On Tue, Oct 23, 2012 at 5:10 AM, Noah Slater <[email protected]> wrote: > On 20 October 2012 22:17, David Nalley <[email protected]> wrote: > >> Following up to this thread: >> >> > Some ways in which the current CloudStack account on GitHub are >> problematic: >> > >> > - CloudStack is used as the account name, in both the URL and the UI >> >> I attempted to solve this with cloudstack-extras - and that thread is >> still ongoing. >> > > I think if we can nail this, the URL is fine. > > >> > - The account advertises itself as the "CloudStack Project" >> >> I changed this to: former home of the CloudStack Project (Now Apache >> CloudStack (incubating) - which obviously has a typo, and doesn't do >> a good job explaining what it is. >> I've changed this a few moments ago to: "Collection of repos that are >> useful for folks using Apache CloudStack" let me know if you see that >> as problematic. >> > > Heh. This looks good to me. (Bit long winded though! ;) > > >> > - There is a repository itself called "CloudStack" >> >> Yes. So here is the problem (and happy to hear creative ways to solve it.) >> The entire CloudStack code base did not move to the ASF. In particular >> we only moved what was, at that time, the most recent version's brand >> and the master branch were moved. (as per the proposal submitted to >> the incubator.) However, CloudStack has about 4 years of history prior >> to that and probably something close to 50 releases. CloudStack, prior >> to April was GPLv3 licensed. So there is a requirement under the GPL >> that we make our source code available for those releases; and quite >> honestly the git repo is the easiest way to do this. > > > How long does this (onerous) requirement apply? Are these active releases? > How long do we have to keep this stuff around? My preference is that we > just wipe it. But there may be good reasons to archive it somewhere instead. > > I just read the GPL, and it says "you remain obligated to ensure that it is > available for as long as needed to satisfy these requirements." Man, the > GPL is a dumb license. So what does this actually mean?
In common usage people assume 3 years because an alternative method of complying (which we aren' eligible for) is to extend a written offer (coupon) at time of distribution for the source valid for three years. CloudStack's last GPL release was in March of 2012 - so theoretically as late as March 2015. There are still people who are heavily using version 2.2 and some of the older 3.0 releases that are GPLv3. > > >> I am open to >> other alternatives to complying with the legal responsibilities of the >> GPL, but I am not aware of any) I have made that repo effectively >> read only (it was acting as a mirror for a Citrix internal repo. >> In short, I do not see a legal way of jettisoning this repo. > > > There has to be, eventually. And if there has to be a way eventually, the > question is when. The GPL *cannot possibly mean* that we have to keep this > repos around in perpetuity. If it meant that, then any software you ever > release under the GPL would mean that for the *rest of your life* you had > to host the source code for it. Yes - in common usage it isn't in perpetuity. Most places see it as 3 years. > > Additionally, in the original releases you made under Citrix, did you > release the source code? Because if you did, then I am not sure this > applies. I think this only applies if you were making binary releases. i.e. > this is basically here to make sure that people who downloaded binary > releases can also download the source code later. Under Citrix - yes, there should have been a corresponding source release. Additionally for Cloud.com - releases from ~February 2011 and later also had corresponding source releases. However, the releases prior to that likely did not (though honestly I don't know.) but I think a function of how the releases were made (yum/apt repos where only the binaries appear to have been made available. I'd go audit this but sourceforge.net (where our pre-ASF releases were housed) appears to be down atm.) > > I think it is reasonable for us (given the wording of the GPL) to remove > the old source code as soon as we consider the non-Apache releases to be > inactive and unsupported.) > > >> But open to alternatives. >> I have added a disclaimer to the description of the repo that says as >> follows: >> DEPRECATED & read-only!! - This repo exists for GPL compliance only. >> CloudStack development has moved to the ASF - see >> http://incubator.apache.org/cloudstack > > > That's great. > > Perhaps the repos name can change to "cloudstack-old", or > "cloudstack-archive" or something? I changed it (and cloudbridge) to cloudstack-archive and cloudbridge-archive respectively. > > >> >> > - There is a link to http://cloudstack.org >> >> I've removed that link (and the link to the ACS project page) I've >> added disclaimer that says 'not affiliated with the Apache CloudStack >> project' >> > > Yep, that's great! > > >> > - The word "CloudStack" is frequently used in repository names >> >> This is a nominative use of the word CloudStack IMO. knife-cloudstack >> is a knife plugin designed to interact with cloudstack. >> puppet-cloudStack is a puppet module designed to deploy CloudStack, >> zabbix-cloudstack is a Zabbix module designed to monitor CloudStack. >> As I said in another email, not every use of the word CloudStack is a >> trademark violation, and indeed I don't know what we would change it >> to if it were. knife-that-IaaS-project-at-the-ASF seems a bit >> pointless. >> > > You're right. I was listing it because it contributed to an *overall* > confusion. With these other things sorted out, I think these repos names > are probably fine. The note above about whether we should consider renaming > the "cloudstack" one is sort of the only one that concerns me at this point > > >> > - There isn't a single mention of Apache anywhere on the page >> >> There is now a mention of apache there, but it disclaims any involvement. >> > > That's great! > > >> What else needs to happen wrt this github account? >> > > Let's wrap up discussion on this thread and then I think we'll be done. > > -- > NS
