time to stick a fork in the gimp?
There’s this article over at newsforge that I cannot resist to comment on. It raises some good points that are worth to be considered. But it also makes some statements about GIMP development and the openusability.org effort in particular that show that the author has fallen upon misunderstandings that should be corrected.
Let me address the good points first. The article says:
The interface seems to irritate users more than the software’s missing features.
The author has a point here. GIMP has some serious shortcomings on the feature side that need to addressed. We are missing some much-needed features such as a decent color management, deep colors and also CMYK and spot colors. There’s a lot of work to be done in those areas. But it also important to make the currently available features more visible and more usable. A feature that is not discoverable is as bad as a missing feature, to the user it simply doesn’t exist. Of course we can improve this by improving the documentation, but in general it is safe to assume that users don’t read any manuals.
We are already trying hard to improve usability of the existing functionality and I think we have done a significant step forward in the 2.3 series. A significant step that was achieved by doing lots of small changes all over the place. But the newsforge article asks for a more radical approach, a user interface redesign. Even though I favor incremental changes, I think we can only benefit from attempting to do such a redesign.
A while ago I started to look for usability experts that can help us with this. If we want to come up with a design that attracts enough developers so that it can actually be implemented, we better make sure that the design is well-done and based on research instead of personal opinions. We need to approach the problem on a professional level. It doesn’t make sense to blindly do what the users are asking for. First, there are too many different voices, and second, the users cannot think of anything better than what they already know. So they will always ask for an immitation of the programs they are using or have been using.
The openusability.org platform hasn’t attracted much experts yet. There were some good proposals but you don’t start a redesign by doing dialog mockups. We first need to find out what we want GIMP to be, we need to define what our target users are and we need to make a clear statement about this. Recently some progress has been made towards this goal. Work has started on a collection of persona that can serve as a vocabulary for defining our user base. So I strongly disagree with the article talking about the dissappointing end of the GIMP redesign initiative. This initiative is only just starting.
Next weekend we are all coming together at the Libre Graphics Meeting in Lyon and I am very happy that Peter Sikking from man + machine interface works followed my invitation to the conference. I know Peter from the regular meetings of the local usability interest group and I am looking forward to accept his help to (re)define GIMP.
GIMP still has a long way to go and everyone is invited to participate.
March 13th, 2006 at 10:23 am
You’ll like these:
http://glazman.org/weblog/dotclear/index.php?2006/03/11/1616-the-gimp-s-lack-of-ui
http://glazman.org/weblog/dotclear/index.php?2006/03/13/1620-the-gimp-s-lack-of-ui-2
March 13th, 2006 at 11:06 am
glandium: I already saw this but I will certainly not respond to it. The authors seems to like to shout at his readers and I refuse to have a discussion on such a low level.
March 13th, 2006 at 11:36 am
Reading your blog post I was about to consider returning to Open Usability and seeing what was happening.
But your reponse above shows me nothing has changed.
Everything Glazman says is right, but out of arrogance you stick your head in the sand. Wake up Sven, find some humility, and let’s start fixing GIMP. Or move over and let someone take over who can actually organize things and work with people without pissing them off.
March 13th, 2006 at 11:43 am
Where am I sticking my head in the sand? With the exception of a few technical details, Glazmann hasn’t said anything that I haven’t already said myself. He is obviously mistaken in his assumption that the GIMP developers would be unwilling to do any changes. But he is ranting and he is shouting and I simply don’t accept this as a form of communication.
March 13th, 2006 at 11:45 am
Sven,
you rock
March 13th, 2006 at 12:19 pm
Those “personas” are completely bogus. The second in particular is a loathesome hug-fest of the status quo.
Creating use cases is not just a matter of deciding how many cats Claudia has or how “technically minded” she is. With a start like that the page is going to go nowhere. I’m tempted to wipe that whole page and start with something sensible.
– Chris
March 13th, 2006 at 12:36 pm
good thing your reaction is positive. The problem though (as I see it) is not talk that has to be done, it’s action and it should happen ASAP. And the UI should mimicPhotoshop UI as Photoshop is defacto standard for the userbase that Gimp targets.
This doesn’t mean that Photoshop has the best UI, but it’s standard so everybody (in the area) knows it
March 13th, 2006 at 12:40 pm
Chris, the personas Wiki page in its current form is a start, nothing more. Noone said that these are good, it’s a start and it will need to be reviewed. The people who are doing this are not user interface experts. Since I have been doing my usability homework for quite a while, I know very well that the personas aren’t as well-done as they could be. But I am confident that we can improve this.
BTW, how technically minded a persona is, is a very important aspect of the persona definition. It is also widely accepted that it does make sense to add some personal details and even small stories to usability personas.
And there is a difference between use cases and personas. We are not creating use cases here.
March 13th, 2006 at 12:46 pm
Nikos, as long as the GIMP development team consists of less than a handful developers who all happen to have a fulltime job and a real life. radical changes are unlikely to happen in a short time frame. That’s why we are currently trying to focus on defining our goals and identifying tasks that can attract new developers.
Do we actually want to target the Photoshop userbase? You are obviously assuming this and I wonder what you are basing this assumption on.
March 13th, 2006 at 1:03 pm
What are you creating, then? Surely if you’re going to try to establish who currently uses the GIMP then the best thing would be to poll actual users and not fake ones.
The second persona simply needs deleted. Aside from being so obviously made up to be perfect (Knuth is not only a mining engineer, he’s a *professional* photographer. What, in his spare time?) it’s hard to see what purpose “Knuth is really happy with the GIMP UI. He’s only concerned about missing capabilities like meta data management or high color bit depth” serves when looking to improve things. If there are enough Knuths, does the Gimp’s OpenUsability project simply shut shop and declare victory?
Are you deliberately being ambiguous here because you personally don’t think the GIMP is the same kind of application as Photoshop or because you’re simply unwilling to commit to stating a direction?
– Chris
March 13th, 2006 at 1:08 pm
Hi,
Personally I would prefer GIMP’s UI to be more photoshop-esque and would like to see major changes. However the main point I want to make is to suggest that you to write all of the GUI code in a scripted or interpreted language (i.e. java, python, C#). I guess the core should remain c-based (unless adding CMYK and whatever required large rewritten).
I believe this will encourage a lot more developers to look at contributing to the gimp.
Adam
March 13th, 2006 at 1:12 pm
Chris, perhaps you want to check the links at the bottom of the personas page to learn what personas are and how they are different from use cases.
Yes, I am deliberately ambiguous about the goals of GIMP because I am currently just collecting input from various sources and will definitely not commit any statements before next weekend’s conference.
Personally I do believe that it would be a major fault to define GIMP as “the same kind of application as Photoshop”, because that would only push the problem to the question “What kind of application is Photoshop?”. We need to define our goals but we should by all means avoid to define our goals in terms of competitors.
March 13th, 2006 at 1:17 pm
Adam, there hasn’t been a single developer yet who has expressed his or her serious interest in working on GIMP and who would have asked if it is possible to use a different programming language. So I don’t think it makes sense to throw away two third of the GIMP code only because you think that it is written in the wrong language.
What I definitely consider, even though it would be a lot of work, is to use libglade for the UI. That would make modifications to the UI a lot easier.
March 13th, 2006 at 1:25 pm
A persona is just a step back from a use case. The serve exactly the same purpose, just at different levels.
Why? Free Software has been defined by its competitors since day 1. The GIMP is no exception, and to be honest I think the current identity crisis has largely been manufactured.
– Chris
March 13th, 2006 at 1:34 pm
> Creating use cases is not just a matter of deciding how many cats Claudia has or how “technically minded” she is.
Actually it is important how technically minded users are, since usability is about people, not robots
As for Knuth, I don’t quite understand what would prevent a 40-50 years old man who holds camera in his hands for last 20-30 years from making professional shots being an engineer.
Sure, personas can be described better. The ones for F-Spot are good.
March 13th, 2006 at 1:51 pm
It is important for the author to know how technically proficient the user is. It is rank amateurism to then go and write “Claudia is slightly technically able” or “James is very good with computers”. It doesn’t describe their day. it doesn’t explain their actions. It’s a get-out by a lazy or inept author.
The page also describes him as spending his spare time installing Slackware and writing Javascript apps. It might as well go on to say he rescues cats from trees and volunteers as a fireman at the weekend.
Those are use cases. They’re good because, unless I’m mistaken, they were written by a professional.
– Chris
March 13th, 2006 at 2:01 pm
Chris, if you would like to follow my advice and actually read a little bit about personas, you would realize that most usability books recommend that a persona definition includes some personal details and even stories like rescueing cats from trees. Why do you insist on ranting about the work that other people have done with good intentions? The page you are commenting on is a Wiki. You could as well go ahead and add your idea of a GIMP persona. At some point we will then go and pick a few of the personas which are well-defined and use them in order to create common use cases.
March 13th, 2006 at 2:31 pm
I think you really exagerate.
the gimp ui is not “that” bad and it is already an imitation of photoshop.
the work is not to improve Layers management for example
to improve plugins menus and dialog boxes
to improve undos function (a way to save incremental modifications or a “tree” based undo )
in the same way I think also it’s BAD to try to be a perfect copy of photoshop ui
photoshop ui is already a mess (cs2 , too many hidden stuff) and IT’S IMPOSSIBLE to perfectly copy a windows/macintosh software in a ‘gnome” context, even if GTK is available in native windows api.
and for real works, a “just little” modifications is a pain, it brings chaos and misunderstandment.
Gimp developpers should only to strive to do a _good_ gui not to simply copy-carbon one.
to imitate ? to be inspired ? yes
to copy-carbon adobe ? no
gimp 2.4 will bring many good improvments in the ui
of course, there will be a gimp 2.6 , 2.8, 2.10 , heck ! even a 3.0 one day, so the work will continue
March 13th, 2006 at 3:05 pm
GIMP UI *is* THAT bad.
If I asked a monkey on crack to design an interface, I’d get better than GIMP.
Sven, you appear to come out with the best of intentions, but from experience, it’s impossible to work with you.
The tragedy of the OpenUsability project was that the usability people and the devs were unable to communicate, and you in particular were unwilling to accept direction and prioritizing from people who are not coders. By and large, you drove them away.
March 13th, 2006 at 3:26 pm
Joachim, first of all, please try not to shout in here. Then, you have obviously misunderstood what the openusability.org effort is all about. It is supposed to take away some work from the developers. You aren’t supposed to give anyone directions or prioritize anyone’s work. That’s not how you deal with volunteers. If you want to see things changing you need to persuade people that your ideas are good and you need to motivate people to actually implement them.
Unfortunately most of the postings on the usability forum were just the usual rantings that we all have heard more than enough. If you start your argumentation with the words “GIMP UI is that bad” in big capital letters, you cannot expect that someone who has put a large amount of his/her free time into implementing this UI is going to listen to you. If you want to work with the developers, then don’t tell them that they suck but try to make them aware of usabilty problems. Show them were the UI gets into the way of common workflows. Show up alternatives. Do some serious usability research.
What we are looking for is advice from usability experts, not advice from users. Users are as bad interface designers as developers are, especially when it comes to an application as complex as an image manipulation program. If we would listen to everything our users are telling us, the GIMP UI would be even more cluttered and inconsistent than it is already.
March 13th, 2006 at 3:35 pm
Out of curiousity, just what amount/nature of work are you talking about here? Because it seems like whatever it is these masses want (I personally have no issues whatsoever), this seems like a useful first step. Then the not-so-technical designer sorts should (potentially) be able to use glade to easily design/try out whatever it is they want.
Or better still, gimp could read different xml files on startup to appease different audiences. “Classic” for the long time users. “New” for every one else complaining.
March 13th, 2006 at 3:53 pm
I’ve gone ahead and Been Bold (to quote Wikipedia) and tidied the two existing personas up. I’ll add to it later, when it wouldn’t look quite as obvious that I’m bunking off work.
– Chris
March 13th, 2006 at 3:58 pm
Just for the record, it’s precisely this kind of thinking which kills good UI. pundit’s whole post is so emphatically wrong (end users playing with glade. that’s, just, wow) as to almost stereotype the kind of thinking that needs to be avoided.
– Chris
March 13th, 2006 at 4:25 pm
The day the UI becomes a clone of Photoshop’s is the day I disappear as a user.
Just remember that complainers are very vocal, and those of us who are actually sitting down and doing real work with the GIMP day in day out (and have actually bothered to get used to the UI before beginning a flamefest) are busy getting stuff done.
Deep paint support and grouped layers is more important than some sort of arbitrary ill-defined unending ‘UI discussion’.
March 13th, 2006 at 4:52 pm
Personally I would stay away from the Glade thing. Big time.
What we need is to actually plan this. We can’t make an OpenUsability project and then just leave it to the developers to sort it out, or leave it to the posters to convince the developers and so on. How are they going to “convince” them? I really don’t think some Wikipedia-style “collective intelligence” techo-utopian system is the way to go here. No Glade. No XML “Classic” modes and such.
Some kind of leader, moderator to sort things out would be prudent. Listen to all the rantings, ravings and sort it out so the programmers don’t have to read “omg gimp sux0rz wtf” rants all day. And, more importantly, so the developers don’t need to think about UI design. If some time could be set aside for the project to focus on implementing the design spec – once completed, of course – then get back to work on CMYK and all that jazz, that’d probably be most efficient.
Simply put, the programmers shouldn’t be worrying about how the designers are doing things. Are the designers counting cats? Who cares? There should be someone who can sort it all out and give the programmers a real design spec.
That said, I’ve gotten non-technical Windows users to try Gimp, and I have recieved VERY different opinions than you typically get from geeks.
Very often from geeks I hear them saying that they want MDI or something – or, at least some implication that MDI is inheritly easier somehow.
I have never heard that from non-technical users. What I have heard are things like this:
*Not all the dialogues behave consistently. Some don’t have previews, some filters don’t even have dialogues, sometimes not all the previews work the same…
*Not all the tools function the same.
*”How the hell do I draw a line?” is a common one. In fact, that one stumped me for years.
No, people do not RTFM to draw a straight line. I can understand a manual for explaining how to do some weird thing with channels, but not to do very, very basic things.
Now, On that last part, I can see exactly how this is organized. The idea is the abstract the tool from the line process. Programmer’s line of thought.
Personally, I would just make a paint tool and a line tool. The pencil tool is just the paint tool with anti-aliasing disabled. Just put that in the tool options for both the line and paint, label it “smoothing” or something to that effect and you’re set.
*Then there’s the beizer curves. It used to be that I could set fuzzy selection on the beizer curve select, but I can’t here in 2.2. (I’m looking at version 2.2 here, by the way.) Why doesn’t it behave like the other selection tools?
Those are the sorts of inconsistencies that drive people out of their minds. Not cosmetic things. The jist of it is that things either aren’t thought out, or are thought out in a fashion totally inconsistent with how non-programmers think about things.
I’ve heard good things about 2.3 but I’m really not much of a beta junkie these days so I’ll have to wait and see.
And let’s stay the shaq away from any XML, glade, collective intelligence Wiki style editing scheme for the frickin user interface. That will make this problem much worse than it is already, and at best simply complicate the code. There needs to be direction and specs.
March 13th, 2006 at 5:20 pm
Joe, the proposal to use glade was actually just made because it could make things easier for developers. Nowadays people don’t want to code user interfaces in plain C and by using glade we could make it easier to create nice and consistent dialog layouts. I never meant to propose it as a tool for users.
March 13th, 2006 at 6:04 pm
Just because Photoshop is the “industry standard” does not mean that the Photoshop interface design is the best possible. I used the GIMP and several other image manipulation programs for many years before I ever saw Photoshop. When I tried Photoshop, I was very happy that I had not paid for it, because I found the interface design to be among the worst I had ever seen. Becoming accustomed to a bad design does not make the design good. Examples of this can be found everywhere.
The QWERTY keyboard, cigarette lighter plugs for 12 volt accessories, and the Mac OS menu bar are all examples of design artifacts which continue to plague users, and yet are unlikely to change, largely due to inertia. Once a standard settles in, no matter how bad it is, or what different and better alternatives are developed, the bad standard will tend to linger on towards eternity.
My personal favorite is Autocad. Autocad was first written for DOS (remember that?), and the original interface design was done for the hopelessly limited capabilities of that “operating system”. When better interface systems were developed, the Autocad interface was held back from improvement _by the users_, who had so much time invested in learning the bad interface that they couldn’t bear any significant changes. Autocad also appears on the Novell list.
From the listings, it does not follow that users want the interface designs of either Autocad or Photoshop, except due to the usual resistance to change. What users want is the functionality first, whether they know it or not.
Free software has the advantage of not needing to bow to the pressure of conformity which afflicts commercial products. Instead, free software is free to evolve the best possible interface design, and that is the path that should be taken.
March 13th, 2006 at 6:48 pm
“Joe, the proposal to use glade was actually just made because it could make things easier for developers. Nowadays people don’t want to code user interfaces in plain C and by using glade we could make it easier to create nice and consistent dialog layouts. I never meant to propose it as a tool for users.”
Oh, yowza. Now I feel stupid. Misunderstood. I’m sorry about all that.
March 13th, 2006 at 7:33 pm
I have no interest in a Photoshop clone, I’m perfectly content with the GIMP. That is not to say that the GIMP is perfect; there are, as Joe Cooper pointed out, inconsistencies, and we, as a community, need to work to make the interface as consistent and intuitive as possible. With that said, I don’t believe that it’s possible to create an interface that makes hard stuff like layer masks and non-destructive editing (using layers) obvious. I have had to invest significant time in learning how to use the GIMP, not because of its user interface, but because editing photos as the professionals do it, is a skill that needs to be learned.
As a user, I would like to see support for procedural layers (adjustment layers), memory consumption optimization, greater speed (Photoshop’s filters are, let’s face it, faster) and layer groups (hierarchical ordering of layers). I’m actually quite fine with the user interface — I have dual screens, and use my laptop LCD to display panels (actually, the entire screen is filled with just panels), and my CRT screen for the actual image. The GIMP’s flexibility allows me to do this easily.
Speaking of the user interface, there is something that Mac Photoshop does that would be handy for GIMP users that, unlike myself, have only one screen — when maximizing an image window, the window managers avoids obscuring the panels (that is, the window isn’t actually maximized to fill the screen).
There are lots of great little touches in 2.3.x, such as the ability to sample all layers when cloning — I used to duplicate the image, flatten it, and use that as the clone source to achieve this. 2.3.x is coming along nicely, but obviously, more developers are needed, especially to tackle the color management issues.
I use The GIMP 2.3.x with UFRaw (and, as a result, don’t have as great of a need for 16-bit per channel editing as I once did; I make all the major adjustments in UFRaw in 16-bit mode, and then fine-tune in the GIMP). It really is a great combo, and it’s awesome to have access to this functionality in Linux.
The GIMP isn’t perfect (yet), but works well, and with the help of more developers and usability people, will get even better.
March 13th, 2006 at 10:59 pm
(Emphasis mine.)
Exactly.
Neither did I. It just provides a mechanism to actually go ahead and make a valid change which might otherwise get shot-down for lack of time on the developers’ part, or knowhow on the designer’s part.
March 13th, 2006 at 11:26 pm
Just to say that I can live with all gimp missing features (16bits, color management, active layers…), thanks to its wonderfull UI. Please, improve it, but do not clone Photoshop one. Using Gimp in a dual-head config is great; using Photoshop in such config is not.
In my photo club, I pushed a few newbies digital photographers to use Gimp, instead of Photoshop Element. They where first disapointed by the UI, but now enjoy it. The learning curve is really better than the Photoshop one. I know it, because newbies using Gimp now do much more things than newbies using Photoshop.
March 14th, 2006 at 1:23 am
Not here to bash anyone or anything for that matter but some of the people here really need to be considerate to the developers. I can understand the frustration of many non-technical users (newbies to linux) with gimp and its floating dialogue boxes which reveal the desktop. Many loyal users will try to tell others that you should use the virtual desktops to solve the problem but frankly that’s not enough. Not everyone uses the virtual desktops and the ones that does, uses it very differently than others. It’s irratating to have the brush window in one virtual desktop and the image in another and switching back and forth. It’s alot more convenient to have it all in one area.
No one here want’s photoshop or its functionality. We all want GIMP to be better.
I love inkscape. I love it’s UI. Maybe not the floating dialogue boxes sometimes which get in the way but overall its a awesome program. It would be nice if GIMP had a similar UI but with better dialogue boxes. Maybe something that can snap to the side like 3dstudio max.
I’m not a serious GIMP user because of my frustrations with it. I don’t blame the developers because I know they are busy with their lives and all. I kinda blame myself for not having enough time to learn C or whatever so I could make modifications myself. But I hope the GIMPs UI is changed for the better.
Remember most people don’t start up using computers with linux. They use windows and then convert to linux. So it helps if people can transfer their skills to linux without needing to re-learn new ways of doing old things.
-just my 2 cents.
March 14th, 2006 at 10:09 am
First of all, I like Gimp the way it is. It works for me, and a lot of other people propably.
And I don’t believe that Gimp can be everything for everyone. Different people need different applications, because they have different needs and different approaches to using software. No matter how hard you try, you could never please everyone.
I think it’s more important for the free software community to provide a number of different apps for different styles of usage, than to cramp every possible UI paradigm into one pice of software.
-Richard
March 14th, 2006 at 12:09 pm
It’s nice to see the discussion that has evolved here, because it shows the diversity of opinions that people have on this subject. Thank you for your input and please excuse that it took a while to approve your comments.
March 14th, 2006 at 4:25 pm
I wonder what Aperture and Lightroom ought to say about where the future of image editing ought to lie. Both have non-destructive editing, and both have new — much better — user interfaces. I’d hope the GIMP community would consider this.
March 14th, 2006 at 4:53 pm
This is just an idea, but it would be handy if the GIMP could have an interface that’s switchable… kinda like Corel Photo-Paint or such, meaning you can change the layout to be e.g. like Photoshop, Photo-Paint native, etc. Would probably stop most of the complaints of users who can’t get used to using right-click for a menu instead of searching a menu bar above image that’s being edited.
More useful changes would be support for native CMYK, LAB and 48-bit RGB images (afaik, the GIMP kernel has been made for 24-bit RGB images only, initially), and revamping some tools so that they offer more functionality or are easier to use with connection with other tools.
March 15th, 2006 at 4:59 pm
I might as well make a GUI contribution here (following development mailing list for some time now).
I look forward to the full-screen editing functions in the Gimp. It kinda gives the image the central role, which it should have. The application (GIMP) could facilitate this nicely.
March 15th, 2006 at 4:59 pm
@Bruce
Aperture is unusable for many people because of its design flaws and hardware requirements. Lightroom seems to be interesting, but is currently available as beta for Mac OS X only. Picasa works on WIndows only, has nice effects (I like its soft focus and b/w focus effects a lot), but desperately needs decent white balance adjustment tool.
I actually like the concept of non-destructive editing with effects chain. The problems is that when it comes to retouching photos you often need painting actions with brushes and this cannot be easily recorded as undoable/redoable actions (there is a mouse events recorder/replayer however).
March 15th, 2006 at 6:24 pm
Bruce:
I think Aperture/Lightroom are really a different thing altogether and we should perhaps be looking at f-spot/digikam fulfilling that role.
March 15th, 2006 at 10:44 pm
I am so only ranting and shouting that last August, I tried to get Linspire sponsor the Gimp in order to reach a better experience and make it a free app for linux at least comparable in terms of usability to other products on other platforms. David Neary can easily testify about that.
I involved in the loop directly the CTO and the CEO of Linspire. Raphaël Quinet’s answer was simple : we have this OpenUsability thing and you can join, MDI is a firm no-go from Gimp developers and it is seen as, I quote, a waste of developer’s time. This was extremely disappointing, I was extremely disappointed, Linspire was it too.
I find it counter-productive when I need someone else to find in the Gimp’s UI a trivial feature that I find alone in half of a second on PaintShopPro 5 on windows, a software for newbies that does not have one tenth of the power of the Gimp.
Now, this is true I do prefer saying things exactly as they are, without any form of political correctness. The guys at Newsforge were not less harsh, and their article, combined to the feedback of the crowd, should ring a bell Sven…
About PS : no you DO NOT have to target PS user base. And Nvu never targeted Dreamweaver use base (this is _not_ irony). That’s just a side-effect of the existence of Nvu that it now competes with it, just like the sole existence of the Gimp makes it compete with PS. Like it or not, whatever you do, your product live on the same market segment. PS target user base IS your target user base. That’s just factual.
March 16th, 2006 at 12:03 am
Robert — I’m not sure. I think that current image editors do not support non-destructive editing will, in say 10 years, be seen as positively archaic. The open source communtiy really needs to get on board this trend.
As for the interface, I also am unconvinced. As someone with a background in professional photography, I really applaud Apple and Adobe for finally breaking the Photoshop UI mold. That they both happen to focus on RAW images misses the bigger point, which is that they are really designed for photographers (and professionals at that).
I’m not saying the GIMP ought to just copy this approach, but I am saying it’s an important development worth studying closely. Given all the complaints about the GIMP UI, it seems all them more important.
March 16th, 2006 at 1:02 am
Alexandre: It is possible to use the non-destructive editing techniques even when you need to clone out bits and pieces of the photo. Simply create a new fully-transparent layer above the one you intend to modify, select the clone tool, tick the “sample all layers” check box (new in the 2.3 series), and clone away! The original layer won’t be affected, but the result will be the same as if you directly modified the original layer.
As GIMP currently lacks procedural/adjustment layers, curve adjustments and the like do require you to modify the original image. When I’m ready to make a destructive change, I usually save the image as, for example, “IMG_XXXXX-pre-curves.xcf” before I continue.
March 16th, 2006 at 1:13 am
I’ve seen Adobe Lightroom’s interface in action and it is quite impressive! I’m pretty sure this is one piece of software that will be very popular among pro photographers. But just like Lightroom will not replace Photoshop, I believe The Gimp should not try and mimic what makes Lightroom so special. If anything, I would expect f-spot to compete with Lightroom. And I don’t expect f-spot will ever replace Gimp.
March 16th, 2006 at 2:44 am
Next time you upgrade or switch blogging software, realize you might spam whatever planets subscribe to your rss feed
March 16th, 2006 at 6:53 pm
The plan to turn GIMP into a non-destructive image editor dates back to the year 2000 at the GIMP Developers Conference in Berlin. But sadly it turned out that we are lacking the developer resources to implement an alternative GIMP core that would allow us to implement these features.
March 17th, 2006 at 3:42 am
I really don’t understand this rage again Gimp’s UI, when i’m looking for a different soft, i don’t want a copy, i want a different think, nevertheless Gimp isn’t so different for being frightened by it… in fact Gimp is pretty sane/logic.
Just know how to take it.
I’m using mainly shortcuts coupled with floating menu in Fullscreen mode thus about the GUI, i don’t even “see” it
The only flaw i can note in Gimp-UI, this is the “Color menu” in Filter section, that’s it. eheh
Please gimp’s Dev, just take care on performance and refinement of current tools, let’s hear with only one ear these ungratiful and lazy – mainly win32- users cause, in fact, they shout more than they show.
Belal: I don’t agree with you, everyone have to make an effort… they want to change, make an effort for THAT, there’s not only one workflow in this world, damn!
About your exemple: “Brush dialog” ?
During your drawing just push shift+ctrl+b (or make your own hotkey) and tada!
What? Oh yeah, how i know this hotkey?
I have just watched it beside the name of the function into the menu “Dialog”. I don’t think i am a genius.
Regards.
March 17th, 2006 at 11:55 am
David Polberger:
That sounds reasonable. However, that layer should automatically be added as soon as you start painting.
Sven:
Is GEGL part of that plan?
March 18th, 2006 at 1:03 am
I’m one of the usually quiet user who like The Gimp’s UI very much! I just can’t imagine using a MDI interface. I just like The Gimp Way and the improvements it got in the 2.x series.
The only thing I’m really missing is more than 8bit capabilites. That might be the reason to switch to another program, even if the UI is not as good as The Gimp’s one.
So please don’t go astray and implement a Photoshop like UI! Improve the current one and implement important features (as I already said 16bit capabilities for one and adjustment layers would be nice too). Thanks for all the great work.
March 19th, 2006 at 6:45 pm
Belal: I don’t agree with you, everyone have to make an effort… they want to change, make an effort for THAT, there’s not only one workflow in this world, damn!
Exactly. There isn’t one workflow so what you find alright I may not.
About your exemple: “Brush dialog” ?
During your drawing just push shift+ctrl+b (or make your own hotkey) and tada!
What? Oh yeah, how i know this hotkey?
I have just watched it beside the name of the function into the menu “Dialog”. I don’t think i am a genius.
Masha’allah.
March 22nd, 2006 at 12:10 pm
It may look like a Troll but I think somebody has to say this.
Most people who have been doing serious bitmap work have been doing it with photoshop. After years, the interface but especially the keyboard shotcuts and the keyboard-mouse combination actions have become like reflexes, we do it without thinking, they are like wired-in to our hands.
Getting to learn a new interface (be it better) is a step most of us don’t have the time or the motivation to take.
So working on a new redesigned interface that is not a mock-up of adobe is a great idea, but proposing alternatively a “classic” interface for those who have been broken in by photoshop would certainly help the Gimp make its way into the stronghold of those of us who have known Adobe for a while.
March 22nd, 2006 at 5:31 pm
Gibouille: GIMP ships with optional Photoshop keybindings for exactly that reason.
March 23rd, 2006 at 12:26 am
Gibouille: I don’t agree. QuarkX is different from Indisign, Maya different from XSI or Houdini etc etc… however, digital artists have to learn those interfaces and philosophy and that poses no problem cause the fact those soft are build on ~heavy financial investment, this is their warranty, but it isn’t a truth. But as OSS are build on free time, with voluntary, people are dubitative.
Also, i don’t know one real artists using the default map/setting in any software, often they adapt their tools.
It’s obvious, Pro who are working with Photoshop, get no interest to use another software aiming at the same field and “less” better but “our” aim is not to convert photoshop user, they don’t make Gimp for them, they fill a lack and provide a different – working well – thing for us, and just for that, they are the best
Users have to make effort, that’s it…
March 24th, 2006 at 4:23 pm
Interesting reading from NewsForge : “Switching art students to GNU/Linux”
http://business.newsforge.com/article.pl?sid=06/03/09/2238246
Excerpt : “For digital painting and image editing tools to replace Photoshop and Painter, the GIMP was the obvious choice.”
March 24th, 2006 at 8:33 pm
I don’t know why amost everyone consider “learning” as a bad thing.
Gimp is different than Photoshop, you have to learn a couple of things, then you realize that they have similar functions, but maybe they are in different places or under different names.
The first time I ran the Gimp, it looked awkward to me. It’s true, it was very different and I was more comfortable with Photoshop. But it only took a couple of days customize some preferences, and everything changed.
I’m currently using Photoshop AND the Gimp, and I’m comfortable with both.
Photoshop is great, its interface isn’t bad at all and it has some features that Gimp hasn’t.
Gimp is great, its interfase isnt’t bad at all and it has some features that Photoshop hasnt.
They’re different, that’s it. Why is it so hard to understand?
Exactly the same thing happened to me with Blender… Clearly it’s not 3D Studio Max (thanks God!
I agree with Sven: There are lots of things far more important (CMYK, Lab, Color Management, high color bits) that keep Gimp out of the professional level and should be done before some interface tweaks.
As a graphic designer, I’d love to have those features right now, but I understand that Gimp is free software, developed by a community of people working hard without a big corporation supporting them, so it obviously goes in a slower pace.
The whole UI thing should be discussed in other terms. Not to make it a carbon copy of Photoshop nor replace it for a completely different one. We should focus in suggest changes and improvements to the current UI.
There are few things that could lead to a better user experience: (in my honest opinion)
- Non intrusive numeric transform type-in windows (currently the scale, rotate and skew windows overlap the workspace)
- When applying an interactive transformation, make the original layer semi-transparent instead of opaque, to make visible the “before and after” of the transformation (having both opaque is a little bit confuse)
- Make the tools and panels persistent on top by default (It is currently possible by setting the windows preferences, but it’s not the default and this is what most people seem to consider annoying).
- The layer limits should be the effective covered surface’s limit instead of the artwork’s size (i think it should be more intuitive: you want to transform what you see in the layer, not the layer itself).
I think most of the other UI complaints are usually of the “Gimp should be like Photoshop” kind, and shouldn’t be taken very seriously. One thing is to improve Gimp usability, but changing an interfase that works just because I don’t want to learn how it works is a completely different one.
Gimp is an excellent piece of software. I have no doubts it will be even better in the future, and this will happen mostly because we can contribute (if we have some programming skills, which is not my case, sadly) and we can share our opinions with the developers (in this blog, mailing-lists, even by e-mail!).
I can’t say the same with Photoshop. If they decide tomorrow that the file menu has to be in the lower left the only thing I can do is cry loud in the streets.
So, let’s be considered with Sven and all the OSS developers around the world. They’re doing a terrific job for us and the whole world, and most of the times they do it for free.
March 25th, 2006 at 1:01 am
sez glazman on his blog:
“Drop the fuc**** gtk+ filepicker and its lame usability !”
What does he have against it? This guy must really hate Gnome too.
March 29th, 2006 at 9:03 pm
Wow, splitting Gimp in two projects : the worst that can be done. Why should i use 2 apps if i can have only one ? I don’t really understand the benefits that we could get from this. Sven is surely right when he writes that Gim has to find its own way and not just follow what users are accustomed to. May be a profile should be better that a split : one profile for Web, one for DTP, one for Photographs, one for designers colorists, each profiles could append to each other if someone wants to have a full program. This way : one code, several use cases. One code, several UI. But this will of course be a hard work to do this.
June 19th, 2006 at 3:39 pm
As a user, I also support the idea of staying away of Photoshop’s UI. My wife has made a full 17mn long animation movie with Gimp and Gimp-gap, “Novecento, pianist”, and has found Gimp really easy to manage with a wacom tablet, thanks to the contextual menu that allows to access everything from a right-click.
I agree however that some windows could be redesigned : I indeed don’t find the filepicker window easy to manage : most people working on professional projects do not put their images in the same directories or use just about ten directories. I almost have every time to click on the button to access the full window, which makes the process of saving or opening requiring more clicks than absolutley necessary. Of course, this is a minor anoyance, but repeated everytime, it becomes boring in the end and that’s the small details that sometimes make the difference.