Hans Reiser Guilty of First Degree Murder (Wired)

Renaming the project would be unethical without the blessing of Hans 
himself.

In the past, he was pretty particular about it, apparently due to a bad 
experience where someone else took credit for Hans' work.  That has been 
why he has been rather insistent on it being ReiserFS.

That said, I can see at least /some/ people being pragmatic enough to 
realize the problems now attached to the name, and wishing to let the idea 
go and be what it can be.

Still, Hans Reiser has not demonstrated himself to be very pragmatic in 
the past, and as I said, I believe it's his call -- and I think most 
developers in the FLOSS ecosystem would agree -- after all, the 
recognition they get for their code is a good portion of the payment they 
get for it, even for those getting paid some money otherwise.  It's not 
like most couldn't work for a proprietary company and make more money, if 
they chose to, but they don't.  

So I do believe it unethical to change the name without Hans' blessing.  
Maybe he'll give it, maybe he won't, but either way, I believe it's his 
decision.

===

I've wondered, however, if it's now actually more likely the code will 
make it into the mainstream kernel.  Nobody involved can seriously deny 
that Mr Reiser was as much an obstruction as help, likely even more an 
obstruction than help.  Every time he got on the mailing lists, fireworks 
erupted.  Then there's the fact that he got reiserfs (3) in, over the 
objections of some, and then pretty much dumped it, refusing to do the 
normal upkeep to keep it up to date with the current kernel.  Others had 
to pick up where Namesys basically went AWOL on maintenance (there was of 
course a basic development strategy disagreement, but it's the mainline 
kernel, and if you want your stuff in it you play by Linus' and 
lieutenants' rules or you're unlikely to get more in, exactly what 
happened), and that's just not the way you go about impressing folks that 
letting you introduce additional filesystems into the kernel is going to 
work.

It was it seems, more the slow steady work of his cooler headed  employees 
that got Reiser4 as close to inclusion as it came, and I don't believe I'm 
alone in thinking we may well have had Reiser4 in mainline by now if Hans 
Reiser himself had simply stayed out of the mainline kernel merging 
process, realizing the realities of the situation, and delegated that job 
to someone else, someone with a bit less history of having EVERY exchange 
blow up.

With Hans Reiser now pretty much forced out of the picture in terms of 
personal involvement, in some ways it seems only a question of letting 
things settle down a bit now, and seeing if there's anyone left to 
actually carry on (which, so far there has been, thanks to some 
perseverance on the part of former employees).  If so, I believe it's 
likely it'll hit the kernel sooner now than it may have otherwise.  
Certainly, not everyone gets hosting on kernel.org, so that alone 
indicates some belief in the technology and the people continuing to work 
with it thru it all.

You know... if Mr. Reiser is really interested in letting his work be all 
it can be... maybe suggesting it be renamed in honor of his employees 
might be the way to go, or if that doesn't work because of the multitude 
and variety of names, maybe honoring their national or regional origin, 
RussianFS, anyone?  Or EuroRussianFS, or pancontenentalFS, or similar?

The thing is, if it isn't introduced soon, regardless of how technically 
great it is, it may still be dead on arrival, simply due to the march of 
technology.  In a couple years' time it's likely that all traditional 
primarily single hard drive block device filesystems will be dropping in 
popularity, squeezed on the small side by filesystems more appropriate to 
solid state devices (zero seek time, limited writes, larger blocks written 
at once), and on the large side by cluster filesystems appropriate for 
SANs and high availability computing.  I know a few years ago I was really 
eager to see reiser4 in mainline, and am still using reiserfs (which has 
been a lot more stable since the introduction of data=ordered by default), 
on md/RAID (RAID-0, 1, 6, depending on the use of the data).  However, I'm 
now looking at some sort of SSD based RAID for my next generation data 
storage, doing away with spinning hard drives entirely.  Some form of 
logged fs (write and append, never rewrite until one reaches the end of 
the media) is better suited to that than journaling (pretty much any of 
them but certainly reiserfs/3 or ext3/4) or reiser4's dancing trees.

Duncan