I just learned about NHSbuntu which I consider a great effort. Seems this implements the idea that was driving Debian Med since its start in 2002. Debian Med is a team inside Debian with the objective to package free software in life sciences and health care for official Debian. Since NHSbuntu is a Debian derivative we would be happy to include into Debian what is missing according to your opinion. Feel free to contact us via the Debian Med mailing list list. Since we can not guarantee that every packaging request is fullfilled either due to lacking manpower or simply lacking competence in very specific software we are running an effort Mentoring of the Month to teach interested people how to package it for Debian on the actual example. That way we helped several interesting packages into Debian while beeing maintained competently by a member of the upstream developers team.
Looking forward to a healthy cooperation between Debian Med and NHSbuntu
Welcome! I always found packaging debs a little cryptic, nice to know there is help. I always wondered if it was the sort of thing amenable to having an IDE help out. I know Ubuntu had some help to do this sort of thing but AFAIR you have to push code to bzr branches in Launchpad.
Definitely we see opportunity for collaboration and cooperation between Debian Med and NHSbuntu, and we certainly are not trying to undermine the excellent work you have already done. In fact, as our ‘all code has already been written’ motto suggests, we’ll be leveraging as much existing work as possible, and delivering that into the hands of NHS clinicians.
Just to explain why we felt we needed to make our own ‘special’ OS when there was already Debian Med - we’ve chosen to make an NHS-specific version of the OS because there are particular technical requirements, and some serious cultural challenges implementing software in the NHS, which we felt might be solvable by making something directly targeted at the NHS. We also felt we could get a bit of a leg up in terms of visibility by doing a lighthearted ‘brand hijack’ on both the NHS and Ubuntu. It saves a hell of a lot of time not having to explain the basics of what we’re doing - with those 8 characters ‘NHSbuntu’ most NHS tech people know immediately what the game is
Marcus (NHSbuntu co-founder)
in Debian Med we use Git (or alternatively SVN for old-schoolers or specific purposes). There is no IDE help, but in the end packaging is crafting a sensible makefile which in the simplest case is a 2-liner. Feel free to come up in a Mentoring of the Month project and evaluate yourself if you know a software project that is not packaged yet but should.
Kind regards, Andreas.
I never assumed any intend to undermine our work (I even see no real chance to do so since noone will stop us to add more medical software to Debian ) We are also following the everything is written (there is no code development in Debian Med - just fixing upstream issues) and we want to bring it to our users as smoothly as possible.
I also understand the need for specific NHS customisation which has also a marketing effect. Rebranding Debian or its derivatives is a very frequent thing. BTW, possibly you want to register on Debian Derivatives Census.
I was just imagining some kind of cooperation as we had established with BioLinux. BioLinux was started several years ago by (at least) three contributors who somehow moved on to other jobs. The last contributor Tim Booth once wrote a message to the Debian Med list which might be very interesting for you. In other mails Tim confirmed that BioLinux is more or less legathy software and new things should happen directly in Debian Med.
My conclusion is that projects should work together from the beginning with Debian Med and exclusively do the final customisation (be it of Debian or Ubuntu). The price might be some delay of new packages showing up in NHSbuntu but the gain is sustainability - I have seen lots of derivatives with a similar focus come and go. That’s why I’d suggest to subscribe the Debian Med mailing list right from the beginning and I ask you to explicitely take influence actions there in your very own interest.
Kind regards, Andreas.
I’ve been discussing NHSbuntu with our technical guys on site since I found out about it a few weeks back. The big stumbling block for us seems to be smart card authentication. Do you know if DebianMed includes this functionality and if it could be incorporated into NHSbuntu?
Chris (IT Trainer)
to be a bit picky about wording: The question should rather be whether Debian (not Debian Med) includes smart card authentication. Debian Med is fully integrated into Debian - you litterally have anything contributed by the Debian Med team if you have a full installation set of Debian. This kind of projects are called “Debian Pure Blends”.
Checking the Debian package pool (I did this via the command
apt-cache search smart card | grep auth
there might be better ways to seek for this) I detected the package libpam-poldi (see
apt-cache show libpam-poldi
most probably also on any Ubuntu system - please check yourself I do not have access to an Ubuntu machine). The package is not maintained by the Debian Med team and I consider this a strength of Debian Med that we are not forced to do anything ourselves since its just a natural part of Debian. The Debian Med team would start playing a role if there would be a medical application that is using the general libpam-poldi pam module and we would try to package this - most favourably with your help since you know how to test its functionality if you are using it yourself.
Hope this explanation helps, Andreas.
There is no technical reason that prevents smartcards being supported in
nhsbuntu. If the US Department of Defence Common Access Card is fully
working, an NHS Smartcard can’t be too great a challenge.
In our development work we’ve successfully read the certificates on the NHS
smartcards using the pcsd tools. The omnipresent OmniCard reader is
supported as are readers that support the nfc chip.
PKS11 and a whole lot of other libs and tools make interaction with
smartcards of many varieties and functions play nicely on Linux.
Right now the ‘missing piece’ is the Identity Agent, the code that pings
the Spine endpoint when the PIN is entered and is used to wrap up token
requests with the certificates. With that piece of code we would be able to
use the smartcard with any browser and access web services that require
Separately, there are Windows clinical apps that use Spine CIS for
authentication. For these we use a Windows VM to publish seamless apps to
the NHSbuntu desktop.