The State of GIMP & Its Future

[starshipeleven]

Pretty much all points in this thread have been covered by the projects's FAQ. So my takeaway from this thread is that the majority of present Phoronix readers don't read about stuff they discuss.

And you haven't seen anything yet. The fun starts when most posters can't even understand the stuff they discuss. But keep doing so nontheless.

[geearf]

Unfortunately, they're not receptive to it. It took something on the order of a decade of people proposing Photoshop-like UIs (and even hacking together things like GIMPshop) before they gave us the single-window mode we have now.

That is unfortunate.

[prokoudine]

Unfortunately, they're not receptive to it. It took something on the order of a decade of people proposing Photoshop-like UIs (and even hacking together things like GIMPshop) before they gave us the single-window mode we have now.

I think you are confusing "receptive" with both "critical" and "not having time to write code".

GIMP is a huge application with little manpower behind it. Some UI change proposals are OK, and we do them, if we have the time.

Some of the proposals we receive make sense and could be done, if there were more developers working on the app.

Other proposals sort of make sense and sometimes even come with patches, but need work before those patches can be applied. This is e.g. where we are stuck with https://bugzilla.gnome.org/show_bug.cgi?id=759601 where the general idea makes a perfect sense, but unless the original developer of the patch makes adjustments, the feature won't go in.

Then there are proposals by people who think GIMP should work like some other different app. The result, if implemented, would make GIMP actually less consistent. As someone who personally made Ps shortcuts for GIMP, as well as quite a few shortcuts schemes from other apps for Inkscape (Corel DRAW, FreeHand etc.) I can tell you that different apps have matching UX approaches only up to a point. So if you change just one part to work like application X and leave the rest untouched, it only makes things worse. And if you change everything to work like in application X, you will a) disappoint the existing user base, b) will have to adapt your application to what application X does, forever. Good luck with that.

Finally, single-window mode was added in 2009, but 2.8 having it was released in 2012. So check your numbers.

[ssokolow]

I think you are confusing "receptive" with both "critical" and "not having time to write code".

Fair enough. I know what that feels like all too well.

Then there are proposals by people who think GIMP should work like some other different app. The result, if implemented, would make GIMP actually less consistent. As someone who personally made Ps shortcuts for GIMP, as well as quite a few shortcuts schemes from other apps for Inkscape (Corel DRAW, FreeHand etc.) I can tell you that different apps have matching UX approaches only up to a point. So if you change just one part to work like application X and leave the rest untouched, it only makes things worse. And if you change everything to work like in application X, you will a) disappoint the existing user base, b) will have to adapt your application to what application X does, forever. Good luck with that.

You'd have my sympathy, but I'm still sore over your decision to force "XCF is the only valid 'save' format". My mother actually has problems with her muscle memory preferring Ctrl+E over Ctrl+S in other apps because she uses GIMP so much and my TODO list actually contains a "Write a GIMP wrapper to auto-close unsaved warnings" because it's such a big glaring exception to the effort I've put into my desktop to have it anticipate my needs. (In my case, because I use GIMP for situations where I'll pull up a PNG, make an edit or two, save the changes, and never touch it again.)

As a UI/UX guy, let me remind you: A UI design cannot force a workflow on an unwilling user!

Finally, single-window mode was added in 2009, but 2.8 having it was released in 2012. So check your numbers.

Ahh, yes. Sorry about that. I'm really starting to wonder what I was doing over the last four years. I keep running into things that feel like they only happened maybe a year ago, then I stumble across the website again and discover they happened in 2011 or 2012.

(I'd also forgotten how long GIMP release cycles are... which I really shouldn't have, given how much I grumbled about the contributor-discouraging evils of long, heavy release cycles back when I read about them in the wake of the 2.8 release.)

[Passso]

There is only 1 modern, stable, and state-of-the art desktop environment available for linux: GNOME. So there is no reason to drop GTK despite the fact Qt's API and tools are a masterpiece of software development.

You crossed the sarcasm's line.

[prokoudine]

As a UI/UX guy, let me remind you: A UI design cannot force a workflow on an unwilling user!

Every single software out there does enforce a workflow on a user. The question is how much flexibility you get along the way. We focus on users who make complex work with lots of layers etc. So we encourage other developers to make software that appeals to a different user base and provides different workflows. And that's one of the reasons we like the Krita project.

(I'd also forgotten how long GIMP release cycles are...

Yes, we do have a problem with long cycles. Developing features in branches kinda helps, but there's little one can do with unevenly balanced workload in the team while having to introduce massive internal changes and general polishing. Besides, you can't really force a volunteer to do just the boring work on the internals. It's free software, it should be fun. So we end up with new features that also delay releases

[FreedomHash]

From my understanding the don't want Gimp as a drop in replacement for PS anyway.Great for armature's but i cant ever see it becoming the industry standard for at least another 10 years to which they will be another 20 behind PS.

Most of what you said didn't make any sense (to be fair due to typos and/or grammar issues - it seems like you were in a hurry and/or on a mobile device).

But to this point: I don't know who "they" is/are, but Gimp has alot of potential as a photo editor / Photoshop replacement. True, when I really need to get work done I do it in Photoshop. Mainly because I haven't yet mastered the Gimp UI. But the vast majority of Photoshop's latest features, go unused in every day photo editing/retouching/processing. Everything that existed 20 years ago, will do almost everything you need, with the exception of limits. (64-bit was an absolutely crucial development, as was the large PDB format, as is the ever-expanding 16-bit capability.) Most pro Photoshop users I know, use custom actions or scripts based on very basic transform primitives anyway (e.g. contrast mask). These can be done just as well in Gimp.

Also, while I'm not a fan of the Gimp interface either, and do know Photoshop's like the back of my hand, let's not forget that Photoshop is no masterpiece of UI design either. It too has a very steep and fairly inconsistent learning curve. In fact there are downright awful choices and inconsistencies. (E.g. the dialog interface for newer filters and operations...completely different than the older ones.) It's just that everyone knows it well.

With Adobe moving to their bullsh*t subscription model, I am highly motivated to make Gimp work. Scripting may get me there. (Note that I'm not inherently opposed to the financial aspect of a subscription model. It's the way it is implemented, and the DRM enforced that to me are too bitter to swallow.) My biggest obstacle to moving photo workflow to Linux is not Gimp, but lack of a really solid Adobe Bridge facsimile. (And DNG converter. I use DNG only for it's single-file metadata handling.)

[ssokolow]

Every single software out there does enforce a workflow on a user. The question is how much flexibility you get along the way. We focus on users who make complex work with lots of layers etc. So we encourage other developers to make software that appeals to a different user base and provides different workflows. And that's one of the reasons we like the Krita project.

True. I should have said a UI design change cannot force a workflow on an unwilling user with established opinions on how they want to work! (Which may be specifically why they chose your software in the first place)

My issue with Krita is that I'm not trying to do digital painting... I'm trying to do some quick photo retouching, using a program I have used for years, and it just so happens that my workflow is centered around one-time editing that takes a PNG or JPEG file and returns a file of the same format.

Moving non-XCF saving to a separate menu entry is a curious choice... especially when there was already precedent in tools like OpenOffice.org for warning if the last save occurred to a less expressive format.

Yes, we do have a problem with long cycles. Developing features in branches kinda helps, but there's little one can do with unevenly balanced workload in the team while having to introduce massive internal changes and general polishing. Besides, you can't really force a volunteer to do just the boring work on the internals. It's free software, it should be fun. So we end up with new features that also delay releases

True. I just wish GIMP could somehow aim to follow Firefox's post-4.x lead in that area.

[liam]

Funny you should say that, given that over the life of the project, Gnome has had far more investment in UX than pretty much any other open-source project. The people who designed Shell *are* UX folks - I can't say I like all their decisions, but those decisions aren't just the result of a bunch of developers arbitrarily copying Apple.

That was back when Sun (NOT GNOME --- the GNOME foundation spends its money elsewhere --- right now, Google may, strangely enough, be the biggest investor in gnome ux via their soc and jim hall) spent the money to put together a ux team (headed by calum benson: https://www.linkedin.com/in/calumbenson, but including Seth Nickell: https://www.linkedin.com/in/seth-nickell and Bryan Clark: http://advogato.org/person/clarkbw/). The HIG wasn't an afterthought. The people involved weren't amateurs playing at ux (here's the HIG released for gnome 2.2: https://developer.gnome.org/hig-book/2.32/index.html.en and here's their old GUP page: http://advogato.org/proj/GNOME%20Usability%20Project/).
Here's G3's "User Experience" leads: Jon McCann (https://www.linkedin.com/in/williamjonmccann), Allen Day (despite numerous searches I've been able to find zilch about his background), Jakub Steiner (https://www.linkedin.com/in/jakubsteiner).
Hopefully you'll notice some differences in backgrounds (i.e.: none of the gnome3 ux people have any formal hci training).
Besides that, it would be worth your while to look at the differences in their approaches (yes, it'll require a lot of searching, but it's worth it if you want to why some people are particularly annoyed with the G3 DESIGN cabal).

Here are a few more links that speak of these things (at varying degrees of separation):

http://www.jonobacon.org/2012/07/30/staring-into-the-abyss-some-thoughts/

https://people.gnome.org/~halfline/g...esign-june.txt ---- that is one of the few gnome-design logs that I've been able to find (they, esp mccann, don't like having records of their meetings)

[liam]

Funny you should say that, given that over the life of the project, Gnome has had far more investment in UX than pretty much any other open-source project. The people who designed Shell *are* UX folks - I can't say I like all their decisions, but those decisions aren't just the result of a bunch of developers arbitrarily copying Apple.

That was back when Sun (NOT GNOME --- the GNOME foundation spends its money elsewhere --- right now, Google may, strangely enough, be the biggest investor in gnome ux via their soc and jim hall) spent the money to put together a ux team (headed by calum benson: https://www.linkedin.com/in/calumbenson, but including Seth Nickell: https://www.linkedin.com/in/seth-nickell and Bryan Clark: http://advogato.org/person/clarkbw/). The HIG wasn't an afterthought. The people involved weren't amateurs playing at ux (here's the HIG released for gnome 2.2: https://developer.gnome.org/hig-book/2.32/index.html.en and here's their old GUP page: http://advogato.org/proj/GNOME%20Usability%20Project/).
Here's G3's "User Experience" leads: Jon McCann (https://www.linkedin.com/in/williamjonmccann), Allen Day (despite numerous searches I've been able to find zilch about his background), Jakub Steiner (https://www.linkedin.com/in/jakubsteiner).
Hopefully you'll notice some differences in backgrounds (i.e.: none of the gnome3 ux people have any formal hci training).
Besides that, it would be worth your while to look at the differences in their approaches (yes, it'll require a lot of searching, but it's worth it if you want to why some people are particularly annoyed with the G3 DESIGN cabal).

Here are a few more links that speak of these things (at varying degrees of separation):

http://www.jonobacon.org/2012/07/30/staring-into-the-abyss-some-thoughts/

https://people.gnome.org/~halfline/g...esign-june.txt ---- that is one of the few gnome-design logs that I've been able to find (they, esp mccann, don't like having records of their meetings)