Re: Bind v. TinyDNS

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sun, 2004-01-04 at 21:01, Rodolfo J. Paiz wrote:
> Whew. Lots of stuff here, of a perfectly valid philosophical bent, and 
> unfortunately I do not have anywhere near the time I would need in order to 
> participate properly in this discussion. Let me just throw out a couple of 
> thoughts, though...

You actually did pretty well. ;-)

>          1. This is not just a philosophical discussion; to a large extent 
> it is a practical one. A large part of the Fedora developers and a big part 
> of the sponsorship and resources and dedication come from Red Hat, Inc. As 
> they have done for a number of years, they seek to foster the growth and 
> development of Free (when I capitalize it, it means "with freedom" rather 
> than "without cost") software, since Red Hat Linux, Fedora Linux, and Red 
> Hat Enterprise Linux all needed or need to draw in one way or another on 
> that development for future tools. Using proprietary or closed software is, 
> to them, being given a fish rather than learning to fish.

Doesn't Red Hat ship some "closed source" packages with its various
Enterprise offerings? Java comes to mind. I have never run a high-end
version, so I don't know for sure, but I thought they and Sun had an
agreement to redistribute Java in one version or another. Given the
server focus, I'm betting they don't ship Flash or other more
desktop-oriented packages.

>          2. Red Hat (and now Fedora) have put forth huge efforts to help 
> standardize and simplify software for Linux. Software which refuses to 
> allow modification of installation paths, for example, works against that 
> standardization.

Agreed. But if I have to add that package back in again later, I'm in no
worse shape. That is, your argument is a good one for why open source is
to be preferred over closed source in general, but not necessarily one
for why closed source packages should not be included in Fedora.

BTW, I'm using the term "closed source" here to really mean anything
that isn't explicitly open source, whether that is truly closed or some
in-between model like DJB's where you can see the source but your usage
is restricted. It's simply too verbose to write out a phrase for what I
mean and "closed source" is simply the most extreme example of what I'm
getting at here.

>          3. They seek to have patches supplied and integrated quickly and 
> reliably, and rely in part on the Open Source philosophy to put lots of 
> eyeballs in front of every piece of code hoping to make the code both more 
> reliable and more trustworthy. Software which does not permit someone else 
> to submit or integrate patches which result in new, released versions works 
> against that mechanism of support for users.

Agreed, but again, if I have to add the package back in later, I'm in
the same spot, but with more hassle. You are correct that closed source
packages work against independent, 3rd party support. But again, is that
a reason to not include them, or just a reason to prefer open to closed
source where possible?

>          4. Resources are short enough for everyone as it is. When you 
> include a closed-source package on a distro, then you face the prospect of 
> bugs in that package (inevitable in all packages) raising Cain with the 
> functionality and stability of other components of your system, but not 
> being able to debug (and therefore support) them. The result is a higher 
> cost in resources for a lower return on that investment.

Yup. But again, I think this is an argument to prefer open source, not a
reason to keep closed source out.

>          5. In some cases, software makers simply will not allow the 
> redistribution under terms which make Fedora's very "business model" 
> possible. DJB is one example, MySQL 4.x is another... for various reasons, 
> people chose to license or publish their software under philosophical terms 
> which cannot coexist with Fedora's. Others do so under terms which would 
> allow Fedora to include them, but which would not allow Red Hat Enterprise 
> Linux to include them. This is less important but still has a significant 
> impact, given that Fedora does represent the technology and stability 
> feeder for RHEL and thus justifies some of the $X Red Hat spends per month 
> on development, bandwidth, and other resources for Fedora.

I understand this argument and I think there is some merit here. Again,
I don't advocate including something which by its license can't be
included. If the package author doesn't want to cooperate, hey, there is
nothing you can do. You work with what you have and do the best you can.
In some cases, this may be as simple as asking the author for permission
to redistribute under a special license. I'm sure that many authors
would love to be able to gain broader distribution for their work with a
mainline distribution if whatever concerns they had about their work was
protected somehow (don't know what those concerns would be, but
obviously something kept them from releasing the software as pure GPL).
My main objection is against any sort of stubborn philosophy versus what
I would term as pragmatism. I'd like as functional of a distribution as
I can get out of the box.

>          6. Philosophically, even if some Free packages are less mature 
> than their proprietary equivalents, including the Free packages in the 
> distro exposes them to a great deal of use, therefore testing, therefore 
> usually interest, therefore usually development. This helps push those Free 
> packages forward and sometimes (ref ATI) convince manufacturers to 
> open-source their software or drivers for everyone's benefit.

Yes, but the question is whether the distribution will suffer versus
distros which are more pragmatic in their inclusion of packages. If I'm
going to run something like a Java-based web server, I'm not going to be
doing it on something like Kaffe for a while. For quite a while, it
simply wouldn't work (though the Classpath guys are doing quite well; I
would have been trying to help them, but unfortunately I have seen the
guts of Sun's JVM). Further, I'm not suggesting that we choose the
closed source in favor of the open, just that both should be provided
with users making the choice. I'm all for the competition and the
general growth of the open source side. Indeed, I favor it where I can.
But again, I have a pragmatic side, too, and I find that products which
force users to build it themselves are great for hobbyists, but lousy
for end users who just want to get something done. The more you can
include, working right out of the box, the better.

I'm wondering if it wouldn't make sense to include a fourth CD that
would contain a bunch of add-on software that is of non-open source
nature, with Anaconda and the various installers knowing about it. All
that software would be non-supported, non-security patched, etc.
Basically, the installer will make it easy for you to use it, but you're
on your own after that point, possibly with pointers to something like a
foreign yum repository if you want updates there.

> These points do not intend or attempt to be all-inclusive, carved in stone, 
> or anything else. They are merely thoughts which suggest that Fedora's 
> objectives have both philosophical and practical grounds and purpose, and 
> why I find those to be a positive thing.

Fair enough. Thanks for the dialog.

-- 
Dave Roberts <ldave@xxxxxxxxxxxx>




[Index of Archives]     [Current Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]     [Fedora Docs]

  Powered by Linux