OpenCL Is Coming To The GIMP Via GEGL

[ninez]

That's a rather valid point. Please note, however, that GIMP is targeting the hi-end market segment either (not there yet, but getting there).

And i will be excited when that happens

The animation related difference is solved via GIMP GAP project which has flipbook, bluescreen etc. For years already it's just the high bit depth that made all the difference. And painting tools in GIMP are far more sophisticated these days, with even more changes coming in 2.8.

GAP, in my personal opinion doesn't really have as good of a workflow as flipbook in CinePaint. (i have used it many times). It's not that i don't think it's a good tool, i just find using CinePaint to be much quicker and straight forward. If I have a folder of photos - i just load them up, edit them all, save and i am done. simple, fast, no fuss. That being said, i am also fully aware that GAP is actually more powerful.

I agree, gimp's tools are getting better all the time. I've been running Gimp from GIT, it's coming along

And that's another problem with that project. Latest release source code was indeed 0.22. Source code for 0.23 and 0.24 wasn't even released, though binary builds exist! As far as I can tell 0.25 is either code from CVS or Kai-Uwe's unofficial Git repository (which is kinda the only reason one could argue whether Cinepaint is dead or alive). BTW, Ubuntu 11.04 doesn't have Cinepaint.

Interesting, the last version of Ubuntu I checked out, still had it ~ but i don't use Ubuntu - not my cup of tea... Im an Archer (Archlinux). The code i am referring to is from CVS ~ there is no official release, but it is what is typically packaged. This is becuase 0.25 is the most stable version of CinePaint - and while some distro's may be dropping it, Arch for instance updated very recently. Suse has also done a nice job of packaging it (they and Redhat probably have some customers, that are still using it professionally (in enterprise) i would imagine - so there are people, who want to keep it around, for now.

And that's a rather valid point as well except the aforementioned animation bit

What I'm saying, essentially, is that today it makes little to no sense arguing (and especially trolling) about history. We know that the GEGL plan hasn't worked the way it was supposed to work, and the FilmGIMP/Cinepaint plan hasn't worked either. What matters is presence and future. There is the first real GEGL-based tool (Cage transform) coming in 2.8, and at least two more GEGL based tools are coming in 2.10 (interactive version of iWarp and seamless paste, plus unified transform tool is in works since a couple of weeks). There's no way back, only forward.

I agree with you. - but for me, anyway. CinePaint has been useful for a long time, and IS 'currently' still valid and very useful - do i see it being relevant 5yrs from now? of course not. I also don't understand why others are in disagreement/trolling/giving you a hard time over the history. From what i know, and from the 'factual infomation' you have posted - I would tend to agree with you. You seem quite knowledgeable on the subject, and i've enjoyed your posts.

and trust me, I'll be absolutely stoked when Gimp is up to the task - i would love nothing more, than to never have to invest money into Photoshop ever again... and will be even more happy, when i am not required to run a graphic suite in a VM, for when i need 'real powertools', not found in Linux (yet)

[Geri]

they should create an usable user interface with usable features, instead of opencl support

[prokoudine]

GAP, in my personal opinion doesn't really have as good of a workflow as flipbook in CinePaint. (i have used it many times). It's not that i don't think it's a good tool, i just find using CinePaint to be much quicker and straight forward. If I have a folder of photos - i just load them up, edit them all, save and i am done. simple, fast, no fuss. That being said, i am also fully aware that GAP is actually more powerful.

Interesting. I'm not what you might call an avid flipbook user (read: I'm talking out of my bum), so I think I might try compiling Cinepaint again and doing a side by side flipbook comparison.

they should create an usable user interface with usable features, instead of opencl support

This kind of mentality is quite dangerous. When you have a team that is too small, start doing just one thing, and lots of other users immediately begin to yell. There are plenty of vocal people who think UI changes are not as important as high bit depth support, and there are extremely vocal people who think that all GIMP needs is proper CMYK support. So what you get really is actual work on UI and laying foundation for the future like this OpenCL project.

[Geri]

This kind of mentality is quite dangerous. When you have a team that is too small, start doing just one thing, and lots of other users immediately begin to yell. There are plenty of vocal people who think UI changes are not as important as high bit depth support, and there are extremely vocal people who think that all GIMP needs is proper CMYK support. So what you get really is actual work on UI and laying foundation for the future like this OpenCL project.

Well, lets see. If i have a team that is too small, i use pure logic. I simply wrote down all the serious problems wich is affect my software, i point on the most heavy problem, and fix that. Currently, gimps biggest problem is that the gui is impenetrable, serious graphics arts are laughing of the unmanageable apperance.

And istead of trying to get rid on this, developers of gimp adding opencl support - wich is needed around 0,01% of the users...

Imho this is the attitude wich is dangerous.

[prokoudine]

Currently, gimps biggest problem is that the gui is impenetrable

Really? Desktop publishing folks think it's lack of CMYK support. Photographers want 16/32bit, LAB and and adjustment layers. How do you know it's you who is right and everybody else is wrong?

And istead of trying to get rid on this, developers of gimp adding opencl support

Try reading the article again, this time — with attention. Only one GSoC student is working on OpenCL, and his mentor isn't even a GIMP developer. The other folks work on UI, better tools and better painting features, just like you wanted them to.

wich is needed around 0,01% of the users...

Which is needed by pretty much everyone who has a decent GPU and needs fast processing of images.

[Geri]

fast processing of the images needs normal multicpu support, not mytchical librarys originated from cthulhu.

i will read the article again tomorow, now i go sleep

btw i know really lot artists from deviantart, and a lots from 3d graphics industry (lot means over 100)... but i will not try to paste they oppinions here from gimp, mostly from the apperance, probably the forumengine itself would even block some words out

at least, opensource community should not ignore the facts.

[liam]

Interesting. I'm not what you might call an avid flipbook user (read: I'm talking out of my bum), so I think I might try compiling Cinepaint again and doing a side by side flipbook comparison.


This kind of mentality is quite dangerous. When you have a team that is too small, start doing just one thing, and lots of other users immediately begin to yell. There are plenty of vocal people who think UI changes are not as important as high bit depth support, and there are extremely vocal people who think that all GIMP needs is proper CMYK support. So what you get really is actual work on UI and laying foundation for the future like this OpenCL project.

Are you a GIMP developer?

Best/Liam

[smitty3268]

they should create an usable user interface with usable features, instead of opencl support

See, I'd actually argue that the kind of developer working on OpenCL support is exactly the kind of developer you'd rather NOT have working on the UI.

That could be horribly unfair to this developer (sorry), but often people with those kinds of interests aren't very interested in the UI, and correspondingly aren't very good at it. Things could always get worse!

It's a mistake to assume that OSS projects can simply hire people and assign them to the areas you would like work on - they have to attract people who are interested in doing it for free.

[prokoudine]

Are you a GIMP developer?

Oh, I'm more in the news writing and community tracking department. Few actual developers have accounts at Phoronix, but I have no idea how often they come to visit. The OpenCL student was in this thread some days ago for sure

[prokoudine]

That could be horribly unfair to this developer (sorry), but often people with those kinds of interests aren't very interested in the UI, and correspondingly aren't very good at it.

Sometimes when I'm in the office I pick up the phone because nobody else's around to answer calls, and 50/50 the person who gives the call assumes I know every single fact about the company and the products It looks like our friend here is one of such people

IMO, it doesn't make a lot of sense to demand people knowing everything. Engineering is engineering, but IMO it's always best to have people who know really much about something than a little about everything.