[Gt-eos] Fwd: Re: gtk source repo location

[Mátyás Selmeci]

On 10/20/17 09:45, Brian Bockelman wrote:
>> On Oct 20, 2017, at 12:46 AM, Frank Scheiner <scheiner at hlrs.de> wrote:
>>
>> Hi Mat, all,
>>
>> On 10/19/2017 08:04 PM, Mátyás Selmeci wrote:
>>> Hi folks!
>>> While this is an important question, we shouldn't take too long in making a decision since we can't get any work started until we've made a repo.
>> I assume OSG won't halt their previous and current work on the Globus Toolkit until we all made a decision.

The big items on OSG's TODO list are:

  * Get the Globus build and test systems working on our infrastructure
    so we can make releases
  * Merge pull requests so EPEL (and in some cases OSG) won't have to
    carry their own patches

(Regarding the first point, I see there is a thread about how to decide 
what goes into a release, but we also need to figure out who will do the 
actual technical work of making a Globus release. I don't want to derail 
this thread with that topic, though.)


>>> I created a Google Form at https://goo.gl/forms/j1RcaoPZUlHt8wnx1 for voting on an organization name and a repo name, incorporating the suggestions I've seen in this email thread. You can add your own responses if you really don't like any of the options, but please use GitHub search to check for name availability (look at the "Users" tab of the search results).
>> On the risk of further delaying this decision, I have a few concerns about the form of the poll:
>>
>> 1. One has to decide for one alternative but cannot select multiple ones. This leaves out consensus on options that are not each one's favorites but close to.
>>
>> 2. But it also looks like one can give a vote as many times as one likes. Which also means that in the extreme I could tilt the result to a name like "Tronc" ([1] ;-)) by just performing as many votes as needed. This cannot be easily detected without further information provided.
>>
>> [1]: https://youtu.be/bq2_wSsDwkQ?t=9m38s
>>
> We're not voting for president, we're gauging interest for the color of a bikeshed!
>
> If we can't trust each other to behave reasonably on a Google poll, we will really hard time making releases...
>
>> 3. One does not have to leave a name or affiliation, so in the end nobody can determine who has voted, who has voted multiple times or who has not voted at all. I mean this is not a vote for a government where a ballot is needed to avoid possible sanctions. Giving name and affiliation would allow for a greater assurance and confidence in the results.
>>
>> 4. Is it one vote per person on the list or one vote per institution/project/etc.? This is more of a regulation we should decide on.
>>
> Let's do it per person...
>
> This is simply a way to gauge interest, not a legally binding vote.  At least from my point of view, I have preferences I like best but see no option so abhorrent that I'd walk away if it won.
>
> If things close and no one likes the results, we can try again.

Also, if we do it per person, we won't have to decide if Brian 
represents UNL, CMS, or OSG :)

-Mat
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.egi.eu/pipermail/gt-eos/attachments/20171020/25e0331d/attachment.html>