On 05/08/2019 10:02 AM, Jehan wrote:
On 2019-05-02 13:28, Michal Vašut via gimp-gui-list wrote:
Great report Flynn, I'm wondering why there are no reactions from the
I had started an answer long ago and never finished it/completely forgot
about it. Just did so now.
Like Jehan, I started an answer awhile ago, but I got bogged down by the
discussion of "pen/path" vs "free select" when the task was making a
rough selection around a stack of cups, and by the discussion of
"panels", which seemed to really be about dockable dialogs.
"Usability/user interface" studies are important but they aren't easy. I
think Flynn Liu did a fine job in the context of doing a class project
with limited time and resources. Below are some general comments for
people (maybe even Flynn Liu?) who want to go beyond what can
realistically be done within the confines of an introductory class
project. These comments do overlap many of the suggestions/comments
Jehan already made.
1. A huge amount of user input and GIMP developer/Team discussions has
already gone into the current GIMP user interface. There also has been
sustained input from at least one person who's profession is doing
usability studies. Nonetheless, improvements are possible and
contributions are welcome. So if someone is thinking about doing a
usability study, a good first step is to locate and read existing
discussions regarding GIMP usability. See Comment 7 below for a list of
places to search for relevant usability discussions.
2. Before beginning a study, it's a good idea to check the status of the
next anticipated release of GIMP, and ask the devs if it might make
sense to use a development version of GIMP for the proposed study,
rather than the current release. Updates on the status of the
development branch can be found in "Recent News"
(https://www.gimp.org/news/). The easiest way to reach the devs is via
#GIMP IRC (https://www.gimp.org/irc.html - use the developer channel:
3. Too often "Usability studies" lump "discoverability issues" together
with "issues with usability once discovered", when really the two types
of issues need to be kept conceptually distinct, and as far as possible
3.1 "Discoverability": How easy is it for someone who doesn't know
where some specific item or functionality is located in the interface,
to actually find what they are looking for?
3.2 "Usability": How easy is it for someone to use a specific
operation or functionality, once they've discovered its location in the
4. "Usability studies" too often fail to distinguish between three very
different groups of users. GIMP is used for quite a lot beside editing
photographs, but here I'm only considering using GIMP for photo editing
- extrapolation to other uses should be obvious:
4.1 Users who are new to GIMP and *also* new to editing photographs
- this group does include people who's only familiarity with "editing"
is applying Instagram filters and the like:
* What people in this first group of users really need is a
general introduction to image editing and to terms commonly used in
image editing. From reading "The Book of GIMP" (which explains the
basics of photo editing even as it explains how to use GIMP for photo
editing), and from comments made by people who teach photo-editing using
GIMP, my impression is that GIMP is actually pretty good software to use
when teaching the basics of image editing to this group.
* I think it is *not useful* to include in the same "usability
study" users in this group and users in groups 4.2 and 4.3 below.
However, people designing classes and writing "how to" books about GIMP
can benefit from usability/discoverability studies that focus
exclusively on this particular group of users.
4.2 Users who've never used GIMP before, but do have experience
with using other software with a comparable range of editing operations,
and so presumably are familiar with the basics of image editing:
* These users already know the meaning of terms and phrases like
"make a selection", "desaturate", and "monotone", and they have at least
a hazy grasp of things like "color spaces" and "color management".
* For this group, what the typical usability studies are
actually testing is a mix of "discoverability" and "usability once
* A study perhaps can be made more of a "strictly usability"
study by orienting users to the interface and making sure they know
where the required operations are found, before asking them to complete
a task. Or perhaps the study can procede in two parts: 1.
discoverability: find the right operation(s). 2. usability: use the
operation(s) to perform the task(s), with all participants informed of
what operation(s) to use, even if they had trouble "discovering" these
4.3 Users who have used GIMP extensively for editing photographs
(and might also have used comparable software). This third group seldom
has "discoverability" issues, and when they do, they know how to quickly
find answers. This is the group of users for whom good usability studies
could be especially helpful, and even more so if the study focuses on
particular parts of the user interface, such as:
* Improvements in Curves and Levels interfaces, now that
floating point images might easily have channel values much higher than
1.0f, and might also have negative channel values, and might also be
encoded using a variety of tone response curves.
* Improvements in the various interfaces for dealing with tool
presets and/or modifying and creating new instances of tools such as
brushes and gradients. Personally I think our current interfaces are
pretty good. But there are many bug reports/enhancement requests with
good ideas that it would be nice to explore, rank, and possibly
implement, ideally with input from usability studies.
5. Generally speaking, proposed changes in GIMP's user interface that
are intended to make life easier for users who are new to GIMP (groups
4.1 and 4.2 above) *won't* be implemented if these changes would
adversely affect the workflow of experienced GIMP users (group 4.3
above). GIMP is designed around the requirements for advanced image
editing. GIMP is *not* designed around the goal of being "easy and
obvious for inexperienced users".
6. Making a useful usability study does require that the person
designing and implementing the study already know how to use GIMP. The
more focused the study is on specific or advanced tasks, the more
in-depth knowledge of performing these tasks the person should have
before beginning even the design phase of the study. Anyone making a
GIMP usability study should already:
6.1 Be very familiar with the current GIMP interface and GIMP
terminology for the various parts of the current GIMP interface. This is
prior to and quite apart from any argument that GIMP interface should be
changed or its various parts renamed or modified. For example, the
person doing the study should understand the difference between panels
and dockable dialogs, even if the participants in the study find these
parts of the GIMP interface to be confusing.
6.2 Be very familiar with the current GIMP-specific terminology for
tools and operations. For example, when writing up the results of a
study, GIMP tools and operations shouldn't be referred to using
terminology that's not used in GIMP, such as referring to "pen" or
"textures" (terms used in PhotoShop?) instead of "path" or "patterns".
This is a separate and prior consideration from noting when and why some
participants might be confused by GIMP terminology that doesn't match
some other image editor's terminology.
6.3 Know how to use GIMP to do whatever tasks might be asked of the
participants. For anyone making a more focused, "specific task"
usability study, this might require knowing (or else being willing to
learn) more than just the basics of general editing tasks. For example,
a study aimed at improving the interface for Curves and Levels for
floating point processing requires that the person designing the study
understand some of the color science/color managment/color mixing issues
that underly issues faced by the user when dealing with floating point
7. As noted in Comment 1 above, people who want to do useful usability
studies for GIMP should be familiar with existing studies and also with
past and current discusions regarding usability. Places to look include:
7.1 Pages in the gimp.org website wiki - this is by no means an
exhaustive list - try searching for specific tools or tool groups, such
as "warp" or "paint":
7.2 Overview of a 2012 usability study conducted for GIMP:
7.3 Various gui.gimp.org website pages that were generated as part
of the above 2012 usability study - this is not an exhaustive list -
other pages can be found by following the links on the pages given below:
7.4 Archives for the GIMP developer, user, and user interface
mailing lists. Links to archives can be found here:
7.5 Open and closed GIMP bug reports regarding usability - search here:
7.6 Forums where GIMP users and members of "the Team" hang out,
including (in no particular order and with no particular implied
8. Regarding the list of resources in Comment 7 above:
* There is no reason to assume that what's already on the
gimp.org website or wiki must be taken at face value/used as starting
points/not be challenged/is set in stone/etc. Referring specifically to
the 2012 professional usability study documented under links given in
7.2 and 7.3 above: Even though I'm sure that study was helpful,
personally I have a lot of reservations about various aspects of the
study's design, notes, implementation, and conclusions. Also, GIMP code
and provided functionalities and interfaces have changed considerably
* For many of the resources, and especially for the bug reports
and user discussions in forums and mailing lists, finding what's already
been discussed requires being creative in coming up with search terms:
Relevant discussions and bug reports might not - in fact often don't -
use the term "usability".
9. Once you feel comfortable with GIMP's current interface and
terminology, and with the various editing operations that seem relevant
to your study, and you've absorbed at least part of what's already out
there on the topic of GIMP usability, then a next step might be to draft
a preliminary study and ask for input, maybe by making a post to this
GIMP mailing list or a user forum, or on GIMP IRC. If users and devs see
that you know how to use GIMP and are well-informed about past
discussions and current issues, they might be inclined to provide input
for improving your proposed study.
Color management and free/libre photography
gimp-gui-list mailing list