[GRASS-dev] Making start of GRASS GIS easier for newcomers

[Markus Metz]

On Tue, Feb 3, 2015 at 4:56 AM, Vaclav Petras <wenzeslaus at gmail.com> wrote:
>
> On Mon, Feb 2, 2015 at 4:39 PM, Markus Neteler <neteler at osgeo.org> wrote:
>>
>> > For newcomers/no rc file: how about creating a new GRASS GIS database
>> > in $HOME,
>>
>> +1
>>
>> > in there a new location with EPSG:4326 (latlong wgs84), the
>> > empty PERMANENT mapset will be created automatically. Does not apply
>> > to MS Windows, because the demolocation is created during
>> > installation.
>>
>> In theory the demolocation could be everywhere since it is part of the
>> source code and even contains a world map.
>>
> In theory, copying the demolocation should be easy. I'm just afraid that it
> will be hard to implement it, to get all the conditions and rights right and
> to debug it.

$HOME is a safe place.

> However, since on MS Windows, the mechanism will not be used
> much,

I give a GRASS course right now, most attendants are on MS Windows,
and they happily use this mechanism.

> it is possible that some bugs will not need solving. See below, for
> one issue which is currently present and is even more important for this
> proposed functionality then for the current one.
>
>>
>> > Why not having a minimal welcome screen with a dropdown list for
>> > existing locations and a dropdown list for existing mapsets in the
>> > selected location, plus a prominent OK button and a change/manage
>> > button, nothing else?
>>
>> +1
>> At FOSDEM, we discussed this issue and most discussion participants
>> thought exactly in this direction.
>
>
> I wish I could be there. I still think that improving what we have now is a
> better way that, at expense of advanced users, try to hide the complexity
> which cannot be avoided at the end.

The idea is to simplify the welcome screen and move out less commonly
used functionality. Obviously we do not agree on what is less commonly
used.

> It seems to me that everybody is so
> annoyed by the 64 startup behavior that changing everything seems as an only
> option. However, I think that just small improvements here and there are
> better in this case. Complete revamp of startup with focus on first time
> users may easily create unsatisfied second time users.

Going through this thread, there are quite some second time users that
agree on such a minimal welcome screen. Long time users often start
GRASS on the command line with the desired location and mapset given
on the command line, no welcome screen at all, nice! I do not think
that this will create unsatisfied second time users.

We can go on forever with this "I prefer this" and "But I prefer
that". Instead I would propose a poll with GRASS users and let them
decide.

Markus M

>
> I don't want to name all changes I've done but let's take example of the
> recent ones. The message which was originally in an error dialog is now just
> a red text in the main window (r64407). In r64408 I removed the text in
> command line which required user to "Hit RETURN" which may be easy to do
> once you understand what that you should press Enter but interacting with
> command line is probably what you don't want the first time user to do. See
> old and new state of GRASS right after running `grass` command (left old,
> right new state).
>
> There is one bug in the startup I was not really looking at, yet. In
> grass.py there is some dummy rc file created and GISDBASE is set to current
> working directory ($HOME in 64 seems a bit more reasonable). This file is
> loaded in gis_set.py where value for Location (<UNKNOWN>) is used to show
> the first-time-user "create location" warning/message. However, GISDBASE is
> used used anyway to fill the corresponding input field which is quite
> confusing. Should GUI respect whatever is in rc file or should it consider
> rc file with <UNKNOWN> for Location as invalid? In other words, how to
> recognize the state of first time user when a message about creating
> grassdata should be shown (or in the future, grassdata and demolocation
> created automatically)?
>
> Vaclav
>
> http://trac.osgeo.org/grass/changeset/64407
> http://trac.osgeo.org/grass/changeset/64408
> http://trac.osgeo.org/grass/browser/grass/trunk/lib/init/grass.py?rev=64408#L502
> http://trac.osgeo.org/grass/browser/grass/trunk/lib/init/grass.py?rev=37863#L331
> http://trac.osgeo.org/grass/browser/grass/branches/releasebranch_6_4/lib/init/init.sh?rev=55970#L435
>