First time Glom usage + database question.

[albinootje]

Hi,

I felt very happy when I found out about the existence of Glom a while ago.

Screenshots look good, the website looks good, and when I tried it the
first time I was happily surprised that dealing with the database was so
easy.

Tonight I made a start with creating my own customized layout (With the
PPA packages on Ubuntu Jaunty), and that went pretty good.

I was wondering about a command-line parameter to directly connect to a
database in the network, as I find opening a connection to a running
PostgreSQL server quite confusing after starting Glom.

Would it be hard to implement a parameter for that ?
Or is there a work-around for it somehow ?

It would make the implementation of Glom in an office setup with 1 central
PostgreSQL server so much easier.

p.s.
I was manually copying the names of fields from a FileMaker database,
and I noticed that the character & gave an error, I tried the + sign
instead and that seemed fine so far.
Are there certain characters in Glom and/or PostgreSQL that needs to be
avoided, just like avoiding characters like ! ? and whitespaces is
best to avoid in directory names ?


_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[albinootje]

Perhaps I should add that everything works fine, except that directly
opening a network connection from the open file window doesn't work.

See screenshot :
http://img223.imageshack.us/img223/7637/open.png

I can connect to the PostgreSQL database later on after choosing some file,
and I can also connect fine with the pgadmin3, but not directly through
the network from the opening dialogue window.

Any idea ?


_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[Murray Cumming]

On Tue, 2009-06-09 at 05:08 +0200, albinootje wrote:
> Perhaps I should add that everything works fine, except that directly
> opening a network connection from the open file window doesn't work.
>
> See screenshot :
> http://img223.imageshack.us/img223/7637/open.png

That allows a second user to open an already-open Glom document. It is
not a list of PostgreSQL servers. I'll think about making that clearer.

> I can connect to the PostgreSQL database later on after choosing some file,
> and I can also connect fine with the pgadmin3, but not directly through
> the network from the opening dialogue window.

--
murrayc@...
www.murrayc.com
www.openismus.com

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[Murray Cumming]

On Tue, 2009-06-09 at 04:51 +0200, albinootje wrote:
> I was wondering about a command-line parameter to directly connect to
> a
> database in the network, as I find opening a connection to a running
> PostgreSQL server quite confusing after starting Glom.

Opening an existing Glom file should just ask you for the username and
password. It remembers the hostname and port. Could you explain the
difficulty in more detail, please?

--
murrayc@...
www.murrayc.com
www.openismus.com

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[albinootje]

Murray Cumming wrote:
> On Tue, 2009-06-09 at 04:51 +0200, albinootje wrote:
>> I was wondering about a command-line parameter to directly connect
>>  to a database in the network, as I find opening a connection to a
>>  running PostgreSQL server quite confusing after starting Glom.
>
> Opening an existing Glom file should just ask you for the username
> and password. It remembers the hostname and port. Could you explain
> the difficulty in more detail, please?

 I did choose both export and "save as example" because I didn't know
how to save my data,

When I then try in "Open Existing Document", then "Select File",
then it sometimes happens that it shows the "Choose a glom file to open"
window with an empty field for the Name field.
And when I choose another glom file in that dialogue it always ask for
confirmation because of overwriting it. That's the point that confuses me.

But I'll see whether I can run a Linux guest VM in VirtualBox and test how
easy it is to connect to a running glom in there.

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[Murray Cumming]

On Wed, 2009-06-10 at 04:09 +0200, albinootje wrote:

> Murray Cumming wrote:
> > On Tue, 2009-06-09 at 04:51 +0200, albinootje wrote:
> >> I was wondering about a command-line parameter to directly connect
> >>  to a database in the network, as I find opening a connection to a
> >>  running PostgreSQL server quite confusing after starting Glom.
> >
> > Opening an existing Glom file should just ask you for the username
> > and password. It remembers the hostname and port. Could you explain
> > the difficulty in more detail, please?
>
>  I did choose both export and "save as example" because I didn't know
> how to save my data,

Data is saved in the database as soon as you enter it. The structure is
saved to the .glom file as soon as you change it. There is no need to
explictly save anything. I'd welcome suggestions to make that clearer.

Examples are just ways to create new databases (and associated Glom
files), so it's normal that it asks you to choose a filename for the new
file. Again, is there any way that we can make this clearer?

> When I then try in "Open Existing Document", then "Select File",
> then it sometimes happens that it shows the "Choose a glom file to open"
> window with an empty field for the Name field.
> And when I choose another glom file in that dialogue it always ask for
> confirmation because of overwriting it. That's the point that confuses me.
>
> But I'll see whether I can run a Linux guest VM in VirtualBox and test how
> easy it is to connect to a running glom in there.
>
> _______________________________________________
> glom-devel-list mailing list
> glom-devel-list@...
> http://mail.gnome.org/mailman/listinfo/glom-devel-list

--
murrayc@...
www.murrayc.com
www.openismus.com

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[Michael Hasselmann]

Regarding the use case of saving data:

That's what confused me as well when I started to work with Glom. You
somehow expect the edit/undo/save/load paradigm (is there a better name
for that?). Well, since undo and save are missing perhaps one could wrap
everything into (database) transactions that get applied when you change
windows (or when you click "apply changes"). That would allow undos and
perhaps be more fitting to what the user expects?

Another possibility could be a flag for each row (in list view)
indicating whether the data in that row is still in a dirty state or
whether it was already made persistent. Or use a signal color (yellow,
red, ...) for the cell background while editing in the details view,
hinting at how the data is only made persistent after finishing editing
a cell. The color semantics (for a cell) could then be:
* white or blue background: persistent data
* red or yellow background: dirty data

regards,
Michael

Am Mittwoch, den 10.06.2009, 11:16 +0200 schrieb Murray Cumming:
> Data is saved in the database as soon as you enter it. The structure is
> saved to the .glom file as soon as you change it. There is no need to
> explictly save anything. I'd welcome suggestions to make that clearer.

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

[Murray Cumming]

On Wed, 2009-06-10 at 11:57 +0200, Michael Hasselmann wrote:

> Regarding the use case of saving data:
>
> That's what confused me as well when I started to work with Glom. You
> somehow expect the edit/undo/save/load paradigm (is there a better name
> for that?). Well, since undo and save are missing perhaps one could wrap
> everything into (database) transactions that get applied when you change
> windows (or when you click "apply changes"). That would allow undos and
> perhaps be more fitting to what the user expects?
>
> Another possibility could be a flag for each row (in list view)
> indicating whether the data in that row is still in a dirty state or
> whether it was already made persistent. Or use a signal color (yellow,
> red, ...) for the cell background while editing in the details view,
> hinting at how the data is only made persistent after finishing editing
> a cell. The color semantics (for a cell) could then be:
> * white or blue background: persistent data
> * red or yellow background: dirty data

This all seems too much like confronting the user with technical stuff
that he doesn't care about. There's also the risk that data would be
lost, or not visible to other users yet, if the user does not do the
step to actually save the data, though a timeout could help slightly.

There's also the problem that various features would not work if the
data is not really there - such as lookups and related records and
related fields. Avoiding problems with that would add more UI
complication and points of failure.

I am fairly happy about the instant saving because it's exactly what
FileMaker does.

--
murrayc@...
www.murrayc.com
www.openismus.com

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list

["Arq. Maximiliano" Meilán]

Agree with Murray, remember I asked the same question, but after work a
little with Glom, I find automatic save is a good choise. (undo is the
only thing I miss)
What about an automatic backup of the entire session?




El mié, 10-06-2009 a las 12:09 +0200, Murray Cumming escribió:

> On Wed, 2009-06-10 at 11:57 +0200, Michael Hasselmann wrote:
> > Regarding the use case of saving data:
> >
> > That's what confused me as well when I started to work with Glom. You
> > somehow expect the edit/undo/save/load paradigm (is there a better name
> > for that?). Well, since undo and save are missing perhaps one could wrap
> > everything into (database) transactions that get applied when you change
> > windows (or when you click "apply changes"). That would allow undos and
> > perhaps be more fitting to what the user expects?
> >
> > Another possibility could be a flag for each row (in list view)
> > indicating whether the data in that row is still in a dirty state or
> > whether it was already made persistent. Or use a signal color (yellow,
> > red, ...) for the cell background while editing in the details view,
> > hinting at how the data is only made persistent after finishing editing
> > a cell. The color semantics (for a cell) could then be:
> > * white or blue background: persistent data
> > * red or yellow background: dirty data
>
> This all seems too much like confronting the user with technical stuff
> that he doesn't care about. There's also the risk that data would be
> lost, or not visible to other users yet, if the user does not do the
> step to actually save the data, though a timeout could help slightly.
>
> There's also the problem that various features would not work if the
> data is not really there - such as lookups and related records and
> related fields. Avoiding problems with that would add more UI
> complication and points of failure.
>
> I am fairly happy about the instant saving because it's exactly what
> FileMaker does.
>

--
                  ____________________________________
                             Estudio MEILÁN
                     (arquitectura y construcción)
                  Avda. Artigas y Zabala - Piriápolis
                  Tel.[phone] - www.meilan.com.uy
                  ____________________________________

_______________________________________________
glom-devel-list mailing list
glom-devel-list@...
http://mail.gnome.org/mailman/listinfo/glom-devel-list