LWN: Comments on "Stallman: Disk, I/O issues delay GNU OS (InfoWorld)"

Re: Why FSF and RMS shouldn't develop software

IkeTo — Fri, 15 Nov 2002 08:56:28 +0000

> It's not nice to troll, but I have to agree. The Hurd isn't
> going to replace Linux, or even be competitive with the BSDs.
> Even after 1.0, I doubt that anyone (read: most people. I'm sure
> it will have a small loyal following (ignorant of it's
> problems)) is going to trust it to be _stable_, let alone
> scalable and feature complete.

It's not nice to troll, but I have to agree. Linux isn't going to replace Unix, or even be competitive with DOS. Even after 1.0, I doubt that anyone (read: most people. I'm sure it will have a small loyal following (ignorant of it's problems)) is going to trust it to be _stable_, let alone scalable and feature complete.

Come on. You are talking about an operating system, not about a small application. Even with applications like Gimp it takes quite a while to get people begin trusting it. Nobody is going to trust an entire platform on its first public appearance. On the other hand, if you go look at its design, it is something that developers and users alike would love to run, even if it gives up a little bit of horse-power of the OS to the flexibility. I'm really looking forward for a day when popularity of Hurd can at least match that of Linux, or at least match that of Mac.

The Hurd effort will continue

coriordan — Thu, 14 Nov 2002 03:52:08 +0000

"He who says it cannot be done should not interrupt he who is doing it"
chinese proverb (supposedly). Posted to kerneltrap.org during a similar
discussion.

Negative opinions about the Hurd are easy to form. Any user who has done
zero research can present such an opion. This takes no effort as the Hurd
developers are the first ones to point out it's short comings.

Actually reading and learning about the Hurd and it's architecture takes
a bit of time but I assure you it is well worth it. The design is brilliant.
Better than Linux, in theory at least.

Ten thousand irrelevant people saying "it will never be released/stable" is
not going to stop the developers from trying. I'll use it, and hopefully
help develope it some day.

Ciaran O'Riordan

Stallman: Disk, I/O issues delay GNU OS (InfoWorld)

Peter — Fri, 08 Nov 2002 21:41:02 +0000

Debian is based on the (GNU/)linux kernel. Not GNU/Hurd.

Really? Then what is this?

I don't agree with Debian on the name thing, though. As Ben Reiter once said on slashdot, "I call it GNU/Linux. Except the GNU/ is silent."

Why FSF and RMS shouldn't develop software

rsidd — Fri, 08 Nov 2002 14:41:58 +0000

Not that I'm a fan of emacs, or evem vi(m), but these, along with gcc, were critical...

Actually, vi (and vim) have nothing to do with the FSF and aren't even GPL'd. vim is under a "charityware" license, and the other widespread free vi clone, nvi, is under the BSD license.

Yes, RMS had a lot of vision and his compiler toolchain is one of the great gifts to the software world, used not only by linux but also by the BSDs and even NeXT and Apple. Unfortunately, he tends to take credit for everything else too -- witness the whole GNU/Linux thing -- effectively minimizing the contribution to the system of other people like Berkeley (BSD), MIT (X), Don Knuth (TeX) and so on. Obviously he's been successful, as your comment shows.

Stallman continues to be important as a provider of a vision for the community, but the FSF's software contribution seems to be a thing of the past. He's entitled to talk about GNU/Hurd, but until it finally works, it's not news. Covering it on sites like lwn only provides ammunition to the RMS-bashers, who already have enough reason to bash him thanks to GNU/Linux and other sillinesses...

Why FSF and RMS shouldn't develop software

tres — Fri, 08 Nov 2002 09:27:14 +0000

I agree that GCC took a little while to get out the door and that Linux will be the kernel of choice in Free software systems (at least for the foreseeable future) but you should really inventory the software on your system. There is NO WAY that Linux would be what it is today if the tools for its development didn't exist when Linus started in 1991. Further, Linux would be stuck in the TOY phase today if it even survived withour the tools that we all consider 'part of the system'. Not that I'm a fan of emacs, or evem vi(m), but these, along with gcc, were critical to the development of Linux, Gnome, XFree86, etc.. Not to mention the truly 'part of the system' tools such as ls, sh, csh, bash, mv, awk, sed, etc., etc., etc.. RMS, and the Free Software Foundation, that he created are owed a great deal of thanks. I 'almost' want to thank the printer company that refused to provide Richard with the source code for their driver. It was that failure that started RMS down the path to free software; and we have all benefited from that!

Regards,
Tres

Stallman: Disk, I/O issues delay GNU OS (InfoWorld)

beejaybee — Fri, 08 Nov 2002 09:03:40 +0000

Debian is based on the (GNU/)linux kernel. Not GNU/Hurd.

Stallman: Disk, I/O issues delay GNU OS (InfoWorld)

alext — Thu, 07 Nov 2002 23:08:05 +0000

Choice is good. Besides whose going to suggest to the BSD people (or conversly the Linux people) that as the other exists they might as well give up?

Stallman: Disk, I/O issues delay GNU OS (InfoWorld)

loening — Thu, 07 Nov 2002 22:44:02 +0000

Wait a second, I thought Debian was on, like version 3.0 or something? What's with version 1.0 being delayed? I thought it got released in 1995....

;-)

Re: Why FSF and RMS shouldn't develop software

nero — Thu, 07 Nov 2002 19:44:41 +0000

It's not nice to troll, but I have to agree. The Hurd isn't going to replace Linux, or even be competitive with the BSDs. Even after 1.0, I doubt that anyone (read: most people. I'm sure it will have a small loyal following (ignorant of it's problems)) is going to trust it to be _stable_, let alone scalable and feature complete.

Why FSF and RMS shouldn't develop software

lseubert — Thu, 07 Nov 2002 19:01:12 +0000

Here's the deal. GNU/HURD can not handle disk drive partitions larger than two gigabytes. It can't deal with high speed serial I/O either, but that's OK, HURD is switching from the Mach microkernel to the OSKit microkernel, which can support high speed serial I/O. Except not yet. That feature hasn't been rolled into the OSKit microkernel. And it doesn't matter anyway, because HURD still doesn't support virtual consoles which is a feature required by OSKit.

Hello RMS? It is now the 21st Century! HURD has been under development for some ten years now; is nowhere near a viable 1.0 release; and is stuck with 1980's era technology. Bear all this in mind dear reader as you also ponder the long delayed fiasco that was the gcc3.x compiler; FSF's refusal to roll improvements back into GRUB; and any number of other long delayed, buggy, and change resistant FSF software projects.

The FSF and RMS should get out of the software development business and turn their programming work over to competent and professional software engineers and systems designers.

The FSF should turn to a much more vital and important role - something that it does very well right now - defending the GPL and the ideals of software libre. Let the FLOSS community take care of the coding - the FSF needs to promote free software, and fight against patents and other liberty destroying IP laws.