On 12/06/18 04:10 AM, Richard Brown wrote:
Chaos is not following.
The implications of what you are saying is that we should be statistically
normal, that we should not be, should not pay attention to, the 'fringe' cases
that 'test' the non-normative conditions.
Normative debates, as opposed to Empirical debates, are debates that require
action on a specific topic and a model to implement the idea put forward in the
topic, and, predominately, can be signified by the use of the word 'should' in
Eventually, Richard, when all the argumentation is done, you are saying the
'Empiricists' have no value. That too is why openSuse is
like ReiserFS, which has proven EXTREMELY reliable and demonstrates the triumph
of good design than the 'cut it an see' iterations that depend on bugzilla
reports and fixes that are needed to make progress.
What else is there in 'modern' file systems?
The fall-back to XFS? Why not JFS?
Well, at least XFS, JFS and BtrFS, like ReiserFS, are proper B-tree systems that
escape from the idiocy of 'preallocation' of the i-node/data division that is
the legacy of the original UNIX file system dating back to V6 days (or was it V5
days; I used V5 for playing around with compiler construction but never looked
at the 'system side'). Yes, ext4 uses b-tree for space allocation but still
the i-node/data preallocation issue. "Backward Compatibility".
The shortcoming of ReiserFS (which is why I've moved to JFS) is that it doesn't
support multi-threading. My multi-core CPU and multiple file systems sow this
up. In a more ideal world one of the distros would be putting effort behind the
development of Reiser4, but it seems the name 'reiser' is tainted. Which is
pretty dumb. It's like refusing to use a Ford F-350 or F450 to pull your
wheel' RV because way back when hank Senior was a rabid anti-semitic jerk.
The strength of Linux was always in its diversity, but a corporate model needs a
higher degree of standardization, normalization and want to focus its 'support'
on a specific (and preferably minimal) set of things. The forking nature of
real FOSS is an anathema to this.
Close alignment with a corporate Linux has advantages, but most of those
advantages favour the corporation. Perhaps the Suse model puts the FOSS and the
corporate too close. certainly when we look next door to the other 'corporate
model' Fedora and RedHat seem a bit more separate and RedHat Enterprise seems to
have abandoned BtrFS
That's RedHat *CORPORATE*.
But let's step back a bit.
Some of the B-School attitudes to 'corporate' rather than FOSS have little to
with functionality, reliability, quality or anything else. There were PCs
before IBM issued its original piece of crap, and some of those were very good,
much more capable, offered more options, hard drives and more. There were even
some from 3rd and 4th tier vendors like Radio Shack. Remember the poster by
Data General: "The little guys say 'Welcome'" ??
This b-school attitude was once explained to me as the 'management' wanted to
deal with a vendor they could sue if things went wrong. If this was a one-man
operation or a small family dealing, they weren't interested. Never mind the
quality. Never mind the functionality. Crap from IBM was 'better' than the
best that Cromemco had to offer. After all, the S-100 was a 'hobbyist bus'.
Yes, in December 1981 Inc. Magazine named Cromemco in the top ten
fastest-growing privately held companies in the U.S. But that article was
written and scheduled before the IBM PC was released in August 12 of the same
year. I don't have a copy of that issue so I can't tell you if there were
reviews of the IBM PC alongside the list that included Cromemco. That Cromemco
was known for its engineering excellence, design creativity, and outstanding
system reliability was beside the point.
Yes, they could be sued, but but IBM had bigger pockets and was more inclined to
just pay off on minor suits.
To the likes of many on this list, "engineering excellence, design creativity,
and outstanding system reliability" matters. But to corporations those are
secondary issues. On the BTDT scale I know that when dealing with a new vendor
or even some customers, I check references, bank accounts, and stuff like that.
Whether thy actually have a better product than the next guy doesn't matter.
The sad thing is that often time Big name Co. might decided to discontinue that
product like for any number of reasons,where as Small Co. HAS to hang on to it's
main product like for its survival.
Corporate needs and private needs are often disparate and when the FOSS gets too
close to the corporation a lot of the corporate attitudes take over.
Yes, some of the things that BtrFS does are nice, but they aren't the only way
of doing things. There are a number of ways I can use LVM and other file
systems to achieve some of the same ends. And I'm sure other file systems can
have the hooks into 'snapper' to do snapshots. But certainly, snapshotting and
handling multiple (even when not RAID) drives, hybrid volumes (such as for
caching on a SSD), thin provisioning (and how that makes snapshots very easy!),
sharing storage space & clustering between machines for HA ... All is pretty
easy. It's not that you can't do such things with BtrFS or ZFS, but it is easy
with LVM and something simple like ReiserFS, or even ext4.
My point is that alternatives, diversification, forking, is natural to FOSS but
not to a corporate model. Having corporate backing for FOSS is great, but don't
get too close.
Q: Are you sure?
> A: Because it reverses the logical flow of conversation.
>> Q: Why is top posting frowned upon?
To unsubscribe, e-mail: opensuse+unsubscribe(a)opensuse.org
To contact the owner, e-mail: opensuse+owner(a)opensuse.org