The problem is that, 100 years from now, if your RAID controller goes down, will you be able to get a repalcement? Will we still even be using disks 100 years from now?
The solution is, it really doesn't matter, so long as we use something that even remotely resembles a filesystem. In fact, you say as much:
Fine, you might say, if there's no disks or Google or whatever you can always copy your data to the latest and greatest storage medium. But there's the rub: it costs, to copy and to maintain this stuff.
Yes, it does. It also costs money to store paper.
If you write something on paper, after you write it once, you're done for the next 100 years or more, as long as you store it correctly. The cost of maintaining your paper data for a long time is much lower than it is for electronic data.
I'm not convinced of that. Let me put it this way: My home directory contains all the files that I care about. Records, notes, papers, old school projects. It is currently 32 gigs, and a significant portion of that is miscellaneous media, downloads, and Windows games. Even the "important" stuff contains files five years old or more, that I have no use for.
32 gigs seems like a lot, doesn't it? Well, I currently have some 350 gigs available on this partition, which I use to store movies, games, work on programming projects, install miscellaneous software, and generally mess around with. It's been suggested that Moore's Law is true just as much, if not moreso, for storage space. Therefore, every time I upgrade my computer, I'm generally going to end up with as much or more space available, much of it used for such temporary things as Doom 3 and Quake 4 (which I probably won't play again).
So, I'll grant you this point:
Granted, you'll want an electronic copy of any important data you create. But who's to say that what people consider important now will still be important in 100 years? Or that someone in the future will find value in something we consider unimportant today?
Well, for one, I sincerely doubt I'll consider a copy of Star Trek: The Next Generation to be important in 100 years, especially when it's not even close to the only such copy.
The point is, I currently have no use for paper, or the kind of storage I'd need for such paper. I do have a use for computer storage, which will be increasing exponentially for the forseeable future. So, for a document to last my lifetime, I just have to carry it with me from storage device to storage device. Since each storage device will have vastly more space than I need for important documents, the cost of maintaining records is negligable, so long as I, personally, have a use for such a device at the moment.
If no one cares to set aside a few gigs of spare storage for my documents, I probably don't have anything worth keeping anyway.
And to answer your question about Google: Maybe not, but you can distribute your documents among multiple online services. In Google's case, they, too, have a current, urgent need for vast amounts of storage. Setting aside 2 gigs per Gmail account, or for your spreadsheets and documents, is really not a big deal for them, even to carry such documents into perpetuity, since they are already paying the power bills, they're already paying for new hardware, all for massively more amounts of data which they will likely throw away -- think Google Cache.
You could say paper is kind of the same way. If you already have a cool, dry, dark place to store your documents, and you're already creating them on paper in the first place, then a filing system doesn't cost much, and you get to carry it 100 years into the future for free. Well, same for storage -- if I already have a redundant digital store that I'll be replacing every five or ten years anyway, then it costs me almost nothing to throw in archival documents.
Actually, on Linux, almost every DVD playing app I know of uses libdvdread for handling DVDs. While it is possible to coerce them into playing a VIDEO_TS folder, or use a DVD image as the "device", the usual method of DVD playback is to not even bother mounting the disc, and simply read directly from the device.
I do not know if they use UDF or the IFO method. I do know that it would probably be pretty trivial to implement a workaround in libdvdread, and that while under Windows, copying a VIDEO_TS folder may be the easiest way to rip a DVD, under Linux, it's trivial to just "cp/dev/dvd foo.iso" and be done with it, meaning that we can rent them and pirate them now, and play them back later.
What's going to be hilariously retarded is when they discover how many hardware players will refuse to play this, and just how easy it is to break with software players. As usual, only the pirates win.
This is just corruption, based on the assumption that "real" DVD players behave one way, and software DVD players behave another way.
I'm really curious to see if this affects me at all. I have a strong suspicion that libdvdread won't care, or could be patched to deal with this. And it STILL doesn't solve the problem of simply ripping the DVD bit for bit.
Also, I can archive my collection of documents, including my email, onto a single 650MB CD, without the attachments. The attachments get archived less frequently on their own CD. If they were embedded in the mails I couldn't do this.
In your case, I'd suggest a simple solution -- buy a DVD burner. You could even reuse the same disc 3 or 4 times, by making it multisession. I'd also suggest incremental backups -- you could backup your everyday changes onto far less than a CD, and much less frequently, you could do a full backup onto a DVD, or a pair of CDs.
So what's your idea to change Thunderbird to automatically extract attachments when mail arrives, save it in a separate folder, and present it as part of the email when viewing?
I don't have such an idea, and I don't think it's necessarily a good idea, which is why I asked what your specific needs were.
Eudora is a little slow sorting, searching, etc through mailboxes with several thousand messages in them.
My solution would be to use IMAP. In fact, I'd like to see Thunderbird using IMAP locally anyway, but in my case, I store all my mail on an IMAP server in the next room, which has the nice side effect of making it available to my laptop. Now, IMAP abstracts away things like searching through mail, and accessing a single message. My IMAP server stores the mail in maildir format, which means it takes exactly as much time to manipulate 50 megs of emails as it does to manipulate 500 megs, assuming it's the same number of messages. As for searching/sorting, my IMAP server doesn't do it properly, but IMAP does have an abstraction for searching, and I believe there's even some sort of protocol in place for creating sorting rules.
Thus, it should be possible to create a GUI in Thunderbird to implement server-side sorting rules, which is more efficient anyway if your ISP lets you do IMAP, or if you run your own server. It should also be possible to modify an IMAP server to do indexed searching, the way an actual search engine would -- Spotlight doesn't seem to have much of a problem searching most of my hard disk, so IMAP should be able to handle it fairly well, if the server keeps an index. Unfortunately, I don't know of any IMAP server that keeps a full-text index, but that's where I'd implement the functionality.
Bonus side-effect: This is no longer limited to Thunderbird. You could keep using the old Eudora, even, if it's a decent IMAP client.
If you want access to your records 100 years from now, put them down on acid-free paper, and store them in a controlled environment.
And when that controlled environment fails, or soemone flat-out steals those records? Paper doesn't inherently offer redundancy. It's also harder to keep up-to-date.
CDs, hard drives, floppies -- all crap, in the long run.
It doesn't have to be a single CD, or hard drive, or floppy, or whatever.
If you're thinking of archives we want to be around for centuries, that's easy enough. Put them on a server with a fairly large RAID array, and replicate it over the Internet to another datacenter or two. If one hard drive dies, you swap it out for another. If one RAID controller or whole box goes down, hard, you build a new one and replicate the data back. If you don't want the hassle of doing this yourself, especially if it's just a small amount of personal data, you get Google to do it for you.
I don't know of a good way to achieve that level of redundancy with paper, not cheaply, and certainly not if you want to be able to keep updating constantly.
For archives and books, paper's still the way to go.
This may be an old-person issue, also -- I actually don't mind reading a book on a decent LCD. eInk would be better, but LCDs aren't bad.
often, companies resort to open-source implementations when the remaining engineers can't properly update/maintain the existing codebase. I've seen it happen; either deadlines force your hand, or there's just too much low-level work to get the engine to support the new features you want.
That may be the motivation, but I, for one, am happy. I'm hoping for the best of both worlds.
Think about it: We could keep bickering about BSD vs Linux, and BSD would have some features, Linux would have others, and neither would be any good. Now, most every distro is on Linux, not BSD, and new features go into Linux. Sad if you liked BSD, but a very good thing if you don't want to have to choose between BSD recognizing your printer and Linux recognizing your scanner, or something equally stupid.
I'd rather have Eudora features rolled into Thunderbird, or vice versa, than have them keep maintaining Eudora separately. Monopolies in open source are generally a good thing, until politics get in the way.
This app is nowhere near ready to be considered an actual spreadsheet.
Have you considered that maybe it's just the import/export? I suspect it's actually very useful as an "actual spreadsheet", assuming it stands alone.
I realize it may be entirely useless to you, but "imports Excel perfectly" is not, repeat, is absolutely fucking not, a requirement to be an "actual spreadsheet".
I've disabled tabs in Fluxbox, because I just don't find them particularly useful. I'd much rather have windows that arrange themselves nicely on my 1600x1200 screen, but I need a good way of cycling through related windows, even if I can see them all at once. I would also like my window manager to do more housekeeping for me.
But basically, everything you're describing already exists in Fluxbox. As far as the app is concerned, it's just a window, but you can group windows into tabs, drag them from one group of tabs to another, or into a separate window, cycle through tabs with one universal keystroke, etc.
It's also not as efficient. For instance, Firefox with a bunch of tabs is much more efficient than Firefox with a bunch of windows. This could be helped with Apple's concept of toolbars that float below the menu bar, and a common menu bar in the first place, but it would require a lot of help from apps.
It was one of only a few clients early on that supported multiple email accounts,
How well, compared to what Thunderbird does now?
because of how it stores email in flat text files (as opposed to Outlook and some others) it was really easy to migrate your mailboxes and settings from computer to computer - even between platforms ie moving from Windows to Mac.
This is why I currently store all my email on a Linux IMAP server, in Maildir format. My settings may have to be moved, but then, the important ones happen on the server -- like filtering, for instance. And my mail stays synced that way between my Powerbook and Linux machine.
The filter tools are starting to show their age, but are still solid. There was a point where I would definitely say Eudora's filtering tools were the best in any commercial email client.
Perhaps. I still use server-side filters, and with good reason, I think.
Hopefully both Eudora and Thunderbird benefit from this.
Here we agree. In fact, I'd much rather not have to choose, if there's simply one, better choice.
We will only be a paperless (or vellumless, parchment, etc) society when a more reliable form of data storage is available. That day is a LONG WAY off.
While it will take awhile to prove, I suspect that digital representations can be more reliable and less forgeable than paper. The problem is, computers are some of the most mishandled tools around.
I wouldn't trust paper either if clueless people were always accidentally burning them ("I thought you could set a book down anywhere! How was I to know the stove would burn it?"), storing them in cool, wet places ("Oh, you meant a cool, dry place?"), using important documents for scratch paper or making paper hats out of them... You get the idea.
Reliable computer storage is pretty much here, for those who care to use it. The problem is not that computers are so unreliable, or even that Windows is so insecure. The problem is that most people have plenty of education surrounding other aspects of their life -- they are required to learn arithmetic and geography in school, they need to pass Driver's Ed to get a license. Nobody uses a paper filing system without having some idea of what's going on, and paper filing systems are generally redundant. But we are not required to take a class on basic computer usage, that explains things like the necessity to keep backups, or when not to reflexively hit "OK".
So, reliable computers are easily possible. Reliable users are a long way off.
True, but I'm suggesting that Reiser4 could become stable, but JFS will never be as fast. Therefore, I'm not second-guessing your decision to use JFS now, but I do think a company like IBM might have an interest in finishing Reiser4. Even Google might, but I don't know at all how well their Google Filesystem performs.
After all, the performance and storage characteristics of JFS and Reiser4 are mostly by design/spec now, whereas any stability issues are bugs to be fixed. And if you change the fundamental design and on-disk format of a filesystem, is it still the same filesystem? If there was, say, a JFS2 which outperformed Reiser4, I'd suggest it would take as much work as Reiser4 has since Reiserfs3.
This only makes sense because Firefox is the default browser of many distros. What about a KDE-based distro that defaulted to Konqueror?
Most current Linux users would accept the new name with hardly a comment. It would, however, be yet another bad thing for marketing Linux to Windows users. It's no wonder Ubuntu caved.
From my own experiences on lkml and reiserfs-list, I would guess he is innocent.
Is he an arrogant asshole sometimes? Yes. Does he have a bit of a temper? Yes. Does he take things personally far too often? Yes. Does it surprise me that he's divorced? No.
I would be surprised (happy, but surprised) if he was still married. But I would be shocked if he killed anyone.
Shocking things do happen, every day. I have no way of really knowing, but until we know more, I call him innocent.
That said, he's pretty much of an arrogant asshole and Reiser4 is crap. Why would IBM pick it up when they sponsor the totally superior JFS?
Because at least some of his arrogance is justified. Reiser4 is the fastest filesystem I've tried, certainly faster than JFS, and more space-efficient, even without cryptocompress finished.
If IBM could make it as stable as JFS, it would be "totally superior" to JFS.
I've made as dark a joke as anyone else, but I've been following the development for awhile, and I was using Reiser4 before it was released -- hell, before anyone even pretended it was stable. For all the furor on Slashdot, it seems reiserfs-list is having a moment of silence.
I can't say I know him well, and he certainly is loud and offensive whenever lkml and reiserfs-list collide, but he does have a vision, and he was encouraging to the young, inexperienced wannabe kernel hacker that I was, and still am. I really hope he didn't do it.
The only possible way this could work is if you managed to sell it to a completely separate group of users. Maybe if you managed to convince them all that it was iNet or iWeb, the shiny new plastic browser from Apple, or maybe if you had a killer app for it, but seriously, I don't see a Firefox derivative replacing Firefox unless the name is moderately similar and the app is measurably better.
Just remember how much time and money it took for Firefox to completely eclipse both Mozilla and Netscape. And Firefox was measurably better than both -- at least, by the time it was actually named Firefox.
So far RedHat, SUSE and Ubuntu have agreed to cede control over ALL modifications, including prior approval of security patches to Moz Corp. Obviously Debian couldn't, wouldn't and shouldn't have done anything of the sort.
Here's the problem: Suppose Mozilla were to give Debian full control, Debian patches the hell out of it, and people say "Firefox sucks! It crashes all the time on Debian!" Now, suppose Debian gave Mozilla full control, Mozilla doesn't allow Debian-specific patches required to make it work properly, so people say "Debian sucks! Firefox crashes all the time on it, but not on Ubuntu!"
Both of them have legitimate complaints.
I already made the change earlier in the year. Done right FF plugins still work so no big deal.
One big deal: A newbie coming over from Windows looking for FF plus plugins/extensions won't find it, and won't have a clue that IceWeasel is really Firefox. They should've at least attempted to make it clear that it's still Firefox, it's just Debian-specific.
I'd have to go with tradition here. Distros get to release derivative versions, and still call them by the original name. In return, distros do the best to make everything play nice, and generally will listen to reasonable requests -- for instance, Gentoo removed the ebuild that built Cedega (then WineX) from the CVS, because although it was technically legal (they allow CVS access, but charge for prebuilt packages), it made it just as easy, if not easier, for Gentoo users to use the free (CVS) version than to subscribe.
Distros have to keep in mind that users will just go and get upstream by themselves if the distro gets it wrong, or they'll switch to another distro that gets it right. Upstream has to keep in mind that if they refuse to cooperate with distros, they won't get distributed, so they should at least make an attempt to play nice with distros and other packages.
In this case, neither is willing to allow full control, and both are paranoid that the other side will tarnish their good name. Because of this mutual stubbornness, both sides lose out. I will likely never prefer Debian over Ubuntu again, and not just for this reason.
I call BS. Net Neutrality is about packet privatization at the ISP or line carrier level.
True, in the same way that the GPL is about always distributing source when you distribute a binary. But the spirit of the GPL runs deeper, and so does the spirit of net neutrality.
Net Neutrality is about Telcos (ISPs and Line Carriers) slowing down or stopping time-sensitive packets (VOIP, music, IPTV etc.) going to or from a competitor that happens to use their lines or because you use their ISP service.
No, that is what the world would be like without net neutrality. The principle of net neutrality is that the network itself should be neutral. The question is whether IM, at this point, is a network of sorts or a service. But still, Google is a service. Suppose Google silently dropped any search results from Yahoo.com?
Packet prioritization is simply one aspect of net neutrality. In its current definition, you're right, net neutrality pretty much only applies to telcos -- but it is not limited to prioritization. I call it a violation of net neutrality when a high school sends you through the N2H2 filter, where it not only attempts to block pornography, it adds an N2H2 banner to the bottom of each page.
Slashdot Mirror
The solution is, it really doesn't matter, so long as we use something that even remotely resembles a filesystem. In fact, you say as much:
Yes, it does. It also costs money to store paper.
I'm not convinced of that. Let me put it this way: My home directory contains all the files that I care about. Records, notes, papers, old school projects. It is currently 32 gigs, and a significant portion of that is miscellaneous media, downloads, and Windows games. Even the "important" stuff contains files five years old or more, that I have no use for.
32 gigs seems like a lot, doesn't it? Well, I currently have some 350 gigs available on this partition, which I use to store movies, games, work on programming projects, install miscellaneous software, and generally mess around with. It's been suggested that Moore's Law is true just as much, if not moreso, for storage space. Therefore, every time I upgrade my computer, I'm generally going to end up with as much or more space available, much of it used for such temporary things as Doom 3 and Quake 4 (which I probably won't play again).
So, I'll grant you this point:
Well, for one, I sincerely doubt I'll consider a copy of Star Trek: The Next Generation to be important in 100 years, especially when it's not even close to the only such copy.
The point is, I currently have no use for paper, or the kind of storage I'd need for such paper. I do have a use for computer storage, which will be increasing exponentially for the forseeable future. So, for a document to last my lifetime, I just have to carry it with me from storage device to storage device. Since each storage device will have vastly more space than I need for important documents, the cost of maintaining records is negligable, so long as I, personally, have a use for such a device at the moment.
If no one cares to set aside a few gigs of spare storage for my documents, I probably don't have anything worth keeping anyway.
And to answer your question about Google: Maybe not, but you can distribute your documents among multiple online services. In Google's case, they, too, have a current, urgent need for vast amounts of storage. Setting aside 2 gigs per Gmail account, or for your spreadsheets and documents, is really not a big deal for them, even to carry such documents into perpetuity, since they are already paying the power bills, they're already paying for new hardware, all for massively more amounts of data which they will likely throw away -- think Google Cache.
You could say paper is kind of the same way. If you already have a cool, dry, dark place to store your documents, and you're already creating them on paper in the first place, then a filing system doesn't cost much, and you get to carry it 100 years into the future for free. Well, same for storage -- if I already have a redundant digital store that I'll be replacing every five or ten years anyway, then it costs me almost nothing to throw in archival documents.
This is true from a personal level on up to
Actually, on Linux, almost every DVD playing app I know of uses libdvdread for handling DVDs. While it is possible to coerce them into playing a VIDEO_TS folder, or use a DVD image as the "device", the usual method of DVD playback is to not even bother mounting the disc, and simply read directly from the device.
/dev/dvd foo.iso" and be done with it, meaning that we can rent them and pirate them now, and play them back later.
I do not know if they use UDF or the IFO method. I do know that it would probably be pretty trivial to implement a workaround in libdvdread, and that while under Windows, copying a VIDEO_TS folder may be the easiest way to rip a DVD, under Linux, it's trivial to just "cp
What's going to be hilariously retarded is when they discover how many hardware players will refuse to play this, and just how easy it is to break with software players. As usual, only the pirates win.
This is just corruption, based on the assumption that "real" DVD players behave one way, and software DVD players behave another way.
I'm really curious to see if this affects me at all. I have a strong suspicion that libdvdread won't care, or could be patched to deal with this. And it STILL doesn't solve the problem of simply ripping the DVD bit for bit.
In your case, I'd suggest a simple solution -- buy a DVD burner. You could even reuse the same disc 3 or 4 times, by making it multisession. I'd also suggest incremental backups -- you could backup your everyday changes onto far less than a CD, and much less frequently, you could do a full backup onto a DVD, or a pair of CDs.
I don't have such an idea, and I don't think it's necessarily a good idea, which is why I asked what your specific needs were.
My solution would be to use IMAP. In fact, I'd like to see Thunderbird using IMAP locally anyway, but in my case, I store all my mail on an IMAP server in the next room, which has the nice side effect of making it available to my laptop. Now, IMAP abstracts away things like searching through mail, and accessing a single message. My IMAP server stores the mail in maildir format, which means it takes exactly as much time to manipulate 50 megs of emails as it does to manipulate 500 megs, assuming it's the same number of messages. As for searching/sorting, my IMAP server doesn't do it properly, but IMAP does have an abstraction for searching, and I believe there's even some sort of protocol in place for creating sorting rules.
Thus, it should be possible to create a GUI in Thunderbird to implement server-side sorting rules, which is more efficient anyway if your ISP lets you do IMAP, or if you run your own server. It should also be possible to modify an IMAP server to do indexed searching, the way an actual search engine would -- Spotlight doesn't seem to have much of a problem searching most of my hard disk, so IMAP should be able to handle it fairly well, if the server keeps an index. Unfortunately, I don't know of any IMAP server that keeps a full-text index, but that's where I'd implement the functionality.
Bonus side-effect: This is no longer limited to Thunderbird. You could keep using the old Eudora, even, if it's a decent IMAP client.
And when that controlled environment fails, or soemone flat-out steals those records? Paper doesn't inherently offer redundancy. It's also harder to keep up-to-date.
It doesn't have to be a single CD, or hard drive, or floppy, or whatever.
If you're thinking of archives we want to be around for centuries, that's easy enough. Put them on a server with a fairly large RAID array, and replicate it over the Internet to another datacenter or two. If one hard drive dies, you swap it out for another. If one RAID controller or whole box goes down, hard, you build a new one and replicate the data back. If you don't want the hassle of doing this yourself, especially if it's just a small amount of personal data, you get Google to do it for you.
I don't know of a good way to achieve that level of redundancy with paper, not cheaply, and certainly not if you want to be able to keep updating constantly.
This may be an old-person issue, also -- I actually don't mind reading a book on a decent LCD. eInk would be better, but LCDs aren't bad.
Not every kind of business is built on DRM. Anyone who employs DRM is a bastard for doing so, in my book.
That may be the motivation, but I, for one, am happy. I'm hoping for the best of both worlds.
Think about it: We could keep bickering about BSD vs Linux, and BSD would have some features, Linux would have others, and neither would be any good. Now, most every distro is on Linux, not BSD, and new features go into Linux. Sad if you liked BSD, but a very good thing if you don't want to have to choose between BSD recognizing your printer and Linux recognizing your scanner, or something equally stupid.
I'd rather have Eudora features rolled into Thunderbird, or vice versa, than have them keep maintaining Eudora separately. Monopolies in open source are generally a good thing, until politics get in the way.
We already call him "The Chair". And then laugh at him.
I would take him more seriously if he would stop his stupid monkey dances and actually get a fucking clue once in awhile.
Have you considered that maybe it's just the import/export? I suspect it's actually very useful as an "actual spreadsheet", assuming it stands alone.
I realize it may be entirely useless to you, but "imports Excel perfectly" is not, repeat, is absolutely fucking not , a requirement to be an "actual spreadsheet".
I've disabled tabs in Fluxbox, because I just don't find them particularly useful. I'd much rather have windows that arrange themselves nicely on my 1600x1200 screen, but I need a good way of cycling through related windows, even if I can see them all at once. I would also like my window manager to do more housekeeping for me.
But basically, everything you're describing already exists in Fluxbox. As far as the app is concerned, it's just a window, but you can group windows into tabs, drag them from one group of tabs to another, or into a separate window, cycle through tabs with one universal keystroke, etc.
It's also not as efficient. For instance, Firefox with a bunch of tabs is much more efficient than Firefox with a bunch of windows. This could be helped with Apple's concept of toolbars that float below the menu bar, and a common menu bar in the first place, but it would require a lot of help from apps.
How well, compared to what Thunderbird does now?
This is why I currently store all my email on a Linux IMAP server, in Maildir format. My settings may have to be moved, but then, the important ones happen on the server -- like filtering, for instance. And my mail stays synced that way between my Powerbook and Linux machine.
Perhaps. I still use server-side filters, and with good reason, I think.
Here we agree. In fact, I'd much rather not have to choose, if there's simply one, better choice.
So, why do you want attachments in a separate directory?
I'm curious, because I think I might have a way for Thunderbird to do what you want it to.
While it will take awhile to prove, I suspect that digital representations can be more reliable and less forgeable than paper. The problem is, computers are some of the most mishandled tools around.
I wouldn't trust paper either if clueless people were always accidentally burning them ("I thought you could set a book down anywhere! How was I to know the stove would burn it?"), storing them in cool, wet places ("Oh, you meant a cool, dry place?"), using important documents for scratch paper or making paper hats out of them... You get the idea.
Reliable computer storage is pretty much here, for those who care to use it. The problem is not that computers are so unreliable, or even that Windows is so insecure. The problem is that most people have plenty of education surrounding other aspects of their life -- they are required to learn arithmetic and geography in school, they need to pass Driver's Ed to get a license. Nobody uses a paper filing system without having some idea of what's going on, and paper filing systems are generally redundant. But we are not required to take a class on basic computer usage, that explains things like the necessity to keep backups, or when not to reflexively hit "OK".
So, reliable computers are easily possible. Reliable users are a long way off.
That's incredibly odd, especially considering AIM talks to ICQ, and vice versa.
True, but I'm suggesting that Reiser4 could become stable, but JFS will never be as fast. Therefore, I'm not second-guessing your decision to use JFS now, but I do think a company like IBM might have an interest in finishing Reiser4. Even Google might, but I don't know at all how well their Google Filesystem performs.
After all, the performance and storage characteristics of JFS and Reiser4 are mostly by design/spec now, whereas any stability issues are bugs to be fixed. And if you change the fundamental design and on-disk format of a filesystem, is it still the same filesystem? If there was, say, a JFS2 which outperformed Reiser4, I'd suggest it would take as much work as Reiser4 has since Reiserfs3.
This only makes sense because Firefox is the default browser of many distros. What about a KDE-based distro that defaulted to Konqueror?
Most current Linux users would accept the new name with hardly a comment. It would, however, be yet another bad thing for marketing Linux to Windows users. It's no wonder Ubuntu caved.
From my own experiences on lkml and reiserfs-list, I would guess he is innocent.
Is he an arrogant asshole sometimes? Yes.
Does he have a bit of a temper? Yes.
Does he take things personally far too often? Yes.
Does it surprise me that he's divorced? No.
I would be surprised (happy, but surprised) if he was still married. But I would be shocked if he killed anyone.
Shocking things do happen, every day. I have no way of really knowing, but until we know more, I call him innocent.
Because at least some of his arrogance is justified. Reiser4 is the fastest filesystem I've tried, certainly faster than JFS, and more space-efficient, even without cryptocompress finished.
If IBM could make it as stable as JFS, it would be "totally superior" to JFS.
It is mostly safe on my own system.
I've made as dark a joke as anyone else, but I've been following the development for awhile, and I was using Reiser4 before it was released -- hell, before anyone even pretended it was stable. For all the furor on Slashdot, it seems reiserfs-list is having a moment of silence.
I can't say I know him well, and he certainly is loud and offensive whenever lkml and reiserfs-list collide, but he does have a vision, and he was encouraging to the young, inexperienced wannabe kernel hacker that I was, and still am. I really hope he didn't do it.
And I hope Nina is ok.
The only possible way this could work is if you managed to sell it to a completely separate group of users. Maybe if you managed to convince them all that it was iNet or iWeb, the shiny new plastic browser from Apple, or maybe if you had a killer app for it, but seriously, I don't see a Firefox derivative replacing Firefox unless the name is moderately similar and the app is measurably better.
Just remember how much time and money it took for Firefox to completely eclipse both Mozilla and Netscape. And Firefox was measurably better than both -- at least, by the time it was actually named Firefox.
Here's the problem: Suppose Mozilla were to give Debian full control, Debian patches the hell out of it, and people say "Firefox sucks! It crashes all the time on Debian!" Now, suppose Debian gave Mozilla full control, Mozilla doesn't allow Debian-specific patches required to make it work properly, so people say "Debian sucks! Firefox crashes all the time on it, but not on Ubuntu!"
Both of them have legitimate complaints.
One big deal: A newbie coming over from Windows looking for FF plus plugins/extensions won't find it, and won't have a clue that IceWeasel is really Firefox. They should've at least attempted to make it clear that it's still Firefox, it's just Debian-specific.
I'd have to go with tradition here. Distros get to release derivative versions, and still call them by the original name. In return, distros do the best to make everything play nice, and generally will listen to reasonable requests -- for instance, Gentoo removed the ebuild that built Cedega (then WineX) from the CVS, because although it was technically legal (they allow CVS access, but charge for prebuilt packages), it made it just as easy, if not easier, for Gentoo users to use the free (CVS) version than to subscribe.
Distros have to keep in mind that users will just go and get upstream by themselves if the distro gets it wrong, or they'll switch to another distro that gets it right. Upstream has to keep in mind that if they refuse to cooperate with distros, they won't get distributed, so they should at least make an attempt to play nice with distros and other packages.
In this case, neither is willing to allow full control, and both are paranoid that the other side will tarnish their good name. Because of this mutual stubbornness, both sides lose out. I will likely never prefer Debian over Ubuntu again, and not just for this reason.
Well, the feature isn't implemented yet, but it's not called defragmenting. It's called repacking.
(cue images of Hans repacking his wife... oh GOD no)
No one should ever have to see this.
True, in the same way that the GPL is about always distributing source when you distribute a binary. But the spirit of the GPL runs deeper, and so does the spirit of net neutrality.
No, that is what the world would be like without net neutrality. The principle of net neutrality is that the network itself should be neutral. The question is whether IM, at this point, is a network of sorts or a service. But still, Google is a service. Suppose Google silently dropped any search results from Yahoo.com?
Packet prioritization is simply one aspect of net neutrality. In its current definition, you're right, net neutrality pretty much only applies to telcos -- but it is not limited to prioritization. I call it a violation of net neutrality when a high school sends you through the N2H2 filter, where it not only attempts to block pornography, it adds an N2H2 banner to the bottom of each page.