ViewVC logotype

Contents of /trunk/community/interview/interview_r.nair-a.tille.log

Parent Directory Parent Directory | Revision Log Revision Log

Revision 13060 - (show annotations) (download)
Fri Feb 22 17:23:21 2013 UTC (2 years, 2 months ago) by tille
File size: 26524 byte(s)
Continued interview
1 Interview between Rajeev Nair <rajeevisonline@gmail.com> (RN) and
2 Andreas Tille <tille@debian.org (AT) at So, 17.02.2013
3 Rajeev Nair is Business head of http://www.healthcafe.in
5 RN: Hi Andreas are you free now?
6 AT: Fine for me to talk right now
7 RN: ok
8 great
9 ok my first question
10 what is debian pure blends..how is it different from a debian fork or custom distribution?
11 AT: Short answer: It is by no means a fork neither it is any custom distribution.
12 It is pure Debian
13 Just adapted to special needs of specific users
14 Please have a look at the doc
15 http://blends.alioth.debian.org/blends/
16 RN: I did go through it...my question is can you actually download a debian pure blend
17 as an iso image
18 AT: Well, for Debian Med nobody so far has done the work to create some iso image containing this part of Debian.
19 In the Debian Edu case there is a separate release called "SkoleLinux" which can be installed in schools.
20 If you want to have Debian Med you should simply install the metapackages named med-*
21 In principle I would be happy to care for such dedicated images but nobody did the needed work so far.
22 RN: yes I think too that it would be a good thing
23 AT: What would you personally expect from such ISO image (just for the sake of interest)
24 RN: good for doctors etc. who dont know too much linux
25 AT: Do you expect that doctors in their practice should maintain their computers themself?
26 RN: a complete install setup for sever and desktop and out of the box apps like freemed, gnumed etc.
27 not completely but just easier so to maintain
28 AT: I would like to contradict here
29 Doctors are specialists in healing people but not in maintaining computers.
30 Health care is too important to let non-specialists fiddle with their IT
31 You might like to have a look at my talk at LSM
32 http://people.debian.org/~tille/talks/20120711_vista-in-debian/index_en.html
33 There is video and the slides
34 Slide 26 of 29 is explicitely handling this idea
35 RN: I will look into it
36 thanks for the info
37 AT: As you might have seen Debian Med has a large part of Bioinformatics which is also a part of health care (not so visible for the patient but in wet labs is done a lot of stuff for your health as well)
38 So installing some clusters for gene sequencing software quite simply makes way more sense to use dedicated ISO images
39 This could be automated easily
40 However, installing plain Debian plus the dedicated metapackage is only one more command line for a bioinformatican ... and those people usually know how to do this.
41 RN: so it is something like apt-get install med-* in debian right?
42 AT: Yes.
43 Biologists just will say
44 apt-get install med-bio
45 or if they want to develop biological applications they say
46 apt-get install med-bio-dev
47 RN: do we have to add an ppa's?
48 AT: No, not at all. That's the point in having pure Debian (as I said in the beginning)
49 If you have Debian than you have Debian Med
50 Debian Med (as other Blends) could be considered specific work groups
51 RN: ok another question will it actually work on ubuntu?
52 because ubuntu is more commonly in use these days
53 AT: Ubuntu takes over all packages from Debian as you know. Because this is an automatic process it simply takes over the complete work of Debian Med and as far as I know (I never tested it) everything should work nicely on Ubuntu.
54 That's one of the huge advantages to stay completely inside Debian - a fork / custom distribution would not have this feature.
55 RN: yeah I agree but the driver support is better in ubuntu atleast is better?
56 AT: Well, the medical software is more or less independend from any drivers - so it works nicely inside Debian and Ubuntu
57 My answer above was not about Debian versus Ubuntu
58 RN: ok
59 I am a huge fan of debian as well
60 AT: It was about creating all stuff right into Debian to let all derivatives from Debian (including Ubuntu) profit from our work.
61 RN: I remember spending money to actually get debian when sarge came out it was simply awesome
62 AT: :-)
63 I do not see any point in Debian - Ubuntu flamewars. I'm just a Debian developer because it was there in the first place (I'm DD since 1998) and I stayed there. Debian Med started in 2002.
64 I never had any reason to change to something else but if somebody profits from my Debian Med work via some Debian derivative (Ubuntu, Linux Mint, whatever) that's perfectly fine.
65 There is actually BioLinux
66 http://nebc.nerc.ac.uk/tools/bio-linux/bio-linux-7-info
67 which derives from Ubuntu and we are working closely together to complete the set of packages they are using
68 RN: yes I heard about biolinux..didnt try it out
69 But Debian-med looks better for now
70 AT: It is probably not about better or worse.
71 As I said we are working together.
72 RN: ok
73 AT: BioLinux has some specific users in BioInformatics with specific needs who get some support
74 We try to make sure that everything that is needed for BioLinux will be available right inside Debian.
75 These packages will migrate to Ubuntu
76 Then BioLinux does some adaptations which are considered nice by their users.
77 AT: I could imagine that these adaptations could be done right into Debian as well and we try to make their work / the diff they need to create as small as possible.
78 Finally we want to make sure that as less work as possible needs to be done twice.
79 RN: ok
80 RN: back to Debian-med again...does it automatically install all the packages marked in yellow on the site?
81 or just the dependencies?
82 AT: No. Unfortunately these packages are not yet ready.
83 The green packages are inside Debian.
84 They can be installed.
85 The yellow packages are our current work.
86 RN: yes but going ahead we will be having those packages as well?
87 AT: Yes.
88 However, there is no guarantee how fast we might proceed
89 Some packages are more or less simple but for the VistA package
90 http://debian-med.alioth.debian.org/tasks/his#vista
91 this requires a huge lot of work
92 It is on our todo list since the start of Debian Med
93 Since last year we made a really big step regarding getting the precondition fis-gtm
94 We might be close to a first upload of fis-gtm in spring 2013
95 Once this is done a vista package comes way closer.
96 RN: can you compile Vista from source as of now?
97 AT: Well, I could imaginge that you can VistA after having build fis-gtm from the packaging stuff.
98 I'm personally no real expert for VistA
99 The thing is: we strongly depend from upstream expertise.
100 RN: but packages like freemed and gnuhealth are working fine..
101 AT: The big step foreward last year was because upstream became involved strongly into the packaging.
102 (I'll continue vista topic and answer gnumed next)
103 Another important idea of Debian Med is to bring together upstream and users (kind of missing link between those parties)
104 AT: Last year I started the "Mentoring of Month" effort to do some formalised education for upstream developers how to learn Debian packaging.
105 http://wiki.debian.org/DebianMed/MoM
106 Upstream knows best where to install what file and how to test - if we teach and help them in the packaging we can guarantee that the packages are properly built.
107 So you see the Blends idea goes a bit further than simply adding packages to the Debian pool but rather create strong links between all involved parties.
108 RN: by upstream therefore you mean IT specialists working in healthcare?
109 or do you mean researchers?
110 AT: Upstream are the authors of the software we are packaging
111 That's Debian "slang"
112 RN: ok
113 got it now
114 AT: BTW, in biology upstream authors are quite frequently researchers and to reward their effort we mention their publications on our web pages - just look here
115 http://debian-med.alioth.debian.org/tasks/bio
116 So back to gnumed and freemedforms
117 Both are systems to manage single doctors practice.
118 You see that more of these are in our focus but the work is actually not that much progressed
119 (unfortunately)
120 Gnumed resides since a long time inside Debian
121 I had strong personal contact to the authors (they also live in Germany like me.)
122 Freemedforms packages were created last years
123 In this case also the upstream author was heavily involved and finalised the packaging under the help of the Debian Med team.
124 BTW, to support my first point about the simple installation by a doctor.
125 If a doctor simply installs med-practice on his system he will end up with two "competing" systems.
126 That's fine for evaluating the systems and having a look etc.
127 However, for real work he needs to pick one.
128 So it makes perfectly sense to have some support company who just uses our preparation in Debian Med to install the computer of the doctor
129 He gets support and just the software he might want to use.
130 RN: ok but there is a problem when it comes to private practice right..if he wants to maintain records
131 do you recommend that he must have an IT specialist to help him?
132 AT: Yes, definitely (that's basically the content of slide 26/29 in my talk at LSM - see link above)
133 RN: ok
134 AT: BTW, that's also the case for proprietary programs
135 Free Software will not change this principle
136 RN: ok software remains free but you always pay for support
137 AT: yes, that's the point
138 RN: what do you think about competing projects like Opensuse medical...I read somewhere that they merely copied debian-med as it is...
139 AT: I would like to support any such project because I do not see any real competion here
140 We are all working for free software.
141 RN: ok
142 AT: So I'm subscribed to Fedora medical SIG mailing list and Opensuse medical list
143 I also try to give hints how things can be done.
144 However, there is no visible work going on on opensuse medical
145 RN: yes I noticed that
146 it says page not found for the gallery
147 AT: In my talk at FOSDEM I was refering to something like
148 http://en.opensuse.org/openSUSE:Medical_packaging_bio
149 It is a simple 1:1 copy of med-bio at some random point in time
150 I simply regard this stupid for one reason:
151 If users stumble upon this page they see something and want to use it.
152 Finally they will realise that there is actually nothing they could use.
153 This leads to frustration about Free Software because of stupid advertising of nothing.
154 It creates bad feelings about any Free Software
155 RN: ok
156 AT: In case they might stumble about Debian Med pages later on they think: Well, I have seen something like this and it was nothing
157 So pages like this do not help in the end.
158 RN: but free software has come a long way with so many packages
159 ok
160 AT: Yes. But you know users might come from some random site and will not know so much. If their first entry point is a failure they might stay away for years
161 So I'm very carefully about advertising wrong things.
162 RN: ok
163 AT: In the same line I do not blindly advertise Debian Med to any doctors I know.
164 I'm not a IT support company and I personally can not do support for them
165 But support is definitely needed
166 As long as there is no professional support I should not drag people into this.
167 RN: but dont you think its about time more people got to know about this blend?
168 by way of articles etc.
169 AT: Yes
170 You are right it is time to let people know.
171 But it is also time to let the right people know
172 RN: totally
173 AT: There should be IT support companies know that they could make a fortune delivering Debian Med to doctors
174 They could get it for free and they could share the money they make with Debian Med (paying for meetings etc.)
175 RN: And its affordable for the doctors as well
176 AT: Sure - they will save the licensing costs
177 RN: yes
178 Actually I was thinking this looks good as an interview.
179 already
180 AT: And we should also make sure that doctors will understand this and will ask the support company to adjust their price accordingly
181 RN: yes that will definitely come about
182 AT: Doctors should understand what Free Software means
183 RN: yes
184 AT: Any other question?
185 RN: yes there is...see I could go ahead and publish this as an interview..
186 but I wouldnt have the expertise or background that you have
187 AT: Yes, you can publish everything.
188 As I told you yesterday I also would like to copy this as is to Debian Med mailing list.
189 I will put you in CC when I write the mail.
190 RN: ok
191 sure
192 but theres one more thing...
193 AT: what else?
194 RN: see people in India always like to read articles coming out from the west
195 particularly written from the west..where most of the research goes on..
196 AT: I wished research in India would get some better credit in its own country ...
197 RN: we have had a lot of articles written by people from different parts of India
198 yes..I think in India it lacks the funding and interest..but its coming up
199 AT: good
200 RN: but it would be really great if you could write an article for us...
201 along the lines of free software, debian med and affordability and need for support from IT specialists etc.
202 I could write but I am not an authority on it..
203 AT: This would be feasible and there are some articles about Debian Med we have created before.
204 We could try something on the line of this interview and we usually work together on articles in VCS (formerly SVN, we are in process to switch to Git)
205 I would really welcome if you would check out the work and would commit your own part of the article
206 RN: I will do that but you get the picture right
207 AT: This would have the advantage that you know best what might be interesting for the audience
208 RN: it would be great for our magazine to have your photo and article for our magazine because you are an expert on the subject
209 AT: No problem
210 RN: I'll put some inputs from my end as well but it would be just better if it went in your name
211 some freshness for the magazine
212 in a software perspective
213 AT: Fine for me - but it might be some teamwork of the Debian Med team and I would like to give credit to all authors (as we did in the past)
214 RN: ok as you wish
215 AT: We will see how it works and how other people will be involved
216 RN: it is a good idea to include their names as well at the end
217 ok
218 AT: ... something like this.
219 RN: once again it is a privilege to talk to you
220 AT: :-)
221 That's to much of honor.
222 RN: no you deserve it
223 AT: I always like to talk to people about Debian Med
224 RN: I think many times that people fail to recognize the work that goes into making free software as well because these days they are as good or even better than their commercial counterparts
225 people should be donating more money to the cause
226 after all theres nothing else in this world that is truly free
227 free to own, modify and distribute
228 AT: I guess this is similar to the Coca Cola effect: It most probably does not taste better or worse than any comparable soft drink. But there is a lot of advertising around and so people believe it is way better.
229 In the same sense there is a lot of advertising about proprietary software
230 RN: yes but I believe now debian-med is getting popular as well
231 AT: :-)
232 RN: and that is a very good thing
233 AT: Now you simply need to make 1 billion people in India believe. :-)
234 RN: thats the idea
235 at least all the people subscribed to our magazine would
236 AT: In Debian slang this is called "the last final step to world domination" ;-)
237 RN: and thats always in a good sense
238 debian for peace
239 AT: :-)
240 BTW, we will have the next team meeting next weekend:
241 http://wiki.debian.org/DebianMed/Meeting/Kiel2013
242 I will try to work on an article when traveling
243 RN: ok that would be great
244 it would be a nice incentive for doctors etc.
245 AT: I hope so
246 RN: btw our magazine is essentially for the layman as well
247 so they will be able to understand too
248 btw does debian-med have any packages that patients can use?
249 AT: Yes
250 http://debian-med.alioth.debian.org/tasks/tools
251 pcalendar
252 workrave
253 quitcount
254 etc.
255 RN: ok
256 maybe we can focus on that as well...
257 AT: Yep.
258 RN: so it makes for a wholesome article for everyone..
259 AT: Perhaps you might like to create some small skeleton
260 RN: and particular advantage to healthcare systems
261 I could do that but would that be ok with you?
262 AT: Sure. That would be pretty helpful!
263 RN: no i don't think my input would be anywhere as good as yours..but I will still try
264 AT: While you consider me an expert in Debian Med I'm by far no expert in writing a magazine article. Here you are the expert and any help is welcome
265 RN: ok
266 but you're presentations are pretty good
267 AT: Thanks
268 RN: I saw the video where you spoke of 'bike shedding'
269 it was light hearted and good
270 very informative
271 AT: seems the lates FOSDEM video ...
272 http://video.fosdem.org/2013/crossdistro/Debian_Med___a_Debian_Pure_Blends_for_medical_care_and_microbiological_research.webm
273 RN: yes
274 you spoke about initial problems with the name etc..
275 AT: yes
276 RN: must see the whole video still
277 I have a slow internet connection thats the problem
278 but it will be the first thing I do to finish downloading it
279 to see it completely
280 AT: On the other hand if you say the presentation is pretty good: I always try to put the focus on explaining that a Blend is no fork / custom derivative.
281 However, your first question in the interview was exactly this.
282 I think you understand it now
283 RN: yes I do
284 AT: But I really hoped that people understand it right from the presentation
285 In previous presentations I experienced that this was the first question in the QA part ...
286 For me the idea is very straightforeward to stay inside Debian but to my big astonishment people very frequently have big problems to get this idea
287 RN: yeah its always good to maintain the same OS
288 less work for the developers as well
289 shall we talk about the rest later
290 I am really sorry Ive got to go
291 AT: Sure this is what I tried to explain. I remember four years ago when I was talking about "Debian Science"
292 After my talk the first question was: "Why do you creating a fork? That's not needed!"
293 well, for sure that's not needed - and that was my talk about ...
294 RN: Andreas I ll be back in fifteen minutes..
295 AT: Yes, we can continue the interview later
296 RN: ok
297 thank you for your time
298 bye
299 take care
300 AT: bye
303 Continued interview at Mo, 18.02.2013
305 RN: lets pick off from where we stopped.. you were talking about "Debian Science"
306 and about creating a fork..
307 AT: If I remember correctly my point was that my presentation are obviosly not as good as I wanted them to be if I intend to explain in detail that Blends (in this case Debian Science) are no fork and the first question in the QA part is "Why do you fork Debian?"
308 RN: yes
309 RN: but what was the presentation about Debian Science?
310 RN: is it related to Debian Med?
311 ok it is a pure blend like debian med?
312 AT: Well, Debian Science is another pure Blend (like also Debian Edu, DebiChem, Debian GIS, Debian Multimedia)
313 However, it is more like an umbrella project from where more specific Blends (like potentially Debian Math, Debian Physics, Debian Electronics - whatever might come out of this if people might start to care about this)
314 For the momen Debian Science assembles a lot of scientists packaging several kind of scientific software.
315 RN: and you are maintaining that as well...
316 RN: I mean maintaining Debian Science too
317 AT: Well, I have a few packages in Debian Science (for instance wordnet) and I was activ in the process of forming the Blend and I'm lurking on the mailing lists - we are just somehow connected. However, my spare time does not enable me to be very active in more than one project.
318 I just had some talks about Debian Science in the past which is not that a big deal because it is handled with the same Blends tools and following the same principles.
319 RN: Is there any government initiative to support Debian Med in any way that you know of? any organization?
320 AT: No, not that I would be aware of.
321 RN: I mean in Germany or elsewhere in the world or any other govt?
322 AT: The only really large scale Free Software in Medicine installations are some hospitals running VistA
323 There came up some universities who are supporting Debian Med in their BioInformatics branch to som extend.
324 I'm not aware of any governmental support
325 RN: do you know of any universities spec?
326 AT: IMHO one of our team members is professor at University of Munich
327 But if you want to be sure you would be best advised to ask on our mailing list <debian-med@lists.debian.org>
328 It might perfectly be that I'm not fully informed about this.
329 Youo know: Free software can be picked by anybody and there is no requirement to let us know.
330 I really can not tell to what extend the support / usage is "official"
331 AT: I can only say pretty sure that there is explicitely no support by German government for Debian Med (I'm working on an institut that is attached to the ministry of health and it is not used here.)
332 RN: do you think reliability is an issue people are scared to move into a new area or lack expertise to handle it?
333 AT: The reasons are complex and you are mentioning only a part of it. I do not think that I could give a concise overview about those reasons here. Debian Med is just another Free Software project amongst a whole lot of other very reasonable projects - non of them gets governmental support. So why do you expect Debian Med to be supportet specifically?
334 RN: because
335 1) It is a medical project..concerns the health of citizens
336 2) could minimize costs in third world countries
337 3) free support from the web..assuming that it is possible
338 4) Debian is freely available and the most popular linux available
339 AT: So do you want to discuss this with your / other governments. I know these reasons but I spend my time in making Debian Med better instead of talking to politicians. I consider lobbying something as a very boring job - specifically if you have no money to spend for your lobbyism.
340 RN: yeah I think we ought to project debain-med to governments...conduct workshops for healthcare etc..
341 would be a good thing...we have done so much already..why not go the extra mile?
342 it is a good thing after all we're not terrorists or anything..
343 RN: :)
344 AT: I will not stop anybody from doing what you are proposing - at the contrary, I'd be happy about this. But this is not my business. I can do and I have fun to do a lot of technical work and I can talk on technical conferences. I will not cut from this time for other things.
345 RN: ok it would be great if we could chalk out a plan of action..some policies that people can follow and how debian-med can help
346 unfortunately for the governement reps must actually sit and discuss with maintanence and packaging team(yourself) and upstream..
347 After all so much of work has actually gone into free software right?
348 I think we must market it more agressively as a matter of principle
349 like spreading non violence... green peace etc...
350 an actual movement with free software...
351 where it stops being an idea and becomes a revolution
352 RN: I dont know what people constitute our govts...but if I saw an initiative like this Id promote it
353 RN: anyway thats how that stands I dont want to bore you
354 let us talk about something else
355 How many people are currently working in your Debian-Med team?
356 Are there any people from India?
357 AT: It is always hard to say how many people are working because you need to define first what amount counts as work and how long the time span of inactivity could be to say qualify as "recent work"
358 I was wondering about this question since some time and cam up with graphics you probably have seen in my FOSDEM talk
359 AT: The graohs shows the ten most active people and I think for the Debian Med team the graph would be interesting even for the top twenty active people. Considering that probably a >30 people graph would not tell much new things you might consider the Debian Med team of about 20 to 25 people.
360 (people = active people in terms of providing code and discuss actively on the mailing lists)
361 AT: Regarding people from India: Yes, my current MoM student (https://wiki.debian.org/DebianMed/MoM) Sukhbir Singh is from India (but moved to Canada last year)
362 RN: How do you go about packaging in debian? Do you actually obtain source code?
363 I mean for Debian-Med..
364 AT: As any normal packaging: Download the source from the internet, check the license, do the packaging (in VCS - we are a team), upload to Debian mirror. Fully normal, nothing at all specific to the topic.
365 We are not developing any software on our own (at least not with the Debian developer hat on - there are some members in our team who are also upstream developers of some software.)
366 RN: Well it is a great thing you're doing...the packages in yellow...they are there because of missing dependencies or because projects are not supported anymore?
367 AT: In most cases it is just a lack of manpower. (Do you want to start working on this ;-))
368 RN: I would love to but I am not an expert you know that...
369 :)
370 And I am spending way too much time bringing my articles to completion for the magazine..
371 I used to work on linux extensively earlier but have been out of touch for a while now
372 AT: Yep. This question was just to make you understand why I do not dive into advertising Debian Med to politicans - I would simply just need to leave my workfield ... and I would hate to exchange something I do enjoy by something I do not like.
373 You just spend your time in what you feel good in ... and I do so as well.
374 RN: but we can always arrive at an amicable solution that is for the good of everyone rite?
375 I agree with you totally..now that I think about it
376 AT: ... by sharing the work to do between experts.
377 RN: ok I agree...hows the article panning out by the way?
378 RN: Do you think that you can write something for our magazine..so as to make Debain-Med more appealing to all of us at least in print..somethings that can be done..anything that you can think up really...:)
379 we can have a picture of the entire team as well on the following page
380 AT: As I said I will try to spend the traveling time to and from our sprint meeting. I'll let you know after next weekend. Everything will be available via SVN so you can check
381 RN: ok...
382 and you would discuss something at least related to policy and how healthcare can be improved just for the sake of the article...
383 please dont say no
384 RN: :)
385 RN: all right let it be as it may..but please let it be so that any health technician reading it is drawn to it
386 thanks and bye
387 we'll talk later
388 take care

  ViewVC Help
Powered by ViewVC 1.1.5