there is an urgent need to move the openEHR community mailing lists from their current location at CHIME, UCL to another location. We have now researched commercial options for handling mailman lists, and have found a number of suitable providers. In all cases, we would move the archives of all lists, as well as automatically resubscribing all members of all lists.
For most of these offerings, it looks as if we will have to change list address, i.e. openehr-technical@openehr.org will become openehr-technical@lists.openehr.org.
The consequences of this are (assuming the automatic resubscription):
taking part in new discussion threads should work normally
responding to existing threads, containing the original list address, will probably not work (unless we make redirection of mail to xxx@openehr.org => xxx@lists.openehr.org work)
any saved copies of openehr list addresses will be wrong. Usually people don’t rely on such things for mailing lists, so this should not be a big problem.
The solution above appears to be the only one available at the moment, and costs money, unless an academic institution in the openEHR network wants to offer to host mailing lists.
This move is part of a larger move of openEHR’s online presence out of UCL. It appears that for most things (SVN, Jira, Atlassian, website), paid online hosting will be the way to go, unless there are academic institutions interested in taking on the job. These changes will mean a) some (hopefully small) upheavals in the user experience, and b) some costs that need to be covered. I would think the board would be interested in any help on the second score.
Hi Athanasios,
I just talked to them... they can't handle domain names (everything ends
up at their own domain) and they don't host mailman, they use LISTSERV,
which might be ok (I have never used it). But I think we want to retain
the openEHR.org domain, which all commercial providers offer.
I used SourceForge before to host projects (yes that’s correct not just software development but collaborative project sites) in past which offers for free lists and many more, such as Web pages, SVN/Mercurial, blog and Wiki and many more. I reckon the licensing might be an issue for non FOSS projects but I believe in the case of openEHR that’s a non issue?? I must admit the SVN is not as fast and there’s limited administration capability but hey it’s still great value for nothing. Plus the platform also allows for ‘donations’ which I think might create few bucks to look after things like administration etc. until an appropriate funding mechanism becomes available.
I used SourceForge before to host projects (yes that’s correct not just software development but collaborative project sites) in past which offers for free lists and many more, such as Web pages, SVN/Mercurial, blog and Wiki and many more. I reckon the licensing might be an issue for non FOSS projects but I believe in the case of openEHR that’s a non issue?? I must admit the SVN is not as fast and there’s limited administration capability but hey it’s still great value for nothing. Plus the platform also allows for ‘donations’ which I think might create few bucks to look after things like administration etc. until an appropriate funding mechanism becomes available.
The established system with mailman is working. It has quite good spam protection, which is hard to get these days, and there is know how that allows the community to keep it running with minimum effort.
A switch would mean moving the archives, ensuring we have the same quality of service etc. We’ll spend time to figure out stuff, develop new maintenance know how, you know the drill.. I say we keep the lists in mailman; all of them.
Seref,
This is the better option in my opinion as well, for the reasons you mentioned and because I believe that having the lists at one place reduces the overhead, for admins and users.
Also, there may be other things like the website that we may want to move to the some place?
Sebastian
we have done a review on Google groups, and our research shows that they are not a safe option at the moment. Problems that have been reported:
some groups lost their entire history, and Google failed to respond to any requests for help
the spam control on Google groups is practically non-existent (I know this from personal experience on a technical group)
we have no idea what Google will do with any of its offerings beyond Gmail and a few other basics - it seems unlikely that google groups will continue in its current form
there is no detailed web-based admin control panel for GG as there is for mailman.
So for the moment, we thought it was best to keep the current lists intact, move them to a safe place (including list archives) and give the community the time to have discussions about moving to different list technology later on, when proper investigations could be made. For now, we are under pressure from UCL to move, and we also have located a reliable list host, that hosts ICANN lists among others. It also doesn’t cost much, and if we get the move done smoothly, noone should notice anything much.
I hope that sounds acceptable, it seemed the best solution in a short time.
I happen to be on a Yahoo group, which does actually work, but contains
advertising, which apart from being annoying visually, makes replying
inline extremely annoying. As far as I know, one has to pay to get rid
of it.
Yahoo groups also seem to suffer from the urge for improving my sex life, while wearing a fake rolex, as I travel to bank to help some guy inherit his late father’s 987 bazillion dollars…
It may well. I personally don't have experience of using more than a
couple of the available technologies, so we need community members help
to make such decisions.
There is no strict necessity for all software projects to be in the same
place, but if they are all in different places, it obviously makes
supplying any common adminstrative and educational support more
difficult. E.g. let's say for example we decided to stick with Atlassian
for everything, then everyone could learn Jira via some webinars. If
some or all projects are on Github, then people on those projects can
help teach each other about using those environments.
One thing I do know is that the work of learning how to use these kinds
of environments should not be under-estimated.
Yes, learning cost is very important. The reason why I am moving my Ruby works
to GitHub is learning and maintenance cost.
The most attractive point in Launchpad is translation related
features, but I think
it does not matter.
Thank you for your reply.
