This week's FLOSS Weekly interview is about ClearOS (audio/video at
http://twit.tv/floss168). Apparently they have taken the CentOS
developer's frequently given advice to go away and do it yourself and
will have a 'ClearOS core' release that is their own rebuild from Red
Hat sources that will be the base for the ClearOS enterprise
distribution instead of relying on CentOS as they have before.
--
Les Mikesell
lesmikesell at gmail.com
>> This week's FLOSS Weekly interview is about ClearOS (audio/video at > http://twit.tv/floss168). Apparently they have taken the CentOS > developer's frequently given advice to go away and do it yourself and > will have a 'ClearOS core' release that is their own rebuild from Red > Hat sources that will be the base for the ClearOS enterprise > distribution instead of relying on CentOS as they have before.Thanks for the link. It makes interesting listening because there are claims that they tried to engage with the CentOS devs to offer support and resourcing, but that relationship was not forthcoming... so they intend to build (as I see it) a direct competitor distribution (i.e. "binary compatible"). Also interestingly, apparently they have recruited help from the SME/Contribs people, so I don't know if that means SME will die because it had precious little resources to start with (and now those resources work for the competition) or SME will still carry on and be rebased on Clear Core. Also stated in the audio is that this was all a direct response to the uncertainties around CentOS. So, it is interesting to see how this is all going to pan out.
On Friday, June 03, 2011 09:06:28 AM Les Mikesell wrote:> Yes, RedHat deserves the credit for denying access to the binaries of open > source work, even to the community responsible for it even existing.[snip]> But when you say that, keep in mind that the 'original packages' part is the > packaging work, not the creation of the vast majority of the code. And that the > Red Hat company made its name and developed its community of users by allowing > free access in the first place up until the EL/Fedora split. Personally I think > everyone who uses free versions would have been better off if they had switched > to Debian the day that Red Hat put the restrictions on redistribution, but I was > too lazy to learn the options to 'apt-get'.Red Hat deserves credit for still provided the source RPM's in buildable form even for those parts of the distribution that are not GPL licensed. They are not required by license to do that; for instance, the PostgreSQL RPM's, since PostgreSQL is BSD-licensed. I mention that particular package only because I have first-hand knowledge of that package. Red Hat deserves credit for providing vast amounts of developer time to the upstream projects, including but not limited to the kernel, glibc, gcc, GNOME, PostgreSQL, and RPM itself. Red Hat is not the only Linux provider who has limited distribution of binaries. And as the CentOS and other rebuild projects have proven time and time again, having the source (and some time and significant effort) is sufficient to build a fully binary compatible distribution. To my eyes it was a win-win for Linux, since without the for-profit model that Red Hat adopted, Red Hat likely would not be around today, nor would Red Hat-funded developers likely have been able to continue to devote as much time and effort as they have done. Perhaps they could have handled the PR in a better way, but then again when someone is used to freeloading they're going to hate having to pay anything at all (and that's not an accusation of anyone in particular, just a simple observation of human behavior). The CentOS developers/rebuilders are to be commended for taking on the significantly difficult task of not just taking at rebuilding the system, but taking on the much more difficult task of making the resulting rebuild 100% ld-level and dependency-level binary compatible, as least as much as is possible with the released source code to the distributed binaries. Not to mention the far more difficult task of then releasing it publicly and dealing with that.... But, I do understand and am sympathetic; I miss the old boxed sets as much as anyone.
On Thursday, June 02, 2011 08:03:34 PM Rob Kampen wrote:> My look at the website shows only i386 versions - this is a long way > away from a replacement or alternative to CentOS.Also, it likely would be a subset, and not the full distribution. This has already been done, and released, as FrameOS 6, back in February. But it is a relatively small subset. Not to trivialize the effort that was taken, however.
On Friday, June 03, 2011 11:21:35 AM Les Mikesell wrote:> I'm talking about what would be > more in the best interest of the community that they attracted by > permitting redistribution of the collated works - and then cut off.It's in the best interest of the community to have Red Hat in a financially stable position to fund all this good stuff in the first place. If the only way Red Hat can be financially viable is for me to give up the pre-EL ways, then that's fine by me, especially since Red Hat is rather accommodating in terms of the source code.> Go back and look at the changelogs of programs in the era between the > RH 4.x and 9 releases if you don't remember how bad the stuff they initially > shipped was or how it got fixed.Got back and look at the changelogs of the PostgreSQL packages.> More to the point, wasn't that the reason you started using Red Hat in > the first place?No. I bought Red Hat 4 because it was the only non-proprietary platform on which I could run RealAudio Server back in 1997 and expect to get support from Progressive Networks. Red Hat being a North Carolina company was a great bonus. And while I would like to be an idealist, at the same time I know without doubt that I, and many others, use CentOS precisely because the binary compatibility for running closed source software is so good.
Marko Vojinovic wrote:> On Friday 03 June 2011 16:21:35 Les Mikesell wrote: >> On 6/3/2011 8:57 AM, Lamar Owen wrote:<snip>> So what? Red Hat created a community by beeing free in both senses, andthen decided to go commercial at some point. And that hurt the feelings of some minor number of hard-nosed community members. Is that what you are talking about?> > I was around at the time of Red Hat going commercial. I heard about that,<snip> I'm having some problems with the way the conversation is going. RedHat *was* a company; to me, the RHEL was aimed as a wedge, to get into corporate America. For that matter, who started offering their distro of RHEL around then? Why, the same company that offered this new o/s on their brand new product, the IBM PC in 1980: IBM. RedHat, at least, has not taken the path to the Dark Side, as the Other Company did.... mark
On Friday, June 03, 2011 03:49:00 PM Les Mikesell wrote:> On 6/3/2011 1:28 PM, Lamar Owen wrote: > > Go back and look at the changelogs of the PostgreSQL packages.> Give me a hint about what to look for. As I recall I always installed > postgresql from source in those days because the disto packages were so > far behind or broken.So, did you provide community-based feedback to the then PostgreSQL RPM packager? Any bugzilla entries? Any e-mails? Anything? Sounds like the packager at the time could have used some good feedback, instead of you bailing out, installing from source. And this is the Community in CentOS; as you have defined it here in this thread, Les. The users, not the developers; the ones who provide good feedback, but don't necessarily build (develop) the system. Your definition was: "[The community is] not the development community that pushes wild and crazy changes into fedora that I'm talking about." (antecedent of your 'it's' in the original is in brackets). This same community is here, and it's vibrant. I see many of the same names I've seen for over ten years. Doing essentially the same thing, and giving feedback if they're not actively developing or packaging. Some are a tad more crotchety than before, but it's a familiar community. Oh, I almost forgot to mention: I *was* the community packager at the time. And I could have used more useful, constructive, non-trollish feedback at the time. Like I got from Sander Steffann, Kaj Niemi, Alvaro Herrera and the tireless developer to whom I handed the packager role, Devrim Gunduz, who is doing outstanding work in that role even today. A vibrant developer community, one I miss, to tell you the truth. The rh.com contact/packager changed a few times, but I was the community packager from 6.1 or so through a good part of FC2's development. Log in to a CentOS 4 machine that has postgresql installed from CentOS-Base repo, and issue a 'rpm -q --changelog postgresql' and scroll up a couple of dozen lines or so from the end (date tagged Fri Nov 21 2003). The PostgreSQL core developer Tom Lane took the Red Hat internal reins, and is still there (employed by Red Hat and in the PostgreSQL Core Team). Tom does outstanding work. PostgreSQL, just to name one project, is very much helped by Red Hat, in upstream Core roles.