This is excellent! Not only does single-window mode look much cleaner and more organized (to me), but we'll finally be getting support for high color depth images. That's one of the biggest reasons many photographers have stuck with Photoshop. These are probably two of the largest, most significant changes to Gimp to come about in a long time. I'm surprised the version number wasn't bumped to 3.0.
edit: Is the GEGL backend port only partially complete? If so, how long until it's in place?
Other tweaks are nice and cosmetic: tooltip progress indicators for transforms, on-image text editing instead of awkwardly doing it in a separate window, etc.
Layer groups are another important organizational touch.
I'm very much looking forward to this. Great work, Gimp team!
Straight from the article, note that it explicitly says you won't have the high color depth image support:
What's really important, but maybe less user-noticeable, is behind the scene: library work. GIMP 2.8 is an intermediary step in the progress for GEGL integration (GEGL is a graph based image processing framework that will provide the core for GIMP, allowing for much advanced features). This process started with GIMP 2.6, advanced with 2.8, will be finished in 2.10. Expect the real meaty features (like high depth color channels) in the release after 2.10 (is not know if it will be labeles 2.12 or 3.0). For now GEGL is used in some UI elements and some optional filters and tools
"At the time this article is written, about 90% of the GIMP application’s core are ported to GEGL, and the only thing really missing are GeglOperations for all layer modes... GIMP 2.10’s core will be 100% ported to GEGL, and all of the legacy pixel fiddling API for plug-ins is going to be deprecated. Once the core is completely ported, it will be a minor effort to simply “switch on” high bit depths and whatever color models we’d like to see... The port lives in the goat-invasion branch in GIT. That branch will become master once GIMP 2.8 is relased, so the first GIMP 2.9 developer release will already contain the port in progress."
The article is a little bit out of date. The 16bit colors will already be available in the next release (2.10). Michael Natterer and Øyvind Kolås had three-week hackathon and now 90% of the work is done.
According to all the docs and pages I see online anywhere, GIMP won’t support 16-bit/channel color until version 3.0, and your source doesn’t really contradict that AFAICT. (None of them give a good estimate what time frame that is.)
But in any event, GIMP’s lack of support for real, powerful color management features, high bit depths, CMYK – much less CIELAB or other perceptually relevant color spaces (ideally they should use CIECAM02 or IPT) – etc. make it pretty much useless for any of my own photographic workflows. GEGL doesn’t look likely to overcome the main deficiencies, and so GIMP remains a cute toy more than a power tool.
Here’s Graeme Gill from approximately 2 years ago:
Graeme Gill:
> Rupert Weber wrote:
>> (Then again, nobody complained so far and probably
>> nobody would ever notice if we did it Right(tm)...)
>
> You've got a catch 22 situation: If Gimp doesn't
> handle color conversions accurately, then noone
> interested in accurate color transforms will use
> it, hence you will get no complaints about not
> doing in Right. So if you want to persuade people
> who value accurate and reliable color
> transformations to use Gimp, you need to do it
> Right, otherwise they will continue to use something
> else. (I certainly can't use Gimp in my work, which
> revolves around checking color space
> transformations.)
>
> Graeme Gill.
I think the source comes close to contradicting that without being explicit:
'GIMP 2.10’s core will be 100% ported to GEGL, and all of the legacy pixel fiddling API for plug-ins is going to be deprecated. Once the core is completely ported, it will be a minor effort to simply “switch on” high bit depths and whatever color models we’d like to see. Oh, and already now, instead of removing indexed mode (as originally planned), we accidentally promoted indexed images to first class citizens that can be painted on, and even color corrected, just like any other image. The code doing so doesn’t even notice because GEGL and Babl transparently handle the pixel conversion magic.'
you read that blog post wrong: 2.10 will have GEGL (2.10 won't be released until GEGL integration is ready), but that doesn't automatically make 16bit color channels, it only enables it. it will happen in a later release.
In 1997, me and my beige box PC spent about a day building the GIMP alpha, excited for a Photoshop killer. Now it's 2012, fifteen years later, and we're high-fiving GIMP for the single window interface.
Many people do not mind the multi-window interface and the reason there wasn't a single-window one isn't some stagnation in progress but probably either opposition by the past maintainers to the idea of switching to a single-window one or not caring about it at all. Those fifteen years are hardly a measure of how Gimp fares in comparison to other image editors.
Can someone please explain to me the advantage of NOT having a single window display in an application such as this? I'm being serious here by the way ...
I remember when I tried GIMP many years ago. I had many little windows scattered across my desktop, and as soon as I misclicked on my background I'd suddenly lose focus of the app and all the sub-windows (toolbar, layers, etc) would disappear.
I either had to run my mouse back to the image's window or alt-tab back to the app to regain focus.
I found this incredibly terrible as a UI approach, but found no resources explaining why this is a good thing.
Partly - the Unix philosophy. Since you can swap out your window manager on X11, you can have your windows managed in any of a wide number of ways. If an application tries to manage subwindows on its own, there is a very good chance that it will do so differently than your global window manager. Assuming that there is a reason you picked the global WM that you did, it's rather disconcerting to be forced to use an inferior WM to manage Gimp's windows.
Add to that the multi-desktop, multi-monitor aspects (although detachable docking panes can help a lot with that).
Desktop environments on Unix systems on which GIMP was primarily supposed to run have always had virtual desktops, and with virtual desktops you can easily distribute the little tool windows to one desktop and the image windows to other desktops (or physical monitors if you fancy), which would not be possible in the single-window mode.
GIMP's multi-window interface is excellent when you have multiple monitors. Couple that with a good window manager and it's a lot better (IMO) than a single window with tools and other dialogs inside.
I have my window manager manage the windows. I can use all my usual keys and shortcuts to modify/move/alter the windows. Eg mousescrolling over a window's title bar to reduce it to just its titlebar if I do not need it (there is a shortcut too, I forgot it though). Being able to make windows always in front. Being able to have multiple images open and move/resize them with alt+(left|right) mousedrag.
- you can move away the docks and have more screen estate for image editing;
- press TAB and the docks go away, you have even more screen estate;
- is easier when you work with more than one image at a time, for example using the clone tool you take samples from one image and move them to another or just when use one image as reference while drawing another.
For an app meant for the creative types, GIMP sure is ugly.
I also thought we were over mystery meat navigation, but it seems to be going strong at GIMP. Not only does having icons for each tab make it look incredibly busy, but it also makes it harder to learn and wastes a ton of screen space. I can't see a single benefit other than it doesn't look like Photoshop.
I wonder how hard it would be to redo the entire interface from scratch. I would love to design it, but no way do I want to navigate their bureaucracy.
I doubt theres much bureacracy, there are not that many core developers - probably going onto the mailinglist and providing some patches might be a start.
I don't understand the appeal of the multi-windows. I use GIMP at home and I often have multiple images open and find the tool windows disappear on me. It's frustrating. What's overrated about the single window interface and what draws you to multi window mode?
I guess the question is whether or not you're using a powerful window manager or not. Most window managers in linux support easy always-on-top and layers, so the multi-window UI really shines.
Ah, this makes sense now :) I haven't tried with a window manager. I'm typically in Windows at home and not always on Linux so I'll have to experiment.
I don't even get how people can claim that there is a big difference anyhow - If we were brutally honest here, Photoshop is really only multi-windowed, too - with the main window maximized in the background and carrying the main menu.
Single window mode is definitely the step in the right direction towards what I think GIMP should be doing: ripping off Photoshop wholesale. Perhaps GIMP has some features that are arguably better than Photoshop, but the problem is that a huge number of people are very used to Photoshop's interface. Just about everyone has messed about with Photoshop at some point either professionally or just for fun (be it legally or illegally) and Photoshop's interface (for better or worse) has become second nature. I would actually argue that the lack of a familiar Photoshop alternative is holding back mainstream adoption of Linux. Open Office or Libre Office are good enough replacements for most people (and they have obviously gone for the wholesale rip off tactic) but GIMP just does not suffice - the main thing being the interface, not lack of features.
There are really simple things that could be done like making the toolbar the same as photoshop's. I mean it's great that you can resize however you like, but having it unresizable makes it easy to remember the button layout and instinctively grab for a tool. The fact that completely different symbols are used for the same actions does not help. eg: knife is crop in GIMP and slice in photoshop.
Once Linux/GIMP has achieved good market share over mac/photoshop then it's time to start thinking about improving the photo/vector publishing paradigm and boy is there much remove for improvement. Adobe has pretty much had the monopoly on 2D graphic design tools for the last couple of decades. With almost no competitors - what incentive is there to truly innovate? I mean compare the sophistication of something like 3D studio max or Maya compared to illustrator. (3D being an area where stiff competition still remains) Why is Photoshop and Illustrator even two completely separate pieces of software? because they can sell two things separately for a combined total price greater than one?
I'm a developer and I desperately want to learn how to design. I gave up on the GIMP and bought Photoshop simply because there are so many more resources (PSD's, tutorials, blogs, videos, etc). Occasionally I pull up the GIMP, but quite honestly there aren't enough hours in the day.
I get the feeling that the GIMP devs wants to be "better" than Photoshop, but the steering wheel, and gas and brake pedals are where they are in a car; it's all about muscle memory. Making it easier for Photoshop users to fire up the GIMP for quick projects will do more to promote the GIMP than being 2x better than Photoshop.
a verbatim copy of the Photoshop UI is a lawsuit waiting to happen. and Photoshop is not the paradigm of UI design, is just something a lot of people know how to use.
(I don't like these quick put downs, they imply a finality that doesn't exist)
That The Gimp has been unsuccessful with a different interface to Photoshop does not mean a different interface could not be successful.
Imagine that an amazing new interface was developed that was so much better than what any product has now. Everyone asks why this has not been thought of before. Now imagine that it belongs to version 3.0 of The Gimp.
Just because one alternative interface didn't work, doesn't mean we should all give up and use the same interface. With that advice, nobody would ever try anything new.
that's not what the grandparent poster said, let me quote "what I think GIMP should be doing: ripping off Photoshop wholesale." - that would be a BAD thing
I'm depressed that so many of the comments here focus on the single window mode, instead of being thrilled about GEGL/Cairo. Those are the killer features that will underpin the GIMP's next few years of awesome features.
It's pretty sexist if you ask me. A general complaint of mine: tech news (and in this case this blog post) include meaningless pictures of woman in a way that is embarrassing to the industry. I'm getting tired of clicking links on HackerNews and LinkedIn today that look like US/People Magazine to any casual observer walking by. We're better than that.
The majority of the article is doing touch up work to a woman, which is a very common task. There is nothing sexist about this work, and it's idiotic to claim so.
Why have the main picture a shot of cleavage? He doesn't touch up the cleavage anywhere in the picture. If he did, it'd probably make for a pretty weird demo. The point of the picture was to illustrate the single window mode. It seems like there would be a common sense test where you think: "Is this really how I want to represent myself and this software?"
It's a female in a bathing suit! There's nothing sleazy about touching up a female or male in photo editing software. In fact, these are probably two of the most common things touched up for photo work.
Out of curiosity why is it embarrassing? I bet that if I'll ask some female friends from the industry what they think of the article, none of them will complain about this.
It's like a sensationalist link-bait title, but more visible. I really dislike that opening industry news has a high probability of looking like I'm reading a tabloid to anyone who walks by my cube. Seriously, if someone walked by and sees a 30 in monitor with a large picture of some girls cleavage do you really think they'll buy that I was reading about open source photo editing software? It reminds me of the old joke: "I read playboy for the articles."
Maybe I'm unreasonable, I'm open to that conclusion.
I just find myself think "really guys, how does that picture say anything about this article" incredibly often. It seems tacky.
Professionalism. The blog lacks all pretense of any.
- The blog's tagline talks about "bullshit and ... how much I suck".
- It leads off with a close crop of cleavage. This negatively attracts males' attention while being untactful to others.
- He then demonstrates Gimp's new features by making people uglier. Have you watched any of Adobe's demo videos for comparison?
I don't know who this Nicu guy is and what his relationship with the Gimp project, but a look at his LinkedIn profile reveals that he was a developer 8 years before changing careers to photography for the last 3.
However his blog post is #11 under the Google search for "gimp 2.8", so the level of professionalism can leave a bad impression for new users.
Speaking of professionalism, I'll quote from the history of the agile manifesto:
In order to succeed in the new economy, to move aggressively into the era of e-business, e-commerce, and the web, companies have to rid themselves of their Dilbert manifestations of make-work and arcane policies. This freedom from the inanities of corporate life attracts proponents of Agile Methodologies, and scares the begeebers (you can’t use the word ‘shit’ in a professional paper) out of traditionalists.
the author (me) is a beauty/glamour/fashion photographer, so he used as an example one of the pictures he usually use GIMP for editing. and such image surgery tools like the cage transform are used mainly for this kind of editing.
Agreed. The article would be better if the examples were less distracting. It's like talking about layout and styling and using a political rant as your example text.
We don't really need anything new in this space, though.
We need a free/open Photoshop clone. We need a direct, one-to-one monkey copy of Adobe's product.
We don't need flights of fancy, we don't need to retrain existing users of image manipulation programs, we don't need to experiment. Adobe already has all the legwork done for us--the answer looks a hell of a lot like Photoshop and its interface.
There is zero--ZERO--reason to try to come up with a new interface when there is a perfectly good one in existence that just happens to be poorly distributed.
The place where GIMP has screwed up (and Blender and the like along with them) is in not parroting exactly the common tools in the ecosystem.
There are weird programs (Wings 3D, Paint.NET, etc.) that are able to significantly depart from conventions, mostly because they have very narrowly-defined functionality (something I don't believe anyone would ever accuse the GIMP of).
The GIMP developers should either focus on making a monkey copy (and ignoring everything else until feature/UX parity is achieved), or stop working on the project entirely so they can do something else for a while. Once the monkey copy is done, then innovate. Until then, stop wasting time with things nobody cares about.
I used the GIMP as my primary art asset creation tool for several years. It was miserable. The filters and procedural image generation stuff was awesome, but everything else was at best hard to use. Eventually I switched to Photoshop because it was a (comparatively) clean, tight tool with all the bullshit removed.
> I used the GIMP as my primary art asset creation tool for several years. It was miserable.
I used the Gimp for < spam > http://rommelhok.com/en < / spam > (48 pages worth of old fashioned adventure comics). Not for the drawing itself, but for the cleaning up, the coloring and the lettering. It was fine.
Everything worked without a hitch, once I got used to some new shortcuts (which can easily be set to the Photoshop equivalents if you want) and some minor differences from Photoshop, some of which were, in fact, improvements (the color selection tool is much nicer in Gimp than the color range selection box in Photoshop, for example).
The only area where Gimp definitely falls short is type setting. In Photoshop, you can tweak kerning, spacing and font size within a text layer for every character individually. In Gimp, you can only do that for the layer as a whole. This wasn't much of a problem for me, however, since computer lettering is already inferior to hand lettering (IMHO) and using Photoshop's superior tools wouldn't have decreased the suckiness of the lettering all that significantly.
The other main shortcoming of Gimp, of course, is that it can't do CMYK colors but that will only become a problem when someone decides to publish my comic, in which case having to redo the colors will be a nice problem to have. I made one printed copy of it for myself at the local copy shop using the RGB Gimp files and that turned out beautifully.
For drawing Inkscape is a better tool, yes, but I did the drawing, both pencils and inks, on paper and then scanned the result for further processing which was mostly removing smudges and correcting mistakes: pixel based operations.
Sorry, but no. Photoshop's interface is not particularly easy to use (discoverability issues), nor is it efficient (too much modifier key usage). Photoshop users are overtrained on a suboptimal way of working. What we need is something with photoshop's engine (which is very powerful), but a better UI. Gimp is not that though, and it never will be (good UI must be designed like a cathedral, and gimp will always remain a bazaar of features).
I agree that GIMP will always be a bizarre of features.
The thing is that, while Photoshop may not be the pinnacle of UX, it's familiar to everyone in the space. It has issues in the UI, but nothing that can't be overcome fairly quickly and nothing that is a gamebreaker.
For design wanks, yeah, you could probably improve it. You then lose the overtraining, though. You also might screw up. You end up trying to develop a moving target.
For a productivity tool, I don't want imagination. I want excellent execution of a proven design. Only in magical fairy pixie land does it pay to reinvent the wheel to be slightly rounder on a tool everyone already knows well.
As I mentioned elsewhere: don't let the best be the enemy of the good-enough.
> I agree that GIMP will always be a bizarre of features
> For design wanks...
> Only in magical fairy pixie land...
Sorry, but need to ask - are you trying to troll here?
I've used The GIMP for about a decade now and it fits perfectly into the UNIX methodology: A sharp tool that does its frigging job. There is no reinvention of the wheel and people who say that without pointing to anything but the single-window/multi-windows difference should really find better use for their time.
Actually, in your original post, you didn't even mention that. Your verdict was: "The filters and procedural image generation stuff was awesome, but everything else was at best hard to use. Eventually I switched to Photoshop because it was a (comparatively) clean, tight tool with all the bullshit removed.".
Any software has a learning curve, The GIMPs curve not being much different from Photoshop, just in a slightly different direction. I don't find any of the tools hard to use and that was from within a month of using it - not just because I've gotten used to it over the years. Actually, if anybody argues GIMP vs. Photoshop with me, they mostly try to convince me that GIMP lacks features, so I really have no clue whatsoever what you mean by "all the bullshit" that is supposedly in there.
"Sorry, but need to ask - are you trying to troll here?"
No more than GIMP has trolled me. : )
First, look at the toolbar. Instead of folding the tools into categories as PS does, they throw it all out there. This is clunky and dumb. There is no reason to have 30+ icons on the toolbar. Photoshop has many more features, and yet doesn't manage to clutter up the interface with them ( see http://www.designlessbetter.com/blogless/wp-content/uploads/... ).
Second, if I close the toolbox pallette, GIMP closes. What the hell is this?
Third, right-clicking on the canvas gives me the same tools as the menubar. This tells me that one of the two is superfluous. Oh, and some of the options are duplicated yet again on the toolbox pallette. So, make that three. Huh. I can also click the upper-lefthand corner of the canvas to get the same menu. Four. Four ways.
Fourth, at least the tool options are layed out--oh, no, no they aren't. There isn't a fast-access tool menu along the top of the screen, and the dialog is coded to not reflow reasonably when positioned there.
Fifth, the goofy mascot is leering at me over the top of the toolbox pallette, as if to dare me and try to make sense of the tool.
Sixth, the scrollwheel doesn't seem to be used, so why not have it zoom in and out? Nope, not done. Instead, it scrolls up and down. Useless. The devs let us drag the canvas with middle-click, but no scroll zoom.
Seventh, color tools are duplicated (yet again) between the Colors menu and the Tools->Color tools menu. Derp.
Eigth, there is both a Layer->Transform menu and a Image->Transform menu. Acessible from something like three different places. So, six redundant menus? I don't even know what the fuck.
Ninth, why are Edit and Select different menus? Even PS doesn't do this.
Tenth, why do Colors->Map and Filters->Map have different mapping options? They're the same? I think they used to be together?
I could go on, but it's giving me brain cancer.
~
I'm not just trolling. The GIMP is an accidentally useful sprawl of bad design. The fact that you have managed for a decade to use it for anything is much more a credit to your skill/persistence/stubborness than any redeeming qualities it has.
First - So "no horizontal lines" is now a complaint?
Second - This is a limitation of the Operating System, where closing the root window closes the application.
Third - No, it just gives you shortcuts to the most important menu. What else would you want there?
Fourth - So have it on the side! It's not like horizontal screen real estate is the rare commodity these days. Also - really? Saying "this horizontal menu is not vertical" is criticism now?
Fifth - I guess you also complained about the "eye" in Photoshop before they replaced it with the "Ps" sign?
Sixth - Scrolling scrolls. It just... scrolls. Not sure what else to say here. Also, Ctrl+Scrolling Zooms, Shift+Scrolling scrolls horizontally. Seems pretty frigging straight-forward to me.
Seventh - Oh boo-frigging-hoo, there is two ways to the same end. There are also keyboard shortcuts, you know? Three ways!
Eigth - Are you trying to be adversary here? Transforming the entire image or just the layer are obviously separate, justified entities. And again - offering them somewhere else doesn't make them less useful, just easier to access. Not sure where else than in the menu (or the right-click copy of the menu) you've found them, though.
Ninth - I'm starting to be lost for words here... How about: Because they are two separate things? Are you even aware of their functionality?
Tenth - These are unfortunately named the same, but I would say this is in the nature of the beast - A color map is a justified concept and a filter that has 'mapping' behavior is a justified concept. So your criticism is more one of the English language.
I was really prepared to debate some of the more reasonable complaints (like color management, somewhat clunky plugins etc.) you could have with GIMP, but sorry - this is just noise.
"Horizontal lines"...? What? No, I'm saying that the toolbox palette is cluttered and takes up more space and is less usable than the PS equivalent.
"This is a limitation of the Operating System," is not actually a limitation of the operating system. All you do is create an invisible root window that handles events for your application, and then all your visible windows are children of it. Or any of several other solutions.
Most of the rest of your points go back to defending duplicated functionality. This is bad design. There is no reason to apologize for that--it's just a failure. It should be redone.
As for the mapping--at least for "alien map", those used to be in the same plugins menu. This is a regression.
Not really sure we would both enjoy further nitpicking on this. I guess the only thing I can conclude here is that even based on your claims, I would not arrive at the same conclusion that you arrived at.
I could probably write or find a list of equal length and severity on the shortcomings in Photoshop. I simply fail to see even the sum of your claims support the strong statement that you made in your original post.
Good shout with Wings 3D! It's definitely weird and awesome. For starters, it's a desktop app written in Erlang. I recommend people bookmark it if they think they will ever need an intuitive way to do basic-intermediate 3D modelling.
I've used Wings3D to do static meshes for over five years. It's very, very, very good at the type of modeling that it does--and miserable for anything else.
But for UV pelt mapping, for box modeling, for those sorts of tasks? It shines like a diamond.
It's got probably the most minimal interface I've ever seen on a modeling program and, despite a few oddities, probably the most discoverable/fastest.
Yeah, I back you up on this 100%. I never use GIMP on my Linux box because it's not enough like Photoshop. I know, ridiculous sentence, I sound like a blasphemer, my apologies. There's nothing wrong with saying Photoshop has got things right. Sure it's nowhere near open, and incredibly overpriced but that's exactly why I want to love GIMP. I've been wishing for a Photoshop clone minus the bad parts of Photoshop plus some innovations but what I've gotten so far is just falling short. It feels to me like MS Paint on steroids. UI is decent but the functionality kills me in that any time I try to do something like I would in Photoshop like, say adding a simple stroke to a shape or adding a gradient or shadow, in GIMP it doesn't come out looking like its even in the same ballpark as Photoshop no matter how long I spend tweaking knobs and sliders.
GIMP may not be for us though. I was under the impression it was a Photoshop alternative but maybe I took that too literally. If I'm the minority then so be it, I'll keep searching for a better fit but I do hope GIMP grants some of my wishes. I open it and I feel like I could use it but then u start using it and it disappoints me.
For those who haven't used GIMP, the big fix of single window mode is probably one of the most important steps they're taking. I have occasionally used the fact that I can stack the windows atop each other for hiding purposes, but much more often I'm saying, "dammit, where the hell is my layer bar? Where the hell is my toolbar now? augh! switching to my reference picture in the browser has upset my window order" and so on.
On-canvas text editing and cage transforms might be useful; we'll see.
Huh? Since I switched to using tiling window managers, I actually prefer multi-window interfaces for the first time ever because, to me, when tiled they essentially work like single window interfaces, except I have a lot more control over it.
(I use a manual-tiling WM, not an automatic-tiling one - I wonder if that makes a difference in this case)
awesome is also dynamic tiling, while I use a WM that does manual tiling (though I can make it dynamically tile by window class, application and other parameters - though I have to write scripts to do it - its not automatic out of the box) and maybe this is the difference: I can lay the windows out precisely how I want, while if I understand dynamic tiling WMs correctly, they will try tile the windows for you, which may not work well for multi-window interfaces.
Layer groups! This makes loading those PSDs the graphic designers send around much more bearable. At least I assume that the PSD import can map the PSD Layers to Gimp Layers.
This needs to come out soon so I can start the wait for 2.10 :)
Seriously though, this is a great step in the right direction, I can't wait till post 2.10 when the geglfication is complete, then some really interesting features should be possible.
I'm undecided about save/export in applications that support multiple formats. I think users should certainly be warned if they're making a lossy save, but perhaps they should all be in the same dialog.
The drawback of making it a single dialog box in the user experience is that the "current name" changes when you use "Save as...". So if you do a "Save" next time, you save it as the exported name and format instead of the original one. If, for example, your export format is .png, and your original format is layered, this could bite you. Also you keep flipping between two names.
Sometimes an "autoexport on save" feature would be nice. Export it to the derived format as well as save the original file with one keypress.
In Gimp, however, if you save to a format that does not support layers when your image has layers, you get a warning. Likewise if you save to a lossy format like JPEG you get a warning. Every time.
Since always. Major releases just about always lead to an influx of users. This may be less true on Linux — I don't know — but it is generally true from everything I have ever seen.
edit: Is the GEGL backend port only partially complete? If so, how long until it's in place?
Other tweaks are nice and cosmetic: tooltip progress indicators for transforms, on-image text editing instead of awkwardly doing it in a separate window, etc.
Layer groups are another important organizational touch.
I'm very much looking forward to this. Great work, Gimp team!