Notes from Paul Dekkers:
All NRENs doing sw dev, lot is open sourced, finding them and maintaining them over time seems hard.
Need better maintained software over time.
Terena conservancy, github account? support? should Terena do/provide something?
Filesender, eating away time - not talented in legal or financial issues
Been fine so far; always been a challenge where stuff goes next year
Hard to (> 3) get people together, and have NRENs also sustain projects
filesender, cat, tcs portal, radsecproxy, simplesamlphp, pysaml, pear, core, eduroam monitoring probes, ...
We make stuff that our nren/community needs, just not usable for others
Building in the open from day 1
Lot of projects with good stuff, manager that has to approve
Default should be open repo on github
Should be made easier?
Also the documenting what's being done
Requirements gathering
Licensing template for our community software, or comply or explain
Practical and managers
Are we reinventing the wheel for software to be open? Apache foundation, …
And licensing, no case for making a new one, but we should have a common approach. Not the problem? The one paying decides the license. Good to list what licences are nice to pick (advantages, disadvantages).
GNU, BSD, Apache? No, much more: geant (bsd) license, ...
Use previously built software? (Was a problem in openconext: not all tools allow distribution.)
freshmeat.net, git, we don't have resources to maintain it?
not just build sw for federation purposes, but help me make this and that?
domestication wiki, apps list
knowledge base, keeping it relevant and not a software graveyard
github is a platform that we can use, we need discovery (besides), for our communuty (idea landscape), and funding
funding as a problem
stopped projects, where final (open) release stage never happened (work got lost)
in collaboration projects, is it also collaborated funding,
scope? refeds is too small (but covers EU and other countries, no paying members, …)
- best practise
- license choice
- social coding?
- governance curation?
productization
short sharp innovation, revolutionize, small forks,
not something (to be around) for 20 years
narrow usecases, small audience
NORDUnet's IPR page
what's created should be as available to community as possible
SURFnet ended up with apache2 instead of un?restricted BSD
analysis of not just license but also ipr, patents (violations), ...
trustmark - don't want to go there? (still someone needs to check responsibility)
no one-size fits all model, tool for:
- listing in repository
- host the software
next steps:
- setting up space on github, paying
- some kind of website
- document
single stream of thought,
not often forks of our projects (for people that disagree outside the stream)
don't sit on stuff and get lost in (own) priorities
if it's in the open, you do get usefull requests and input
flesender: dayjob evaluating what's coming in, requirements, everybody wants 1:1 interaction
need form of maintenance mode
(feature complete -> graveyard)
fear for projects not living on its own without funding from one
if it's on social coding: with people using it, it will be around
people using the code can't code per se, indicator for popularity?
we need running code, criticall mass should follow, at least it does what a small community does
internationalization, accessibility,
taskforce software development?
what process, scrum? is maintainable? aren't we all struggling with this?
what kind of outcome would be expect?
conversation could be in emc2
doesnt have to be internationalized, as long as it's possible to internationalize: no hardcoded stuff
not just coding, q&a, systematic testing
mediamosa had to build their own foundation
someones time is also financial contribution
but internal time is often easier assigned then money
ask terena to do something like apache foundation?
buy another service/software from another nren? ppl apparently rather buy from terena
(you want to buy it from someone that has it as core business, would nren sustain: rather filesender from filesender foundation then aarnet)
rather buy box.net from box then from a uni with barely svn repo
bikeshed, if terena runs something discussion on what's best,
now a lot of solutions running at nrens, convergence
terena software foundation? embracing apache/...? are we special enough?
quality mark? security audited/internationalization/...
lacking funnel for the money
terena shouldn't be software company
some certifications that stuff is ok (or are we waiting for emergency)
shouldn't be prequisite for being part of the greenhouse
Notes from Niels Van Dijk:
We all dev software, the same kind of tools, same usecases
Terena has been contacted by several.
Resources and financie is an issue?
What tools are good candidates?
Mindset change needed: build in the open.
- make software location open,
- make stuff reusable, readable, openly available.
- licence is not the main issue, however having only a few within the community would really help, also for management and lawers to make it easier
- technical storage of teh code is not the problem. (github, svn etc)
- - Discovery is the problem
- - Funding and resources is a problem
- - how to work with management on this?
- - How to move from startup project to 'stable' product?
- - How to take care of incubation? -> Running code
SUNET IPR statment: - goo.gl/dDz5v