>
I'd love to see some ABI standards. A standard driver model.
That has been done. Two problems: the driver interface to the (micro)kernel would soon become inefficient as the kernel evolved, and it would make it even easier to ship proprietary drivers for free software.
>
Subsystem support, like NTs kernel has
That's one of the Hurd ideas: migrate to Lisp by sporting both a POSIX and a Lisp personalities.
>
things like POSIX support are needed in the former (and, as I mentioned, adding and enhancing it is an ongoing process for VSTa) -- but for a microkernel, these can all be userland changes, and quite confined ones at that
Granted, but when one compared the Hurd (and Linux for that matter) with VSTa or QNX, one is comparing also POSIX, drivers, scalability parts... just not a fair comparision.
Now if QNX or whatever was free software and had a roadmap to take on Linux and Hurd goals, that would be very interesting... only that they would face the same problem as the Hurd, to wit lack of critical mass.
>
Supposing that you're right about our culture,
and you may well be, shouldn't you account for
that in your design?
Well, it's a decadent culture. Designing for it is counter-productive in the long run. Ultimately customers should ask for something better, and other than that free software provides a way out in a limited scope.
>
Linus thinks he can keep advancing the 2.6.x kernels without causing stability problems
He's wrong, Linux 2.6 isn't stable enough, and worse it shows no signs of dealing with reported issues like the data corruption problems with the LVM2 + RAID5 + ext3 stack.
>
Our manager had been one of the original Mach developers. HURD still uses Mach BTW.
They are moving away from it. It was a proof of concept, and only recently the Hurd got enough attention to get to a more definitive microkernel.
Just like with relational databases, it is cheap to get it wrong, and expensive to get it right, so in our short-term culture getting it right gets forever sidelined. It doesn't matter how much resources you have, it needs time to mature, and once it matures the benefits are huge.
We have enough system resources to pay for the cost of glue. Yes, the initial implementation is more complex, but further modular development gets simpler.
>
Compared to its peers (QNX, VSTa), the HURD is tremendously bloated and slow
Problem is, these are not peers. They do not aim for full POSIX compatibility, they do not aim for multiple personalities nor for the level of development flexibility of the Hurd. They may fit on floppies, but so does GNU/Linux if you trim it enough.
It has been a time since I've started to consider Linux too complex. 2.4 also took quite some time to stabilise, and 2.6 still isn't production ready in some quite relevant situations.
For example, trying 2.6 + LVM2 + soft RAID5 + ext3 is asking for data loss. I and several other people reported this, but seemingly either we were statistical noise or we weren't well connected enough, and the kernel hackers never paid attention up to at least 2.6.5 - I just gave up following up nothing at all since 2.6.6.
I know it has taken some bad decisions and now lacks critical mass, but perhaps the Hurd is the way to go... it should enable better isolation of disruptive development, and enable kernel development to continue adding features.
In fact the Hurd was conceived between a bridge between ad-hoc POSIX and cleaner Lisp systems.
Better leave it
on
Game with God
·
· Score: 0, Flamebait
Assuming there is an intersection of game producers and sci-fi fans, better leave it like it is.
Sci-fi usually has lousy depictions of religion, except when made by deeply religious people, and even then...
For example, Anglican CS Lewis' Space Trilogy - _Out of the Silent Planet_, _Perelandra_ (or _Voyage to Venus_) and _That Hideous Strength_ - are great, even if the last one is usually considered over the top. So is _A Canticle for Leibowitz_, by a Romanist whose name I forget.
Even if it never rises from commonplaces, Mormon Orson Scott Card's Ender books are also fair. Perhaps its picture of Romanism is too wishy-washy, either from his sympathy for Romanist Brazil or from fear of being seen as a Romanist-bashing Mormon, but it is not so bad.
Other than that, in otherwise well-regarded works like _Duna_ or Asimov's _Foundation_, religion is just some hierarchical, sacerdotal clone of Romanism.
Not to mention the worse I've ever seen, Arthur C Clarke's _2001: A Space Odyssey_. The movie is useless, the book actually finishes, and it finishes in a kind of mystical scientificism that had been ridiculed by CS Lewis 30 years before, in _Out of the Silent Planet_.
>
I havent had an unintentional reboot since I started using Win XP
OK, but how often do you intentianally reboot? Most MS Windows users I know just shut down every evening, and boot at the morning.
Meanwhile if you are conservative with GNU/Linux, say sticking to some version of the kernel and not caring about the latest and greatest versions of everything, you can run for months - or years, if you have good, redundant hardware with uninterruptible power supplies.
Even if you run something risky like Debian testing you can run for weeks at a time.
SmartFolders? Why do companies have to reinvent nomeclature so often? Couldn't they have stuck to vFolders for virtual folders, which is much more descriptive anyway?
MySQL forces one to do much coding due to its incompleteness and misimplementation of the SQL standard. PostgreSQL should save some, potentially lots of, coding.
A truly relational system should be even better, like Alphora Dataphor, but this is not free.
>
As much as I applaud Gnumeric for their great implementation, it's still a Linux/Unix only implementation.
No, it is a Gnome implementation. That means it can be relatively easily ported to MS Windows (Red Hat Cygwin) or Mac OS X. There were some Gnome ports around, don't know if they live still.
Slashdot Mirror
That has been done. Two problems: the driver interface to the (micro)kernel would soon become inefficient as the kernel evolved, and it would make it even easier to ship proprietary drivers for free software.
That's one of the Hurd ideas: migrate to Lisp by sporting both a POSIX and a Lisp personalities.
Granted, but when one compared the Hurd (and Linux for that matter) with VSTa or QNX, one is comparing also POSIX, drivers, scalability parts... just not a fair comparision.
Now if QNX or whatever was free software and had a roadmap to take on Linux and Hurd goals, that would be very interesting... only that they would face the same problem as the Hurd, to wit lack of critical mass.
Granted, but that's the point: they have different goals, therefore the comparision was useless.
Well, it's a decadent culture. Designing for it is counter-productive in the long run. Ultimately customers should ask for something better, and other than that free software provides a way out in a limited scope.
He's wrong, Linux 2.6 isn't stable enough, and worse it shows no signs of dealing with reported issues like the data corruption problems with the LVM2 + RAID5 + ext3 stack.
They are moving away from it. It was a proof of concept, and only recently the Hurd got enough attention to get to a more definitive microkernel.
Just like with relational databases, it is cheap to get it wrong, and expensive to get it right, so in our short-term culture getting it right gets forever sidelined. It doesn't matter how much resources you have, it needs time to mature, and once it matures the benefits are huge.
No, but once it is in place the development of modules become much easier and less riskier.
The problem with Linux is that it is becoming more and more difficult of avoiding breakage.
So what? If your team took the wrong approaches, it does not invalidate the concept.
We have enough system resources to pay for the cost of glue. Yes, the initial implementation is more complex, but further modular development gets simpler.
Problem is, these are not peers. They do not aim for full POSIX compatibility, they do not aim for multiple personalities nor for the level of development flexibility of the Hurd. They may fit on floppies, but so does GNU/Linux if you trim it enough.
You know just enough to be dangerous
It has been a time since I've started to consider Linux too complex. 2.4 also took quite some time to stabilise, and 2.6 still isn't production ready in some quite relevant situations.
For example, trying 2.6 + LVM2 + soft RAID5 + ext3 is asking for data loss. I and several other people reported this, but seemingly either we were statistical noise or we weren't well connected enough, and the kernel hackers never paid attention up to at least 2.6.5 - I just gave up following up nothing at all since 2.6.6.
I know it has taken some bad decisions and now lacks critical mass, but perhaps the Hurd is the way to go... it should enable better isolation of disruptive development, and enable kernel development to continue adding features.
In fact the Hurd was conceived between a bridge between ad-hoc POSIX and cleaner Lisp systems.
Assuming there is an intersection of game producers and sci-fi fans, better leave it like it is.
Sci-fi usually has lousy depictions of religion, except when made by deeply religious people, and even then...
For example, Anglican CS Lewis' Space Trilogy - _Out of the Silent Planet_, _Perelandra_ (or _Voyage to Venus_) and _That Hideous Strength_ - are great, even if the last one is usually considered over the top. So is _A Canticle for Leibowitz_, by a Romanist whose name I forget.
Even if it never rises from commonplaces, Mormon Orson Scott Card's Ender books are also fair. Perhaps its picture of Romanism is too wishy-washy, either from his sympathy for Romanist Brazil or from fear of being seen as a Romanist-bashing Mormon, but it is not so bad.
Other than that, in otherwise well-regarded works like _Duna_ or Asimov's _Foundation_, religion is just some hierarchical, sacerdotal clone of Romanism.
Not to mention the worse I've ever seen, Arthur C Clarke's _2001: A Space Odyssey_. The movie is useless, the book actually finishes, and it finishes in a kind of mystical scientificism that had been ridiculed by CS Lewis 30 years before, in _Out of the Silent Planet_.
OK, but how often do you intentianally reboot? Most MS Windows users I know just shut down every evening, and boot at the morning.
Meanwhile if you are conservative with GNU/Linux, say sticking to some version of the kernel and not caring about the latest and greatest versions of everything, you can run for months - or years, if you have good, redundant hardware with uninterruptible power supplies.
Even if you run something risky like Debian testing you can run for weeks at a time.
Performance is all very good, but I'm a professional and what I care most about is reliability.
That's why we have menus of commands. Nothing smart here.
Quite different, once locked a smart bomb will self-guide and follow.
Then I missed something. They look just like Ximian Evolution's vFolders to me.
Smart is not descritive at all. Folders aren't smart, sorry... no AI yet.
SmartFolders? Why do companies have to reinvent nomeclature so often? Couldn't they have stuck to vFolders for virtual folders, which is much more descriptive anyway?
Which is a nice way to get out of the rat race. I fail to see a problem here.
MySQL forces one to do much coding due to its incompleteness and misimplementation of the SQL standard. PostgreSQL should save some, potentially lots of, coding.
A truly relational system should be even better, like Alphora Dataphor, but this is not free.
Actually everyone would benefit if MS was included, but it probably wouldn't participate anyway.
This is a contradiction in terms.
There's already Gnome 2 for Cygwin, but Gnumeric wasn't ported (or at least packaged) yet.
It would be nice if Debian MS W32 went forward, so that we'd have a better installer than Cygwin's.
No, it is a Gnome implementation. That means it can be relatively easily ported to MS Windows (Red Hat Cygwin) or Mac OS X. There were some Gnome ports around, don't know if they live still.
Obviously. My point is that you shouldn't even need that code, it should be all in the RDBMS already. SQL robs us of that.