On Mon, May 3, 2010 at 1:23 AM, A Hettinger <[email protected]> wrote:
> (clippage) > > "We have just heard these days that a staging server is down and that's why > b138 isn't available. We also heard that it can't be made available on an > external server because of legal reasons. How can an open source project > work smoothly when a development build can't be made available easily ? It > seems hard enough to get all the technical pieces together and it still has > to go through a legal bottleneck to release an experimental build ? This > shows that there are bureaucratic obstacles holding the project back." > > you seem upset that there are legal reasons for some things. We have to > follow the law, and I'm sorry if that offends you. Until all code which > others have a claim to is out, there is going to be hangups. The > emancipation project is working on this, but there is a large number of > utilities to work through, especially considering the need to maintain > standards compliance (something else most linux distros don't bother with). > I don't see how you can assume I'm offended by the fact that we have to follow the law. I'm not and I do think we have to follow the law. I never said the contrary or anything even remotely near that. All I'm saying is that the OpenSolaris project has bureaucratic and legal obstacles that hold it back when compared to other open source projects. > > (clipped) > > "IMHO, the whole "codebase" vs. "binary distribution" talk is a silly and > stupid idea. There should be an OS called OpenSolaris and that's it. People > would participate in creating the best OS they had in their minds, > committing code and seeing the results. The way things are now it looks like > what OpenSolaris is is a "consortium" of some kind. I remember reading an > article from 2006/2007 where someone from Sun said how great OpenSolaris was > and mentioned all these companeis that were oficially partnering up with > them to develop code. Right... that could happen without it being open > source in the first place." > > We in many ways ARE a consortium. There are a number of different agendas > to get OS optimized to perform different tasks. I have two different ones > myself (both my personal one, and my employer's). Personally I'm interested > in desktop/workstation usage, my employer is interested in only storage > servers. > I'm not talking about conflicting priorities. We are (or pretend to be) an open source project and that means anyone should be encouraged to contribute regardless of commercial relationships. By consortium I mean a group of companies sharing a common interest and contributing between themselves in a closed fashion. You don't need open source to make it happen. A few weeks ago there were some talks from Intel developers saying they were committing code to opensolaris.org that should remain private until the sorted out the legal issues. alanc@ reminded them that opensolaris.orgshould not be used for private collaboration between Sun and Intel and they should take that it somewhere else until they are able to make the code public. > > The OSOL the code-base vs. OSOL the OS is a historical anomaly. "Open > Solaris" was already in use for the codebase, but some in Sun felt the need > to reuse it for a binary distribution as well. It probably should have been > branded something else, like "Solaris-Next" or some-such, and there was > quite a bit of argument from the community over it. I believe the archives > still remain, if you are interested. In any case those of us who started > before "Open Solaris" was a distribution are not going to change our > language because of some historical revisionists, besides we don't have > another word to use anyway. > As much as I think Oracle's "OpenSolaris" distribution has an inherent advantage since the average user expect the OS to be called "OpenSolaris" and will thus consider this distribution the main one (which I think it should)... I was NOT arguing about naming issues. What I'm saying is that I find it less than optimal to have have a main code base and encourage people to create distributions on top of it. IMHO encouraging people to split work when we don't have a big developer base is wasteful. If you read that paragraph again I was arguing that we as a project should have the end goal of delivering a complete operating system, not a "code base" by itself. That means you drop the "distribution" nomenclature and start delivering an "OS". What I'm try to say in practical terms is that I find it more likely that a model like FreeBSD's complete OS will work better than trying to emulate Linux and its distributions. Besides, I don't think the current model has worked so far. What do the other distributions have that is so different from the one released by Project Indiana. To me they look more like a different collection of packages than anything else. That would easily be accomplished by providing Indiana's respin just like Fedora does. There was another thread in discuss@ where I asked about the distinction between forks and distributions and what they were contributing back to the main "code base". > > "AFAIK, FreeBSD releases what they call "snapshot" ISOs every month as > clockwork. Fedora let's you download version 13-Alpha any time you feel like > doing so and they even encourage people to put a silly countdown clock on > their blogs to tell the world when a new release is coming. So does Ubuntu. > We can't even be sure when b138 will be out ? And more importantly, we can't > even know the technical issues behind 2010.x delay so we can help fixing > them ? People have asked for that (about show stoppers) and the answer was > "there is someone inside that makes that call, it's not discussed publicly > though". That's not very welcoming to outside contributions." > > Very rarely are the /dev releases delayed significantly (which is the > equivalent to what you describe). The reason I have been given for them > being posponed are quite reasonable. > There is no problem at all in things getting delayed. The problem is when external help is offered we have legal/bureaucratic barriers that prevent cooperation. > > The delay of /release is unfortunate, but expected in the case of some > types of bugs. Ideally there would be no bugs, but some of us have to live > in the real world. The reality is delaying /release to deal with a major bug > is far better then the alternative. > Again, no problem in taking the time to work on bugs and released a polished product. If you read that paragraph again I was talking about the show stoppers not being easily communicated and the community being kept in the dark. IMHO, if there was a public list of show stoppers people would be encouraged to try to replicate them, provide more details and even fixes. >From my original email I think you can understand that I'm advocating that there should be an "OS" called "OpenSolaris" and this project should do all it can to encourage external contribution. There have been many complains that Oracle's "OpenSolaris" binary distribution is not really an open source project but their product. This implies we shouldn't bother being part of it. As an open source project you should be welcoming contribution. > > "In the end this all boils down to: does Oracle/Sun view the Fedora/RHEL > model as something worth it ? We don't know and we shouldn't be the ones > begging for answers... they should make their plans clear." > > From their press releases and announcements, it seems clear enough what the > general intent is. If they presented exact timetables for everything there > would be the same number of people complaining when things where missed. > I disagree. There is nothing clear at all. > > "People have said it is very hard to open source a huge project like > OpenSolaris and I agree. I just get the feeling Sun came late to the party > and Oracle has no intention of hurrying up. In fact I think in a few months > we are likely to see a "open core" model slowly taking place. > > I'm probably going to be flamed for even mentioning this 3-word company > here but I was reading an article some weeks ago about a few things IBM > learned while doing Linux kernel development [1] and the following paragraph > made me think of OpenSolaris: > > "We spent far too much time behind the IBM firewall, discussing things, and > we tried to polish our external communications" Frye said. "So we banned > internal IBM communication on the Linux kernel. Anyone working on the kernel > at IBM was not allowed to talk to anyone else inside the company. All > communications had to be external."" > > That's vary easy for IBM's linux contributions, because they aren't talking > about things that others have claims to. > You seem to be missing my point again. In that paragraph I was talking about decreasing internal communication and making it more external to get the community more involved. I don't see where I was talking about anything being easy or that other have claims. > > "Like I said in another emails, people shouldn't be trying to shut down > these talks while saying "wait and see." Despite occasional trolls, most > people are complaining because they care about OpenSolaris and not because > they would like to see it die. It's going to be a sad month when nobody ever > asks when 2010.x will be released." > > And what is it accomplishing? Yes, they care, fine, great, wonderful even, > but the complaining just adds to the "noise" part of the sig/noise ratio. > It's not helping, It's not going to change anything, and at this point > nothing is even being added. We have all already heard it. > <sigh> > > "And If I'm getting everything wrong... I would be very happy to hear > *officially* from Oracle what they envision for OpenSolaris in details. That > way we can all adjust our expectations and consciously conform to the model > being proposed or go away." > > It would be nice, but (and this has been said before) Oracle does not tend > to announce things until they are ready to go. The up side of this is less > vapor-ware. The downside is useless speculation and FUD. > <sigh> That can probably be how you run a traditional closed project but not how you run an open source project. And I totally disagree with you here: when there isn't a clear communication between the main (only?) sponsor and the community what you get is exactly speculation and FUD. > > (clippage, if you are really interested read gtirloni's full messge) > Yes, anyone really interested please read my original email before coming to similar conclusions. -- Giovanni
_______________________________________________ opensolaris-discuss mailing list [email protected]
