new version, just considerably easier to download. As I mentioned, a kernel update is 35mb of patches. I would be surprised if an equivalent incremental patch were more than a megabyte.
Why isn't it possible to produce incremental binary patches containing just the diffs? Not only would it vastly increase the chances of people downloading them (which is good for everyone), it is good for Red Hat too since their bandwidth for up2date is slashed.
Now obviously there are times when incremental diffs are not useful, but with the proper safeguards (e.g. checksums and backing up originals etc.) I don't see what the problem is. If anyone from Red Hat or RPM land is listening, please consider implementing this feature. Pretty please with a cherry on top.
MozillaQuest Magazine: So why are we are such a bad source for news?
Linus Torvalds: Because you're an utterly clueless one man operation that has nothing pleasant, informed or useful to say about anything. Having demonstrated woeful ignorance of your namesake Mozilla and after being shunned by that community you're now expanding your ignorance about other topics too. In fact I am surprised I am even talking to you - perhaps someone should have warned me.
But Mandrake is French. Let the assholes boycott that if they want to make an ineffectual protest at another nation that dares to have its own foreign policy and interests.
It's better to rewrite everything to use iptables, though this does require some effort since the syntax is not quite the same. The biggest hurdle is figuring out how to log and drop a packet. In ipchains it is one command, in iptables you must create a new chain that does both actions and redirect packets to that.
Essentially the situation was that unless you were some no-life loser who spent 10 hours a day on the system you were doomed to slowly, very slowly slog, slog away watching that exp bar crawl up one pixel at a time. After an eternity you raise a level, learn a few new skills or spells and repeat. The process for the casual player (as in a few hours a night) was just an exercise in tedium. There was no 'balance' here - the game was tuned to make progress as slow and as painful as possible. Worse, it was tuned assuming folks were twinked with unfeasibly over powered armour and weapons. So unless you wasted a disproportionate amount of time raising funds to buy uber gear you stood no chance of progressing because the mobs would murder you in a second.
It wasn't just the game that was the problem. Patch after patch and expansion after expansion demonstrated beyond a doubt that Verant didn't give a shit about the casual player. Every single expansion without exception has been deliberately aimed at the high level player. Sure you might see some 'newbie' zones but by and large expansions were developed for 30+ players, i.e. those already 'hooked'.
So casual gamers could basically fuck off. If you weren't constantly running EQ fullscreen for 10 hours a day there was little chance of progress. After the abysmal Shadows of Luclin expansion followed by a price rise I dumped the damned thing and I'm glad I did. It was a wrench to be sure, no doubt from the Skinner box like reward model, but I'm happy to be rid of the bloody thing. I don't think short of some extraordinarily positive reviews I would ever touch an online game written by Sony or Verant again.
Now addressing your points of EQ / AC players moving back from version 2 to 1. I would not be surprised if Verant and Microsoft offered 'migration' paths that enable characters to upgrade from the old version to the new one. I would be extremely surprised if they offered a route in the other direction.
It seems odd to me that it was getting any press at all. Underpowered and overpriced would be apt descriptions. The only thing going for it would be the weight, but that is a high price to pay for all the things your money didn't get you such as a CDROM even.
But it is also a tax on everyone else. Whether you think people are stupid or not, it is still a crime and still requires investigation as does any other. Not to mention that money flying out of the country never to be seen again also affects the economy in other detrimental ways.
Like any database (or configuration), settings should fall into logical categories and the underlying file structure should follow that. It's like designing an SQL database where every field is a blob and employee records are kept in the same table with the number of toilet rolls and the price of a cup of tea. Something crashes the system while the price of a cup tea is being modifed and the whole thing is screwed!
That alone would make the whole thing more palatable. Of course text files also mean that if you are ever unlucky to find yourself stuck on the command line that you can attempt manual repairs with a text editor. Not so with the tens of megabytes of registry.
Personally I think
In the same period of time, my Linux firewall has been screwed a couple of times, each time I was able to recover quite handily. Now as my Linux firewall is currently running on my old NT/W2K box (I have a new machine for XP), I know it has nothing to do with the hardware. Simply put, Linux is far more recoverable and part of that is the way it stores files.
I have yet to form an opinion of what OS X is like (another machine I have) since it throws out a lot of the Unix configuration files for its own thing, so possibly it is a half way house. So far though, I have seen two kernel panics and a dozen or so lockups in two years with full recovery from them all which isn't bad going. Compared to XP it is still twice as reliable.
The simple fact is that a system consisting of many config files, which are all static are less likely to be trashed and if they do, there is still a good chance that you can boot up and fix whatever corruption there is. Even kludging the registry with hives and transactional logs still doesn't approach that reliability.
In my day to day experience of Windows I have seen ten or so serious screw-ups which have been registry related. This isn't machine specific, but has been due to bluescreens and power failures from which the computer has been unable to reboot from. Now if you're lucky the computer will manage to recover, but if it doesn't... then you're hosed since you have no idea what bit of the registry is broken and you just have to pray a backup will fix things for you. This is in stark contrast to multiple config files where you can normally figure out where the problem is from the boot sequence or what service isn't running.
It took me hours to fix a system I broke this way. The sad part is, the XP online help actually recommended I do this in its best practices section on using encryption.
Your choices after that boil down to - restoring from a backup registry and praying that it works, or reinstalling. The recovery console is a joke and a last ditch effort. The only times I've required it are when I foolishly marked my temp folder as encrypted and a service pack used it before peppering my system32 dir with encrypted files and during recent filesystem data corruption. On neither occasion was it particularly useful and I was sorely pushed each time to recover to a working system.
At least Unix gives you a fighting chance since configuration files are all individually named and occupy different places on the disk. It is quite possible to identify the precise problem and fix it if necessary. Those files might be messier, but at least its easy to back them up (since they're not 'live') and *much* easier to restore them. It is my opinion that the registry is quite possibly the most awful things about Windows, even before considering the mess of registry keys it actually contains.
Why can't Microsoft do the same? The answer of course is they could, but then they wouldn't have the luxury of knowing what software is installed on machines, drivers, unpatched versions of IE, and so forth. Whether they use it in an underhanded way is debateable, but one could easily envisage that they could use it in various nefarious ways such as being able to monitor how successful their competitors are, how many people are using their software (to determine piracy rates etc.), geographic usage of their software and other nuggets of info.
Well if you want to sit down, you might as well buy a disabled cart. Do a Homer Simpson and become morbidly obese and you might even qualify to get one for free.
As for ambient light, I find it hard to believe they paid any attention to that at all or they would have used a matt finish on the screen to diffuse the light rather than the shiniest reflective surface they could find to bounce most of it back into your eyes.
Frankly, I might have bought more games for my GBA if I thought I could actually see them. Instead it sits in a drawer except for occasional forays in good lighting. Nintendo can go to hell if they think I'm going to make the same mistake twice.