Re: [Gnumed-devel] Re: Texlive and Gnumed

[Sebastian Hilbert]

Am Montag 08 März 2010 21:02:18 schrieb Gour:
> On Mon, 8 Mar 2010 20:07:38 +0100
> 
> >>>>>> "Sebastian" == <address@hidden> wrote:
> Sebastian> I predict that MeeGo will not take off (unless they get a
> Sebastian> competitive app store going). Every mobile shop except Nokia
> Sebastian> will go for Android or Windows.
> 
> Hmm, don't forget that MeeGo is going to be full & open development
> platform offering choice of developing apps not just in single
> language (Java).
>

That makes it attractive to the developer. The user does not care one bit 
which programming language was used or if there is more then one. There are at 
least four players in the market. Google has enough money to pull it off. 
Apple has enough design to sell users anything. LiMo is trying to get a bite. 
Now Nokia and Intel want in as well. 

Will be fascinating to watch. Tell you what. Operators want it easy. They 
don't care about the user. They will go for Android. Apple will sell a few 
more apps but will go down as soon as the first iphone look alike phone is 
out. They will fail as soon as someone in China or Taiwan gets the message 
that users buy design.
 
> Sebastian> Nokia will defeat itself by agonizing over the choice of
> Sebastian> Symbian (now FOSS) or MeeGo. They switched Maemo from GTK to
> Sebastian> QT. Still unless you get apps out there you will go down -
> Sebastian> no matter the technology.
> 
> I'm sure Symbian is dead and will be replaced by Maemo. 

Don't count on that. Symbian is a stable platform. Remember operators don't 
care about you being able to install apps. They want you make calls or produce 
traffic they can charge you for.

> There is
> already Qt SDK available for Maemo 5, while Maemo 6 will probably be
> MeeGo.
> 
I have an N810 myself. Nokia will sell a few N900 but unless they market that 
thing as a phone only devs will care. Users buy design. They don't care if you 
got root or not.

> That's why I see the need to use Qt as the platform of choice since
> although there is some interest within wx community to provide wxQt,
> they simply do not have resources to compete with Qt.
> 
Qt has evolved as a mainstream platform. I am not sure it will make inroads on 
phones.

> And in my case, MeeGo is interesting for me as developing platform
> enabling me to use my favorite language for both desktop and
> 'desktop-lite' (aka mobile) apps, not depending on the MeeGo store.
> 
As said before it seems a good choice for the desktop and has the added 
benefit to run on mobile platforms. You future product will not depend on the 
MeeGo store. How many users for your software do you expect if the platform 
never takes off ?


> Otoh, even present Maemo & Moblin offer more apps than I need even
> today, so having support for ARM + X86 will just improve things.
> 
We have yet to see how they are going to pull that off. The idea might be 
catchy. Until one can download anything MeeGo is nothing more then MeToo (TM)

> Sebastian> Same is true for GNUmed. Users judge features. Since they
> Sebastian> are used to buggy software (trained by MS) they will always
> Sebastian> chose a fully featured crashing app over a technically
> Sebastian> superior feature lacking app. Simple as that.
> 
> I will repeat what did many times before: provide simple(r)
> installation for GNUmed and it will become more attractive.
> 
Ok. Lets assume for a second that you are right. How hard is it to install 
GNUmed ?

apt-get install gnumed-client
apt-get install gnumed-server
gm-bootstrap_server

Show me one single software product (or other EMR) that uses PostgreSQL that 
is easier to install. I vaguely remember the discussion about arch linux. I 
know there were some points. None of them were convincing (to me).

The preferred method to install software on GNU/Linux is distribution specific 
packages. We have those for all major distributions. The smaller ones never 
indicated any interest.

Find me one distribution which cannot run GNUmed in 5 minutes from a 
downloaded tarball (dependencies are up to the distribution) and I will 
rethink my statements.

> However, in Croatia MD having his own clinic can use only 'certified'
> software and all those are written for the buggy & popular platform (can
> you guess?), so there is simply no choice for poor doctors than to choose
> the one which is the least bad amongst bad ones. :-(
> 
That is why GNUmed will need to be certified. This is a major effort. Many 
doctors here in Germany have written their own programs and even got them 
certified but the will fail because one man shows don't fly anymore. I have 
yet to see another EMR that is run by one person or another EMR that will 
attract a community.

Take a look at elexis. It is written in Java and Eclipse. They care little 
about technical design and have a 'get the job done' approach. They produced a 
running software which can be used today. Despite using JAVA they have trouble 
attracting a community. There are simply not enough computer literate docs out 
there. And the ones out there are attached to one project or the other.

> Be assured, that after speaking with many of them and having
> experience of my wife while working with those packages (one of
> client/server app, and when the vendor's server went down, all the
> clinics in the Croatia using it were down as well) many doctors are
> cursing the software choice.
> 
Well cursing is one thing, taking action is another. Let me tell you that only 
the buck in your hand or bank account is a buck you can count on. Whining is 
common for the majority of the population. Change is never easy. Don't count 
on it.

> I'm not sure if the situation will improve when we will be allowed
> into EU,

It will not.

> but that's why having GNUmed stripped from some of the MD
> backage and tailored for other specialized uses in the medical field
> is worth of investigating/evaluating.
> 

There is nothing to investigate. It is generic enough. One cannot expect one 
GUI to work for all. So every profession will have to rewrite the GUI. We have 
good contact with physiotherapists. One of them started his own software 
because he already knew JAVA and need something working now. There is little 
you can do about that. Think about yourself. You were to use Haskell. How many 
codevelopers did you hope to attract. If you want to go mainstream you only 
got C++ or JAVA. Guess C# might be one lesser known option.

Again. Unless you punch the keys it will all be a huge dream.

> Sebastian> I also predict that desktop apps will become obsolete in the
> Sebastian> eyes of most doctors and users. Doctors will be happy with
> Sebastian> every webapp you throw at them as long it is free and will
> Sebastian> allow billing.
> 
> I threw some webapps in front of my eyes, but not being satisfied has
> come to GNUmed...ohh, I'm not the doctor. :-)
> 
I admit I do not exactly understand what you mean.  GNUmed has the potential 
to reuse large chunks of backend and middleware and provide a webapp on top of 
that. We have always stated (but never proven) that it is easy to write 
another GUI (QT , web or whatever). We will see if we are correct. Noone ever 
tried it.

> Gour (still believing in desktop apps)

Good thing. How about having both with the same codebase. I sure hope you are 
successful. Because that means more code and more happy users.

Sebastian