[Debian-med-packaging] Bug#639996: Bug#639996: fastdnaml is not able to start in GNOME

[melchiaros]

Am 01.09.2011 16:55, schrieb Andreas Tille:
> On Thu, Sep 01, 2011 at 04:21:16PM +0200, melchiaros wrote:
>>> "These programs are simple command line programs so what exactly should
>>> we expect?  The freedesktop menu should just open a terminal and call
>>> the program - nothing else.  Is there a way to implement this?"
>> It may be helpful to look at biolinux6 where it is implemented.
> I tried a comparison
>
> $ diff -u fastdnaml.desktop_biolinux fastdnaml.desktop_debian 
> --- fastdnaml.desktop_biolinux  2008-08-01 00:00:00.000000000 +0200
> +++ fastdnaml.desktop_debian    2009-09-04 14:09:29.000000000 +0200
> @@ -1,9 +1,11 @@
>  [Desktop Entry]
> +Version=1.0
> +Terminal=false
> +Exec=fastDNAml
>  Name=fastDNAml
> -Comment=fastDNAml
> -Exec=/usr/local/bin/fastDNAml
> -Encoding=UTF-8
> -Icon=/usr/share/pixmaps/dna.png
> -Terminal=true
> +GenericName=Phylogenetic tree construction
> +GenericName[en]=Phylogenetic tree construction
> +Comment=Tool for construction of phylogenetic trees of DNA sequences
> +Comment[en]=Tool for construction of phylogenetic trees of DNA sequences
>  Type=Application
> -Categories=Bioinformatics;
> +Categories=Biology;Science;Education;
>
>
> and tried to set the obvious difference 
>
>    Terminal=true
>
> in the Debian desktop file which just opened a terminal.  Would you
> consider this a fix for the problem reported?  Regarding severity I
> would call this at best minor if not wishlist.  The rationale behind
> this opinion of mine is that I can not really imagine that anybody
> honestly will start a command line program that way and then tries to
> input full path to the data he want to process with this program.  I
> admit that I'm not a user of fastDNAml - but I would strongly advise
> not to use the program in this manner.
>
> In my opinion the only reasonable thing to do in the menu is to fire up
> a documentviewer / browser and *teach* people how to reasonably use
> command line programs like this.  (I actually implemented this way in
> the Debian Menu system for CLI programs in the Blends framework.)
>
>> (Also with icons for all programs in the GNOME menu: Well a standard
>> icon, but there are no free spaces).
> Having some kind of a standard icon for such programs - possibly using
> some common icon in the metapackage and linking to it might make some
> sense.
>
> Kind regards
>
>        Andreas. 
>
:)hah:)
yes,.. I can understand what you say:" ...

browser and *teach* people how to reasonably use
command line programs like this...."

o.K what I have to say about this is that you have to know that the most biologist are pure noobs on computers(pure bioinformatic ones are exceptions - I mean mainly all the other:many:eg.geneticists;which also need some tools) and I agree that it would be better to teach them what they need, but the thing is you have to get them where they are.

When some of them went to linux it was a big step for them, when they find debian it is a bigger one.
Give them a clear view on the system:

You have two choics:

1.Separate command line programs completely from the GUI one: -> than there should be no entries in GNOME menu for the terminal programs -> only entries for GUI ones. Who need terminal programs have to go to the terminal by self and start them there.

2.Give them access to the command line programs from the GNOME desktop, than there should be working wrappers(and Icons -> you know they came from windows and associate missing picture very easy with low quality).


As far as I am affected: Choice 1. is absolutely o.K to me. The think is I do not want to see it mixed a semi working.

So I stay in my opinion that it is not a low severity, because a unskilled user recognized it as not working.

A suggestion: Choose choise 1 for cleanup the debian desktop and go succesive to choice 2.