Sign up now! · Forgot password?
RSS/Atom feed identi.ca Twitter

Gimp for casual users

This discussion is connected to the gimp-user-list.gnome.org mailing list which is provided by the GIMP developers and not related to gimpusers.com.

23 of 23 messages available
Toggle history

Please log in to manage your subscriptions.

Gimp for casual users Andreas 20 Jul 19:44
  Gimp for casual users isabel brison 21 Jul 09:17
  Gimp for casual users Alexandre Prokoudine 21 Jul 09:56
Gimp for casual users Andreas Lemke 21 Jul 16:59
  Consistency in Selection Steve Kinney 21 Jul 17:26
   Consistency in Selection Richard Gitschlag 22 Jul 01:39
  Gimp for casual users Andreas Lemke 22 Jul 15:21
   Gimp for casual users Steve Kinney 22 Jul 15:44
    Gimp for casual users Ofnuts 22 Jul 18:22
     Cannot enter Japanese text? Daniel Hauck 22 Jul 19:47
      Cannot enter Japanese text? Liam R E Quin 22 Jul 21:20
       Cannot enter Japanese text? Daniel Hauck 22 Jul 21:52
  Gimp for casual users Andreas Lemke 28 Jul 16:02
   Gimp for casual users Liam R E Quin 28 Jul 16:07
   Gimp for casual users peter sikking 28 Jul 17:20
    Gimp for casual users Andreas Lemke 29 Jul 08:18
     Gimp for casual users Guiu Rocafort 29 Jul 09:37
     Gimp for casual users peter sikking 29 Jul 11:39
      Gimp for casual users Andreas Lemke 30 Jul 20:21
       Gimp for casual users peter sikking 30 Jul 21:44
        Gimp for casual users Andreas Lemke 31 Jul 20:12
         Gimp for casual users Andreas Lemke 02 Aug 20:34
   Gimp for casual users Andreas Lemke 28 Jul 18:51
Andreas
2012-07-20 19:44:05 UTC (over 2 years ago)

Gimp for casual users

I am assuming that Gimp is not only targeted to power users but also to casual users. Do we have any evidence how well casual users can deal with the Gimp GUI?

Gimp has come a long way but when I look at the interface, I can still identify numerous issues that appear to go against GUI best practices and make the learning curve steeper than it should be. (I am not talking about the save/export issue ;-)). I am talking of principles of simplicity, affordability, WYSIWYG, modes, consistency, etc.

For example the concept of selection in Gimp is highly complex. I doubt that beginners can quickly make sense of the many different selection types. But this is of course an empirical question.

(I have read about the good work of Peter Sikking and team at m+mi works.)

I would be interested to start some work on analyzing the existing GUI and designing possible improvements.

Any comments?

Andreas

isabel brison
2012-07-21 09:17:13 UTC (over 2 years ago)

Gimp for casual users

In teaching the basics of GIMP to non-professionals I have found one of the more problematic concepts to be the floating selection. Even to users who have no previous experience with Photoshop (where pasting something on an image automatically creates a new layer), it seems to be non-intuitive. What often happens is they paste something, and begin to manipulate the floating selection before turning it into a layer. Then an accidental click with the move tool on the background fuses selection with background, and what they thought was a layer is suddenly immobile. I always have to go over this (the difference between floating selection and layer) several times before it really sinks in.

The selection tools are comparatively easy to explain, except maybe scissors select, where losing sight of the first anchor makes it impossible to close the selection and where, unlike free select, anchors cannot be deleted by pressing back space. This, and having to click inside the line to make the selection, makes this tool unnecessarily hard to use IMO.

isabel

Alexandre Prokoudine
2012-07-21 09:56:32 UTC (over 2 years ago)

Gimp for casual users

On Fri, Jul 20, 2012 at 11:44 PM, Andreas wrote:

For example the concept of selection in Gimp is highly complex. I doubt that beginners can quickly make sense of the many different selection types. But this is of course an empirical question.

http://blog.mmiworks.net/2011/08/teaching-interaction-10.html

Alexandre Prokoudine http://libregraphicsworld.org

Andreas Lemke
2012-07-21 16:59:31 UTC (over 2 years ago)

Gimp for casual users

(Hi this is Andreas - I am sending this from a different email address)

I analyzed the following different selection types. It seems that in Gimp each tool uses its own way of making selections. Which introduces subtle inconsistencies. I am not saying they are critical, but this is a place I started my analysis.

1. The Selection a set of pixels of the image

Multi-selection: Press Shift or select mult-select icon The Select menu seems to apply only to this type of section.

2. Layers Multi-selection: Gimp uses a concept of chaining. Click chain icons to select more than one layer.
This multi-selection only works for moving, scaling, flipping etc. It doesn't work for e.g. deleting or duplicating.

3. Layer Groups When a layer group is selected and the move tool is active: Dragging moves not the layer group but the layer under the mouse.

4. Alignment tool: Uses its own selection mode. Multi-select: Press shift. No toggle action.

5. Paths There are different selection types for paths: - In the path dialog a path can be selected (grey background). Path on canvas is shown in red
- or with a blue background through clicking the path in the dialog - with the path tool active, a path visible on the canvas can be selected with a left click. Path becomes red and control points become visible. I guess I am now in path edit mode. - to select another path, I can Alt-click or select it in the path dialog So, when the path tool is activated, I can be in or out of edit mode. (Why are there these two modes?). Once in edit mode I cannot leave it without de-activating the path tool (?)

6. Path Edit Mode Multi-selection of control points: Shift click (has toggle action); drag moves the selection

I am a bit inspired by the original work on GUIs at Xerox PARC and the later work on Windows and MAC, etc. The idea there was to have a consistent selection mechanism across all types of objects. Once again, I don't think these inconsistencies are critical but fixing them could be a contribution to an overall improvement.

Best,

Andreas

gimp-user-list mailing list
gimp-user-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gimp-user-list
Steve Kinney
2012-07-21 17:26:44 UTC (over 2 years ago)

Consistency in Selection

Hey,

I changed the Subject: here because I think Andreas' comments jump out of the "For Casual Users" thread as something new, worth paying attention to.

All us users will benefit from being more conscious of the differences in "selected" behaviors in different parts of the GIMP as outlined below. I generally "approve of" the design choices involved, but where and as possible I am also "all in favor" of making these things as consistent across different functions as possible.

:o)

Steve

On 07/21/2012 12:59 PM, Andreas Lemke wrote:

(Hi this is Andreas - I am sending this from a different email address)

I analyzed the following different selection types. It seems that in Gimp each tool uses its own way of making selections. Which introduces subtle inconsistencies. I am not saying they are critical, but this is a place I started my analysis.

1. The Selection – a set of pixels of the image

Multi-selection: Press Shift or select mult-select icon The Select menu seems to apply only to this type of section.

2. Layers Multi-selection: Gimp uses a concept of chaining. Click chain icons to select more than one layer.
This multi-selection only works for moving, scaling, flipping etc. It doesn't work for e.g. deleting or duplicating.

3. Layer Groups When a layer group is selected and the move tool is active: Dragging moves not the layer group but the layer under the mouse.

4. Alignment tool: Uses its own selection mode. Multi-select: Press shift. No toggle action.

5. Paths
There are different selection types for paths: - In the path dialog a path can be selected (grey background). Path on canvas is shown in red
- or with a blue background through clicking the path in the dialog - with the path tool active, a path visible on the canvas can be selected with a left click. Path becomes red and control points become visible. I guess I am now in path edit mode. - to select another path, I can Alt-click or select it in the path dialog
So, when the path tool is activated, I can be in or out of edit mode. (Why are there these two modes?). Once in edit mode I cannot leave it without de-activating the path tool (?)

6. Path Edit Mode Multi-selection of control points: Shift click (has toggle action); drag moves the selection

I am a bit inspired by the original work on GUIs at Xerox PARC and the later work on Windows and MAC, etc. The idea there was to have a consistent selection mechanism across all types of objects. Once again, I don't think these inconsistencies are critical but fixing them could be a contribution to an overall improvement.

Best,

Andreas _______________________________________________ gimp-user-list mailing list
gimp-user-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gimp-user-list

Richard Gitschlag
2012-07-22 01:39:09 UTC (over 2 years ago)

Consistency in Selection

Date: Sat, 21 Jul 2012 13:26:44 -0400 From: admin@pilobilus.net
To: gimp-user-list@gnome.org
Subject: [Gimp-user] Consistency in Selection

On 07/21/2012 12:59 PM, Andreas Lemke wrote:

3. Layer Groups
When a layer group is selected and the move tool is active: Dragging moves not the layer group but the layer under the mouse.

That one bugs me too and I am very tempted to file it as a bug, because "select item to move" is the default setting for the Move tool. You can, however, activate the Shift modifier and then move the layer group as a single unit.

On the subject of groups versus individual grouped items, I really like the way Inkscape handles it: - Normal click selects single items relative to the current context (parent layer/group). If the clicked item is in a group, the group itself reacts. - Ctrl+click bypasses groups and selects individual items whether they are grouped or not.

Though of course Inkscape's behavior isn't necessarily applicable to GIMP....

-- Stratadrake strata_ranger@hotmail.com
--------------------
Numbers may not lie, but neither do they tell the whole truth.

Andreas Lemke
2012-07-22 15:21:18 UTC (over 2 years ago)

Gimp for casual users

From the beginning, Macs had a certain WIMP interaction style based on selection of visible objects and (mostly generic) operations that could be selected from menus or invoked through keyboard shortcuts on the selected objects. In particular, there was a set of generic operations (cut, paste, ...) that would apply to all applicable types of objects. In a file manager these generic operations apply to files. In a text editor they apply to sequences of characters. I have never used the old Xerox ALTO computers but I guess that was already the case back then.

Gimp uses a different model and I am trying to understand the pros and cons. Each type of object comes with its private operations. I delete selected image pixels with the Edit->Clear command. I delete layers with the Layer->Delete Layer command. In the Text tool, I delete selected text with a Delete command from a special context menu. I cannot use the operations from the main Edit menu for editing the text. When in path edit mode, I delete control points with Ctrl-Shift-Click. This deletes the point under the mouse irrespective of which points are currently selected. You get the idea. BTW, when I hit the Delete key while in path mode, guess what happens?

An avantage of the Mac model is that there is a lot of consistency. The user learns once that if they want to act upon an object, you select it and then you pull down a menu and click on an action. The select-cut-paste paradigm works througout, whatever the type of object. Keyboard shortcuts are always the same. But you always need two steps: select, act.

With Gimp, you do not first have to select the object to act upon. Ctrl-Shift-D duplicates the current layer in whatever mode you are in. You can duplicate a layer while in text editing mode, no problem. You can thus do certain actions more rapidly without first having to change the mode or tool.

I would be interested in knowing if Gimp users consciously prefer the model with "private" non-shared operations. Or has it evolved like this for historical reasons? Clearly, the choice of interaction model affects the Gimp learning curve.

Regards, Andreas

Steve Kinney
2012-07-22 15:44:12 UTC (over 2 years ago)

Gimp for casual users

On 07/22/2012 11:21 AM, Andreas Lemke wrote:

I would be interested in knowing if Gimp users consciously prefer the model with "private" non-shared operations. Or has it evolved like this for historical reasons? Clearly, the choice of interaction model affects the Gimp learning curve.

I think the big difference here is that different types of objects in the GIMP are fundamentally different in nature. It may be possible to make "generic" operations on objects more consistent, but there is a limit to how far this can go - lots of judgment calls to be made. The one thing I am sure will not happen, is simplification past a point where functionality is lost. That's something that happens to commercial software...

:o)

Steve

Ofnuts
2012-07-22 18:22:33 UTC (over 2 years ago)

Gimp for casual users

On 07/22/2012 05:44 PM, Steve Kinney wrote:

On 07/22/2012 11:21 AM, Andreas Lemke wrote:

I would be interested in knowing if Gimp users consciously prefer the model with "private" non-shared operations. Or has it evolved like this for historical reasons? Clearly, the choice of interaction model affects the Gimp learning curve.

I think the big difference here is that different types of objects in the GIMP are fundamentally different in nature. It may be possible to make "generic" operations on objects more consistent, but there is a limit to how far this can go - lots of judgment calls to be made. The one thing I am sure will not happen, is simplification past a point where functionality is lost. That's something that happens to commercial software...

:o)

Another difference is that you have several active objects cooperating at the same time (layer, selection, path, text input..) so it owuld be hard to tell which one get the "generic" action...

Daniel Hauck
2012-07-22 19:47:11 UTC (over 2 years ago)

Cannot enter Japanese text?

The causes of this could be all sorts of things so I do not expect other users to have this trouble at all. But here are my circumstances:

1. Running CentOS 6.3 (running GNOME 2.x and all that) 2. I have mozc installed as my Japanese input method 3. I compiled GiMP 2.8 (and a boat-load of dependencies) from source into /opt/gimp-2.8

One annoying symptom which may or may not be related is the fact that my GTK themes do not apply or carry over from my normal operating environment.

But the reason I'm writing this is that I cannot enter Japanese text into GiMP. When I open the text tool, I try to switch over to mozc and I cannot when using either CTRL-SPACE or ALT-~ (alt-grave). And I also cannot do it by clicking the IME selection thing in the systray as it says "(i) No input window" when I put the mouse to it.

I can, however, open up gedit or some other text editor, create Japanese text, copy and paste it into GiMP.

Was/is there a compile option to enable other input methods? Or GTK themes? Or for that matter command line switches or environment variables I need to set from a shell script launch routine to address either of these problems?

Liam R E Quin
2012-07-22 21:20:09 UTC (over 2 years ago)

Cannot enter Japanese text?

On Sun, 2012-07-22 at 15:47 -0400, Daniel Hauck wrote:

When I open the text tool, I try to switch over to mozc and I cannot when using either CTRL-SPACE or ALT-~ (alt-grave). And I also cannot do it by clicking the IME selection thing in the systray as it says "(i) No input window" when I put the mouse to it.

You may need to open tool options for the text tool and choose "use input editor."

Liam

Daniel Hauck
2012-07-22 21:52:16 UTC (over 2 years ago)

Cannot enter Japanese text?

On 07/22/2012 05:20 PM, Liam R E Quin wrote:

On Sun, 2012-07-22 at 15:47 -0400, Daniel Hauck wrote: You may need to open tool options for the text tool and choose "use input editor." Liam

Did that... should have added that to my original message. The results are the same.

I believe the problem is connected with my lack of theme support in that the program and libraries living in /opt/gimp-2.8 are not finding something which connects it to the rest of the UI which, I'm guessing, wants to see GNOME/GTK things and falls back to X11 things which produces the bland/generic results I am seeing.

Andreas Lemke
2012-07-28 16:02:09 UTC (over 2 years ago)

Gimp for casual users

Hi,

I have done a simple usability study with a novice Gimp user (only one, not 1000 :-). Would it be ok if I post it on gui.gimp.org? But I would need to have an account. There doesn't seem to be an option to register.

Best,

Andreas

Liam R E Quin
2012-07-28 16:07:04 UTC (over 2 years ago)

Gimp for casual users

On Sat, 2012-07-28 at 18:02 +0200, Andreas Lemke wrote:

Hi,

I have done a simple usability study with a novice Gimp user (only one, not 1000 :-). Would it be ok if I post it on gui.gimp.org? But I would need to have an account. There doesn't seem to be an option to register.

That site is more of a way for the gui team to communicate with the world than a general community site. But if you send a link to your study, even on this list, I'm sure it will be reviewed with great interest, and maybe could end up on that wiki or on the brainstorm site.

Liam

peter sikking
2012-07-28 17:20:24 UTC (over 2 years ago)

Gimp for casual users

Andreas Lemke wrote:

I have done a simple usability study with a novice Gimp user (only one, not 1000 :-). Would it be ok if I post it on gui.gimp.org? But I would need to have an account. There doesn't seem to be an option to register.

to be a usability study it needs to be a usability study.

this does not take 1000 participants, one-way mirror labs and video cameras.

but the bare minimum needs to be there:

test participants recruited from the core user groups (these are well defined for GIMP, see: )
yes, the participants can be serving their apprenticeships / be on the learning curve to the groups defined there, but on their way they must be.

six or more participants (not the more the better, above 12 really the law of diminishing returns kicks in).

beyond that a solid report with a clear test protocol description and clear outline of findings would convince me that it is really a usability study.

usability folks know that, apart from the effort that goes into recruiting, what I ask for above is really easy, practical and minimal.

when really a usability study was done or will get done, we can talk about how that can go into gui.gimp.org

--ps

founder + principal interaction architect man + machine interface works

http://blog.mmiworks.net: on interaction architecture

Andreas Lemke
2012-07-28 18:51:05 UTC (over 2 years ago)

Gimp for casual users

So here is my report - as an OpenOffice doc. I hope I can post it here. Otherwise please let me know how I can best share it.

Andreas

Am 28.07.2012 18:02, schrieb Andreas Lemke:

Hi,

I have done a simple usability study with a novice Gimp user (only one, not 1000 :-). Would it be ok if I post it on gui.gimp.org? But I would need to have an account. There doesn't seem to be an option to register.

Best,

Andreas
_______________________________________________ gimp-developer-list mailing list
gimp-developer-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gimp-developer-list

Andreas Lemke
2012-07-29 08:18:47 UTC (over 2 years ago)

Gimp for casual users

Hi Peter,

I didn't mean to imply it is a fully controlled, scientifically valid study. So let's call it an experiment. Just a single data point. As such it doesn't prove anything. But it can be the beginning of more solid work. Maybe in our community, others are willing to contribute more experiments in a true open source effort. If there is interest, I might do more myself.

Your second point is about the target Gimp user population. It is stated that Gimp is not made for those you just wish to touch up a few holiday pictures. That seems to imply that Gimp is not for casual users. My personal feeling is that that would a pity for several reasons. One of them is that casual users might grow into power users/professionals. Something like that was the case for myself. I started by touching up photos and then discovered the power of the software to do more, e.g., Web art. I was in my apprenticeship but I didn't know it at that time.

So, what does it mean: software for professionals? Clearly it means powerful tools, fine-grained control, efficient working. We cannot trade these characteristics for support of casual users. We cannot simplify the program to the point that professionals are hampered. Nevertheless, we can set a challenge for ourselves. Every professional will start as a beginner. The challenge is: keep the power and still lower the learning curve. User interface designers know that there are many subtle ways to help beginners. One way of doing this is through affordances as they have been pioneered by D. A. Norman already in the 1980s.

Does this make sense?

Regards,

Andreas

Am 28.07.2012 19:20, schrieb peter sikking:

Andreas Lemke wrote:

I have done a simple usability study with a novice Gimp user (only one, not 1000 :-). Would it be ok if I post it on gui.gimp.org? But I would need to have an account. There doesn't seem to be an option to register.

to be a usability study it needs to be a usability study.

this does not take 1000 participants, one-way mirror labs and video cameras.

but the bare minimum needs to be there:

test participants recruited from the core user groups (these are well defined for GIMP, see: )
yes, the participants can be serving their apprenticeships / be on the learning curve to the groups defined there, but on their way they must be.

six or more participants (not the more the better, above 12 really the law of diminishing returns kicks in).

beyond that a solid report with a clear test protocol description and clear outline of findings would convince me that it is really a usability study.

usability folks know that, apart from the effort that goes into recruiting, what I ask for above is really easy, practical and minimal.

when really a usability study was done or will get done, we can talk about how that can go into gui.gimp.org

--ps

founder + principal interaction architect man + machine interface works

http://blog.mmiworks.net: on interaction architecture

_______________________________________________ gimp-developer-list mailing list
gimp-developer-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gimp-developer-list

Guiu Rocafort
2012-07-29 09:37:33 UTC (over 2 years ago)

Gimp for casual users

Maybe in our community, others are willing to contribute more experiments in a true open source effort. If there is interest, I might do more myself.

What about doing a "template" with certain conditions and rules that allows anyone to repeat the experiment made by andreas ? I would be able to do one. I guess that other people is in a similar situation and would like to contribute doing just a few. With all the ones made by different people we could get a serious amount to be able to take more funded conclusions.

I'm actually not an usability expert and i'm not sure if that would be viable, but it makes lot of sense to me.

On Sun, Jul 29, 2012 at 10:18 AM, Andreas Lemke wrote:

Hi Peter,

I didn't mean to imply it is a fully controlled, scientifically valid study. So let's call it an experiment. Just a single data point. As such it doesn't prove anything. But it can be the beginning of more solid work. Maybe in our community, others are willing to contribute more experiments in a true open source effort. If there is interest, I might do more myself.

Your second point is about the target Gimp user population. It is stated that Gimp is not made for those you just wish to touch up a few holiday pictures. That seems to imply that Gimp is not for casual users. My personal feeling is that that would a pity for several reasons. One of them is that casual users might grow into power users/professionals. Something like that was the case for myself. I started by touching up photos and then discovered the power of the software to do more, e.g., Web art. I was in my apprenticeship but I didn't know it at that time.

So, what does it mean: software for professionals? Clearly it means powerful tools, fine-grained control, efficient working. We cannot trade these characteristics for support of casual users. We cannot simplify the program to the point that professionals are hampered. Nevertheless, we can set a challenge for ourselves. Every professional will start as a beginner. The challenge is: keep the power and still lower the learning curve. User interface designers know that there are many subtle ways to help beginners. One way of doing this is through affordances as they have been pioneered by D. A. Norman already in the 1980s.

Does this make sense?

Regards,

Andreas

Am 28.07.2012 19:20, schrieb peter sikking:

Andreas Lemke wrote:

I have done a simple usability study with a novice Gimp user (only one,

not 1000 :-). Would it be ok if I post it on gui.gimp.org? But I would need to have an account. There doesn't seem to be an option to register.

to be a usability study it needs to be a usability study.

this does not take 1000 participants, one-way mirror labs and video cameras.

but the bare minimum needs to be there:

test participants recruited from the core user groups (these are well defined for GIMP, see:

)

yes, the participants can be serving their apprenticeships / be on the learning curve to the groups defined there, but on their way they must be.

six or more participants (not the more the better, above 12 really the law of diminishing returns kicks in).

beyond that a solid report with a clear test protocol description and clear outline of findings would convince me that it is really a usability study.

usability folks know that, apart from the effort that goes into recruiting, what I ask for above is really easy, practical and minimal.

when really a usability study was done or will get done, we can talk about how that can go into gui.gimp.org

--ps

founder + principal interaction architect man + machine interface works

http://blog.mmiworks.net: on interaction architecture

______________________________**_________________ gimp-developer-list mailing list
gimp-developer-list@gnome.org
https://mail.gnome.org/**mailman/listinfo/gimp-**developer-list

______________________________**_________________ gimp-developer-list mailing list
gimp-developer-list@gnome.org
https://mail.gnome.org/**mailman/listinfo/gimp-**developer-list

peter sikking
2012-07-29 11:39:31 UTC (over 2 years ago)

Gimp for casual users

Andreas Lemke wrote:

I didn't mean to imply it is a fully controlled, scientifically valid study.

I am not asking for any of that. I work now 19 years in user interaction, in the software industry. all of that highly practical and unscientific, just create software that (for once) does not suck (usability) and delivers value (product realisation).

since ages I cooperate regularly with usability experts, I brief them, observe some tests, analyse the results with them. all very practical. that is where it comes from what I ask for.

So let's call it an experiment. Just a single data point. As such it doesn't prove anything.

it is disinformation. information gathered with a faulty test set-up. to call that a usability study is really not helping the GIMP project in any way. and as the vice-chairman of openUsability, I can say that you are doing the usability community a big disfavour, as this will ultimately reflect badly on them, because you are implying you are one of them.

I can deal with this, every day I have to separate information that matters from that that doesnt. but now we have not one, but two mailing lists with predominately usability and interaction design laymen that think you sent some useful information, and I will have to deal with the fallout.

if you found gui.gimp.org, why did you not asked for a briefing from me, I would have gladly given one, and discuss the test setup.

I could have told you from the beginning that you could have skipped the file dialog thing, it is a GTK thing, GIMP has no control over this. I could have told you that the floating selection is a known problem, nothing to prove anymore. next useful test of that is when there is a prototype of the new pasting/compositing design.

and that test has to be an ease-of-use test, not an ease-of- learning one. I do discuss this difference with usability folks (like when Tobias Ehni got in touch with me) and they see the point, and the challenges.

Tobias then proposed to do a new, experimental method he was studying at that moment. I completely supported him in that, just putting the boundary conditions what user insights I want out of it for the GIMP project.

But it can be the beginning of more solid work. Maybe in our community, others are willing to contribute more experiments in a true open source effort. If there is interest, I might do more myself.

I really wished you would have done the work to make this solid work. create a framework for collaborative testing, cool. but not gone public with what amounts to a single non-test.

Your second point is about the target Gimp user population.

there is a big difference between who can use GIMP (everybody, it is free software world) and who it is designed for. who it is designed for is very clear. all user research has to be conducted in these core user groups, not to be useless or disinformation.

this speaks 100% for itself for usability experts and interaction designers. zero discussion. it is usability research 101.

So, what does it mean: software for professionals?

did you actually read the vision briefing? there is a whole section called professionals? I quote the first sentence of it: Saying a piece of software is for professionals is meaningless.

Clearly it means powerful tools, fine-grained control, efficient working. We cannot trade these characteristics for support of casual users. We cannot simplify the program to the point that professionals are hampered.

here you are showing you get it.

Nevertheless, we can set a challenge for ourselves. Every professional will start as a beginner. The challenge is: keep the power and still lower the learning curve. User interface designers know that there are many subtle ways to help beginners. One way of doing this is through affordances as they have been pioneered by D. A. Norman already in the 1980s.

I would say that researching that asks for the right participants, people who are moving to GIMP (a commitment of months that is) and are aspiring to be part of the core user groups. everything else does not matter.

then there needs to be a research process that tracks them over, say, two work-weeks (i.e. 80 hours of GIMP use) what their learning curve is (that takes observation of a competent usability researcher) and how it feels to them (that takes conversation between the participant and the researcher).

that, times at least six, and well analysed, would really help.

--ps

founder + principal interaction architect man + machine interface works

http://blog.mmiworks.net: on interaction architecture

gimp-developer-list mailing list
gimp-developer-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gimp-developer-list
Andreas Lemke
2012-07-30 20:21:04 UTC (over 2 years ago)

Gimp for casual users

I will remain silent about most of what you write except one point:

Am 29.07.2012 13:39, schrieb peter sikking:

and that test has to be an ease-of-use test, not an ease-of- learning one. I do discuss this difference with usability folks (like when Tobias Ehni got in touch with me) and they see the point, and the challenges.

Clearly, there is a difference. (and it should be interesting to dig into that deeper.)

I would be interested in getting the sentiment of the Gimp community on this statement. Does it mean to imply that ease-of-learning is not important for the "target user community"? If so then, wouldn't it be better to just offer keyboard shortcuts and CLIs and ditch all the wimpy menus, palettes, dialog boxes that mainly are there for ease-of-learning - or, as I like to call it: reducing the need to learn. We could save a lot of screen realestate, and keyboard shortcuts are faster than point-and-click anyway.

No, learnability is a key quality criteria for any program. Best practices describe how this can be achieved. Sometimes there are trade-offs, but we should make these trade-offs only if all else fails.

Regards

peter sikking
2012-07-30 21:44:41 UTC (over 2 years ago)

Gimp for casual users

Andreas Lemke wrote:

I will remain silent about most of what you write except one point:

you do not have to go directly into what I wrote, but since you insist on pushing GIMP for casual users agenda here, it would really help if you come clear about your usability background.

up to now I have data points about you that are all over the place: beginners mistakes in setting up a test (and calling a single test a study) on the one hand, on the other hand: getting interaction concepts and using very learned words about usability.

I presented my background, can you be clear about yours?

Am 29.07.2012 13:39, schrieb peter sikking:

and that test has to be an ease-of-use test, not an ease-of- learning one. I do discuss this difference with usability folks (like when Tobias Ehni got in touch with me) and they see the point, and the challenges.

Clearly, there is a difference. (and it should be interesting to dig into that deeper.)

I would be interested in getting the sentiment of the Gimp community on this statement. Does it mean to imply that ease-of-learning is not important for the "target user community"?

you are painting it too black and white.

let me say that in the redesign work that is being done in my team making the learning curve more gradual _is_ a theme, all the time.

but I can also say that all the time when something gets designed or redesigned for GIMP, the trade-offs are apparent between ease of learning and ease of use and choices have to be made.

that are simply the facts presented to those who do the design work, they are very enlightening.

it is clear that the choices go the way of ease of use for GIMP, while again keeping an eye on that learning curve.

--ps

founder + principal interaction architect man + machine interface works

http://blog.mmiworks.net: on interaction architecture

gimp-developer-list mailing list
gimp-developer-list@gnome.org
https://mail.gnome.org/mailman/listinfo/gimp-developer-list
Andreas Lemke
2012-07-31 20:12:25 UTC (over 2 years ago)

Gimp for casual users

I notice a touch of respect. I appreciate that. Isn't it much more fun to deal with each other that way?

And yes, I did use the term usability study loosely. I thought that was clear to the reader and not a damage to anyone. But if not, I apologize.

Am 30.07.2012 23:44, schrieb peter sikking:

I presented my background, can you be clear about yours?

I didn't think it was necessary to present credentials before being taken seriously. But I can write about myself. There is no secret. I will do in another email.

Andreas

Andreas Lemke
2012-08-02 20:34:55 UTC (about 2 years ago)

Gimp for casual users

Hi Peter,

you asked to write a bit about myself, which I will be glad to do. I studied computer science with a focus on human computer interaction and cognitive science in Stuttgart and Boulder, Colorado. This is now a long time ago and since then I have not directly worked in the field since. Our focus at the time was on knowledge based HCI and specifically on knowledge based design environments. The objective was to empower non-professionals (sic!) to do certain design processes and benefit from design knowledge embedded in the computer environment. One of the technologies we developed were so-called critics, basically condition action rules that can trigger when a certain suboptimal design state was entered. The critic would then propose a change in the design. But these would not necessarily be executed automatically; the users should keep the autonomy to control their design process. We also represented task knowledge, i.e., structures of tasks and subtasks. From one perspective, it was the goal to reduce the knowledge necessary to do design tasks. We applied these methods to user interface design and kitchen design, for instance. For example, we were also influenced by Ivan Illich's tools for conviviality, i.e., tools to give people the freedom to engage in models of production that are different from classical industrial production.

Many of us thought that knowledge based systems would change the character of human computer interaction. We know now that this has not really panned out and it is doubtful if it will ever happen at a significant level.

[Just to avoid any misconceptions: I don't think that Gimp should be a knowledge based system]

Nevertheless, I am still intrigued by the topics of computer human interaction design and empowering people to do things they otherwise would not be able to do. I would be glad if, in my spare time, I could contribute in a small way to make Gimp an even greater tool than it already is.

Best regards,

Andreas

Am 31.07.2012 22:12, schrieb Andreas Lemke:

...
Am 30.07.2012 23:44, schrieb peter sikking:

I presented my background, can you be clear about yours?

I didn't think it was necessary to present credentials before being taken seriously. But I can write about myself. There is no secret. I will do in another email.