Linux developer Hans Reiser's conviction for first-degree murder has re-ignited the debate about the future of Linux's various file systems. Prior to his arrest in 2006, Reiser was best known for devising the ReiserFS file system, which was included with several Linux distros. Reiser's contributions to open source are now being …
Not an open issue...
Removing the chief architect from the equation would have similar implications whether the project in question was open source or proprietary.
The fact that viable Linux FS alternatives exist at all is an endorsement of open architecture imho.
Not only the file systems are affected
I remember once I had a problem with a Raid controller driver. I tried to contact the author and unfortunately he had died in a helicopter crash. Help was not forthcoming from anywhere, not even from the OS vendor with whom we had a maintenance agreement. So for me, these particular raid controllers died with the author (In the case of Linux).
What this means is that if you rely on community support and it is not available after the author's gone, you're in trouble. At least I understand that this shouldn't happen in commercial s/w. If the author of some software leaves a company, he is replaced and support continues. It's not always the case with Open Source, much as I've come to depend on and love it. :(
I can't see this changing the filesystem's standing very much in either direction. Reiser may be in chokey before long, but that doesn't rule out his involvement in development at least on some level. And what with it being open-source and all, those other folk who had the (guile|chutzpah|masochism) to work on it -- both of them -- can of course continue to do so.
Will anyone who's invested significant effort in this project now boycott it because the lead dev is a big crook? Doubt it. I'd bet a higher sum on a few more people taking an interest as a result of the Killer Chic this filesystem now enjoys.
Tux icon? Rumour has it Nina R. was last seen mingling with the mackerel in his belly...
At least as a last resort, you can always hire a competent coder to kick the driver into shape. Expensive, difficult, but not impossible. This is rather tricky to do with equivalent closed source drivers.
On a vaguely related ranty note, the worst thing about many raid drivers for linux, et al, is that the hardware manufacturers are seldom helpful and forthcoming with hardware documentation. This isn't unique to raid cards of course, but it is a pain when one of the largest manufacturers behaves this way. Releasing buggy hardware makes reverse engineering extra tricky. Thankyou, adaptec.
Lots of spare time on his hands now....
I expect Mr Reiser will have the odd hour or two to spare over the next thirty years! There's probably not much of a better way of removing distractions than prison, and if he's on good behavior he'll probably have access to a computer and email so he can carry on working with the community.
"....It seems likely that one of these will emerge as the preferred file system for future versions of Linux...." Ermm, just like Linux settled on one desktop...? I don't see a convincing reason why any of the available file-systems should dominate to such an extent as to stop development with the others, and it may be better in the long run to have more than one option.
This is certainly not just a problem with Open Source. There are plenty of proprietary software packages out there that have heavy dependencies on an individual.
I think the best possible situation is one where you have a company like Red Hat supporting a driver for some hardware. You have the source code to it, and if for some reason they stop supporting it, you're not completely SOL. But you also know that there isn't one person who everything depends on.
The expectation with proprietary software is that the company doesn't rest on one person's shoulders, but it is the case surprisingly frequently. This is especially true in very technical, very specific solutions.
A filesystem is not that important
Filesystems can be replaced, people can't. It's a pain, but nothing compared to kids losing a mother as happened in this case.
There are umpteen different options in Linux anyway, just like there's umpteen of everything else. I've found XFS to be very good in the past and as an added bonus I believe it was developed by a team of non-murderers.
The File System Of Choice
Well it was an (admittedly minor) issue on a large Linux project I worked on last year.
"We've got enough support problems as it is without putting all our data in a filesystem designed by an axe murderer."
Besides, if the bloke isn't bright enough to figure out the rubber bungs in the floorpan of his mother's car when he gives it a good going over with the hosepipe (as one does from time to time), how good could his filesystem be?
Dead penguin because, well, it's obvious.
Re: Not only the file systems are affected
Ah yes, the “bus factor”. (Now there's a thought…)
If Hans Reiser can kill his wife....
...imagine what being married to an Exchange developer must be like.
How did Hans kill Nina
Just a niggling little detail. There is no fucking way they can prove forethought without knowing how she died thats one of things you have to know to prove a charge of first degree murder. You might as well have convicted him of being a freakish little programmer because thats why he was convicted.
Support and missing persons
Anybody who thinks that commercial products and "closed source" software are immune to this must be judging from inexperience. I've been on the bad end of that stick "I'm sorry, but frankly we don't anticipate _ever_ fixing that because it's purchased IP and the company we bought it from is out of business", "Sorry to here you have a problem with out hardware, but we can't really help you. Perhaps we could find out who wrote the Windows driver, at the HongKong shop we contracted to". And at my own employer at the time, a Very Large IT company, where a product was withdrawn almost as soon as it was released, because the principal architect just got tired of the debugging effort and quit.
Some people have been passing sarcastic remarks and demonising Reiser. While he may have been convicted by the court, that does not make him guilty. There have been miscarriages of justice before and there will be in in future too. Also, by all reports Mrs.Reiser was not exactly Mrs. Goody Two Shoes.
I read Mr.Hans' post at the link given in the article and it seems to be a very polite and decent way of writing, and shows that he is much concerned for the Linux community in general.
This contrasts with lazy bums commenting on how Reiser would have lots of free time on his hand...
Shame on you and fuck of.. you prats...
Makes no difference
ReiserFS might have had his name, but there was a group working on it. Unless it comes out very soon that ALL of them were involved with killing his wife, I can't see it making any difference.
I used ReiserFS a while back before ext3 become stable, but these days it's ext3 for me, partly because of this and partly because it's better known.
ReiserFS just isn't so terribly significant for Linux, and never was. It is just one of the filesystem choices. Few distros offered it as the default at installation time, and only for some time (as noted in the article, even SUSE dropped it years ago).
As for murderer-developed software, you really cannot know. Given the thousands of people involved in Windows development over more than two decades, I think it likely that some of was written or modified by a murderer, convicted or still uncaught.
Work in prison?
What, can't they give him a laptop or something so he can still work on the thing? Instead of being an unproductive piece of flesh rotting at the expense of everyone...
Its not true that MS killed her
I dont know where that conspiracy theory came from....
Almost any system can have a dependency on one designer. The thing about open source is ... its open so if someone feels its usefull then they can at least develop/fix bugs etc.
It plain to see from MS OS development and problems that the closed nature of their stuff makes it very very hard for them to even develop at a crawl.
Good graphic design doesnt make an OS.
... if convicted, he will be sent to an open prison.
@Mike ... missing persons
You are absolutely right, support goes AWOL all the time. I remember spending weeks tracking down a bug in a chip from a very large US semiconductor company and when I finally got a proof for the problem, it turned out the guy who designed that part of the chip had retired.
In the words of the company rep, "we had to go get him off a golf course".
To be fair, in that case they fixed it. But generally I would say commercial pressures are so great nowadays, you are usually supported better with open source than with closed.
Re: Its not true that MS killed her
"""Good graphic design doesnt make an OS."""
Are we talking about the same MS? Since when has their software looked anything other than bland or irritating? Not that I don't completely agree with you, I just feel it'd be a comment better used against OS X.
It would have more chance if it wasn't called ReiserFS
Enthusiasm will be dampened by having a filing system called after a convicted murderer. More seriously Linux already has other journaling file systems which are properly integrated into the kernel and actively supported, e.g. ext3, ext4, jfs, xfs and probably more.
ReiserFS v3 *was* integrated into the kernel but 4 never was. Reiser was criticized for losing interest in v3 when it needed proper maintenance and this concern was why v4 never got integrated. v3 was certainly a very good FS but it lost out to the simpler ext3 for mindshare.
If v4 was still experimental, available only as a patch and no one steps up to support it, it will bitrot and will never get integrated. That would be a shame, but in a way it would be Reiser's own fault. The kernel mailing lists tell the story of his difficulties working with the other kernel developers. If it was in there from the start, the chances are that the work would be picked up.
Maybe someone will branch it, rename it and continue work, but it really depends on the state it is in now. I guess if you want to know the answer, read the lkml archives and see what happens to it.
"While he may have been convicted by the court, that does not make him guilty."
- Yes it does. He's legally guilty, which is why we are allowed to call him a murderer in the Reg comments pages without the moderators getting all jumpy about libel and stuff.
"There have been miscarriages of justice before and there will be in in future too."
- You're right. Although to be fair, even though a body was never found and the evidence was 'only' circumstantial, there was an awful lot of it. Generally, if something looks like a duck, walks like a duck and quacks like a duck ... it's a duck.
"Also, by all reports Mrs.Reiser was not exactly Mrs. Goody Two Shoes."
- Maybe, I don't know, but that doesn't mean she deserved to die. You are sailing perilously close to the 'she was asking for it' argument with this one.
"I read Mr.Hans' post at the link given in the article and it seems to be a very polite and decent way of writing, and shows that he is much concerned for the Linux community in general."
- That's probably true as well, but murderers are people too. They can be nice or nasty like the rest of us. Perhaps he's a great guy most of the time and just topped his wife on an off day.
"Shame on you and fuck of.. you prats..."
- Point taken, but I think you are allowing your personal enthusiasm for the guy or his technology to cloud your thinking.
To Be Fair
To be fair, Ext 4 was allowed into the kernel in a VERY incomplete state shortly after Reiser4 was turned down and is still not production-ready. Namesys of Reiser4 developed it separately, got it to a feature complete state and then submitted it. The response was basically, "Sorry , we couldn't be arsed to read through the code, so its not going in." and "Strip in down to a couple of hundred lines and we'll maybe think again." I think only one kernel developer actually bothered to look at it and make suggestions. It's not surprising that Reiser's response was vitriolic in the extreme.
And yes, I know submitting a stripped down version for inclusion is rightly the prefered method and Reiser should have followed that. The problem is that Ext 4 wasn't stripped down for it's submission, that is basically all there was at that stage. The Linux kernel basically became its development tree (OK I'm exaggerating).
I used to work with MS Word documents written with specialist non-European fonts. Occasionally we would have to email one of these documents, whereupon we would use the save option "embed TrueType font". But for some reason, some of the font sets wouldn't embed, no matter what I tried.
When we finally bought a copy of Adobe Acrobat, it provided the answer - the fonts were locked against embedding. I checked out the TrueType file specification and found that TrueType fonts could indeed be locked against embedding.
Email correspondence with the manufacturer of the fonts followed. They also had no idea that TrueType fonts could be locked against embedding. They also had no idea that any of their fonts had been so locked, and not even the vaguest idea why that might be. They hadn't designed any fonts or employed any font designers in years - they had all been let go shortly after their product lineup was completed. They now had no technical knowledge about the product they were manufacturing, and had no way of helping me.
To their credit, when I said that I could go in with a disk editor and reset the necessary flags, they agreed to this immediately and without reservation.
(Another foible with this font package/vendor: we, and all our branch offices, had to keep a stack of boxed copies of the font set each, because they wouldn't sell us site licences.)
Thank god this is only open source
Can those who reject RFS on the basis he's a killer not separate the man from the product ?
If some scientist invents the cure for cancer and the common cold then, for whatever reason, bumps of his missus are they going to reject his cure ?
I don't see it ethically difficult to accept work done before a murder which is unrelated to the murder, but then I personally wouldn't reject using the benefits of experiments in which people were wrongly killed or suffered in order to gain that knowledge.
As to whether RFS can survive or not without his 'leadership', and how good it is anyway, that's an entirely different matter.
@ Jason Bloomberg: RFS != ReiserFS
RFS is not ReiserFS, it is a network file system, developed as an alternative to NFS.
US legal system
For once, Slashdot had some good commentary on the case. The system provides a reasonable chance for an innocent Hans to go to jail for being a intolerant, elitist nerdy freak dude that weirds out the jury. We wouldn't have heard about it if he was some unknown black guy.
We stopped using ReiserFS ages ago as it was already niche then and ext3 did a bunch of stuff in a far safer way. Try finding recovery tools for ReiserFS. For a long time, the advice has been to ext3 unless you know why you want to use something else.
"Some people have been passing sarcastic remarks and demonising Reiser."
That may be true, but no matter what the filesystem, if you have a head crash it's absolute murder to fix.
THose "won't use a killer's fs" posters
I hope none of you drive Volkswagens.
As others have said: suck out the juice and spit out the pips.
- Vid Hubble 'scope snaps 200,000-ton chunky crumble conundrum
- Updated + vids WHOA: Get a load of Asteroid DX110 JUST MISSING planet EARTH
- 10 years of Facebook Inside Facebook's engineering labs: Hardware heaven, HP hell – PICTURES
- Very fabric of space-time RIPPED apart in latest Hubble pic
- Massive new AIRSHIP to enter commercial service at British dirigible base