In many ways they compare, but they don't compete. To be in competition the economic models would have to radically change. For now any 'competition' is subjective thought.

I was thinking (maybe I'm wrong).... The OpenSource will always be at a disadvantage against to closed source. In OpenSource anyone can analyze the code, read specifications, study the algorithms and technologies. From the OpenSource, we can not analyze almost nothing in the closed source software. They can copy us (not literally), we can not do the same.
So better not to pay attention to closed software, and try to improve GIMP following our own way.

Adobe doesn't care much about laymen, it's the professional community they are after.
So is GIMP, as it follows from the product vision statement: GIMP is a high-end photo manipulation application and so on...
In this sense the comparison of gimp to its commercial counterparts is inevitable.
All things considered, GIMP has a lot of catching up to do to live up to this statement.

From my perspective this goal, this ambition to make gimp a complete and competitive tool for the professional audience is a good thing,
and if you think about it for a moment, it might be the only driving force for a small group of talented coders and enthusiasts, which
the gimp team has actually been. And no young programmers in sight lining up to take over. Take that ambition away and who knows what happens next.

I don't know why we take for granted that GIMP will always be there? After all, people get tired, find other goals, leave the project, get on with their lives.
Happens all the time. Just an example, not long ago Brecht Van Lommel left the Blender Foundation to work for SoftAngle Arnold Renderer.
One of the most talented programmers on the team (Cycles render engine is his contribution and those who are familiar with Blender should know
how Cycles has changed Blender and made it a true competitor among the big hitters in 3d industry).
GIMP needs your support. So, do you, did you?

And yeah, notions like 'I love gimp, it has everything I need, PS is a piece of crap, we spit on it' is a fanboyism.
Not only it is not helping GIMP, it is downright harmful to its reputation and alienates the potential converts.
The same applies to this forum.

@prokoudine The reverse psychology isn't necessarily a bad thing. No offence intended, on the contrary, was meant to be a compliment of sorts
but my apologies if you felt offended anyhow.

Well, I think there are pros & cons to each. Linus's Law states that "given enough eyeballs, all bugs are shallow". Meaning that open source has the advantage of many developers seeing the code. With so many eyes on the code, bugs are more easily detected than with a proprietary, closed source app, in which only a few developers, working in secret, are viewing the code.


I agree. With only a hand full of unpaid developers and no marketing budget, there is no way GIMP can compete with a 7 billion dollar company, like Adobe. Adobe has branded their product so well, it's now part of the lexicon. (i.e,. it's been Photoshop'ed). Most people have never heard of GIMP. I usually get blank stares from people when I mention GIMP. Let's be honest here. Photoshop is an excellent name and provides a good description of the what the product does. GIMP brings up images of a crippled person or someone wearing a leather mask and a ball gag.

I have always viewed GIMP as a hobbyist app and assumed that professionals would fork over the money to purchase Photoshop or some other high-end commercial graphics editor. For what I need to do, which is the occasional photo manipulation and web graphics/logo/art design, GIMP is a perfect tool for the job. It's the "being creative" part of using GIMP that is the attraction for me, as well as the "Free" part. I always assumed this was the case for most other GIMP users, as well.

Even with all the disadvantages GIMP has, I think the developers have done an outstanding job with the app, so far. Yes, there is room for improvement and it will be interesting to see what the future holds. Let's just hope it doesn't follow the same path as Gnome3, which is destined for the bit bucket.


Well, Chrome OS does run on a Linux kernel but that is pretty much where the comparison ends. Google stripped away all the great things I love about about Linux and replaced them with layers of privacy invading, proprietary software, designed to track every keystroke, datamine and control every document and charge me for the privilege. I say; "no thanks" to that.

where have you read this? could not be in this thread ...
i told that the chrome-PS-version has nothing to do with linux (many out there think so). it is, in fact, a streamed M$-version of PS

@Odinbc

It's a generally accepted knowledge that users tend to use your software in ways you probably wouldn't want them to It's also easy to see how developers of GIMP wouldn't want an uphill battle to compete against 200-pound gorilla called Adobe, what with both technical and marketing challenges looming over. But try saying that to users who need to get their job done and are actually choosing between free (in both meaning) GIMP and commercial Photoshop. They really do choose between the two. They evaluate features, usability, file formats support etc. That's competition all right

@YAFU

Frankly, I don't know how you got the impression that we cannot analyze and learn from proprietary software. We can do that. It's just that often it's a somewhat boring work that few are interested to do. Unicorns and rainbows should not be expected.

@K1TesseraEna

I'm really not good at subtle irony, am I? Of course, no offense taken.

@prokoudine

Well, just a feeling after being a Linux user for a long time. For example a lot of suffering and incompatibilities to open closed formats in OpenSource applications. I prefer to think it's because the difficulties imposed by the closed code (or not clear specifications published), instead of thinking that these things happen for OpenSource developers incapacity or lack of talent.

@YAFU

Would you like a short or a long answer about specs and file formats?

It was just an example. Another example would be the failed attempts to achieve an OpenSource alternative to Adobe Shockwave/Flash player.
Does that have the same explanation to the file formats issue?

Short, of course. Anyway I would not understand a long and technical explanation, I guess

I think you're right, YAFU. It's definitely not a lack of capabilities or talent that's an issue for the Free Open Source community. Some of the best devs freely contribute to the effort.

It's just that in many cases, a company needs to keep their software closed and proprietary. If millions are spent to develop an algorithm, say a fast rendering engine, a company isn't going to give that away, and who could blame them? They are going to hold it close to the vest, copyright or patent the technology and hope it doesn't get stolen and reverse engineered by the Chinese.

Take for example, nVidia. There has been an effort underway for years to provide an open source version of nVidia drivers in Linux. Even after all this time, and with a little help from nVIdia, the open source nouveau drivers aren't nearly as good as nVIdia's proprietary Linux drivers. nVidia isn't going to give away their secrets by providing all the technical specifications for their latest video cards and why should they? They are in the business to make money for their stock holders, provide jobs and good salaries to their employees and generate enough revenue to pay for the research it will take to develop the next generation of video cards. Everything can't be FOSS. We need capitalism too.

Millions are rarely spent to develop an algorithm (at least not by private companies). It's more like marginal improvements to algorithms developed in public organizations (of financed by them: DARPA, etc...). Most of the technology advances by making token improvements over something that has been developed elsewhere.

To patent something, you have to publish it (that's one of the original intent of the patent, ensuring that knowledge doesn't get lost...), so there is no need for reverse-engineering, everything is explained in the patent claim (and whatever isn't in there is still patentable, so the explanation goes into a lot detail ). You can also just publish without patenting, if you think patenting isn't worth the effort but you don't want some competitor to patent it.

What you are saying is simply not true, ofnuts. Take for example, encryption ASICS used in smartcard technology and a myriad of other applications. The technology is patented and the inner workings is NOT public information. If it were public, everyone would be getting free satellite TV from DirecTV, DishNetwork and other satellite TV providers. This is because the devices would be copied and sold for millions on the black market.

Also, algorithms do not have to implemented solely in software. Hardware and software are logically equivalent. Millions are routinely spent to develop and protect algorithms. There are cases where companies choosing not to patent but to instead seek protection using a "Trade Secret".

You can have patented stuff and secret stuff. But they cannot be the same thing. To patent something you have to describe it in public.

You do have to describe something for it to be patented, but what I'm saying is that there are many devices which contain algorithms, that were developed at great cost, in which the technology is patented, and in which also contain trade secrets. An algorithm can contain both patents and trade secrets. Smartcards containing hardware ASICS are an example of that.

I am a huge supporter of FOSS and I try to use it whenever possible. The point I'm trying to make is that there are cases, such as I described, where FOSS is not a feasible solution and it takes a lot more resources (than just a few unpaid software devs) to develop and protect an algorithm. Commercial, proprietary algorithms are sometimes necessary and capitalism provides the incentive and resources to make it happen.

@YAFU

Well, I'll try to keep it short.

There is a problem I have with this statement that you made:

This is the typical FOSS mantra that keeps being read on and on. A TL;DR answer to that is: in some/many cases FOSS developers don't even make an effort. Sometimes because they lack time, sometimes — because they are just being lazy or ignorant. In some extreme cases I heard a rather funny claim that it's unfair to support proprietary file formats, if respective proprietary apps don't support open ones.

Here is a more verbose reply, with real-life examples.

1. Visio. For years there had been claims how difficult it was to reverse-engineer its binary file formats and support them in free apps. E.g.:

http://dia-installer.de/doc/faq.html.en

I'm really unsure who wrote that, but much of VSD was reverse-engineered, with actual parser code written in Python, by someone who isn't even a professional programmer. The guy who did it is a Cisco-certified network engineer, for god's sake.

2. PSD. The specification has been open and actively maintained by Adobe for two or three years now. No free software — neither GIMP nor Krita, despite the fact that both had respective GSoC projects — support it in a more or less complete manner. Krita simply had less code to begin with before their GSoC project started last year, and GIMP is missing a) public text layers API to map text layers parser's output to, and b) non-destructive editing implementation to map layer effects parser's output to.

3. ABR. About 4 years ago someone complained on Flickr that GIMP couldn't read brush dynamics from Photoshop. What I was told by the GIMP team is that there is no open spec for that part of ABR files. So I teamed up with the aforementioned guy who reverse-engineered VSD and single-handedly created the spec in about a week in my spare time. While at that, we also reverse-engineered GRD (gradient) files from Photoshop just for fun — we only spent two lazy evenings on that one. Obligatory disclaimer: I haven't written a single original line of code in my life other than back at school (in LOGO and Pascal) some 15 years prior to that moment. To get this done, I only had to be patient and methodical. That was all that was required.

Here's what happened next, or rather, what didn't happen next. Both specs were never used by the GIMP team — they never found the time for that. A couple of years later someone in the Krita team (AFAIK, Lukacz) stumbled upon the ABR spec and used it to create a new brush engine that mimicked the one from Photoshop. Whether they made an ABR loader or not I have no idea. The GRD file format support was added to SwatchBooker by Olivier Berten who had the time to take care of that.

4. Corel DRAW. Same story as with Visio, really. All that was needed to get this supported in end-user free software was a developer willing to team up with someone who can do reverse-engineering. The net outcome: support for CDR files (v1 through v16) in both LibreOffice and upcoming Inkscape 0.91 (probably in upcoming Scribus 1.5 too).

All the examples above have two single common features:

1. One doesn't have to be an uberprogrammer to reverse-engineer a proprietary file format.
2. Having a spec means pretty much nothing if there is no programmer willing to build upon it in end-user software.

Finally, here's an example of a case where even an open spec and a dedicated developer would have little effect on the state of affairs. Please read the comment from Maeric at the bottom of http://libregraphicsworld.org/blog/entr ... esign-tool.

Well, that was my experience. Perhaps yours is completely different.

P.S. Also, just to keep the fun going. OpenRaster, as you probably know, is a completely open file format. MyPaint makes a somewhat special use of it: it dumps another file into the container. This file keeps a record of all brush strokes made, so that at any later time one could reopen the ORA file, click on the canvas where a brush stroke is present, and reload the very brush that was used to make it. If you open such a file in GIMP and re-export it, this file will be erased from the container and lost forever. So much for open specs and open source code.

Implementing a format fully takes time, because in many cases it's not just a format. As you pointed out, using Photoshop brushes may require a specific brush engine. Supporting PSD in full will likely require new layer blend modes (which itself has an impact on the XCF format, if I remember a comment by one of the Michaels), some way to emulate adjustment layers, etc... Even something simple like PNG requires your Internet browser to support alpha-transparency (why is IE coughing )

LibreOffice Draw opens VSD files pretty well... but I'm not sure it has the same algorithm for "connectors", so these may not have the same shape in VSD and LO.

However, as a FOSS developer, you can/should remain a bit paranoid about these proprietary formats. They can be changed for no specific reason, and if you can read them, you are never sure that you are writing them properly...

The situation with ORA, MyPaint and Gimp seems to be equivalent to the debate on actual support of OpenDocument by MS-Office.

@Ofnuts

To be sure.

Not in case of GIMP. Apart from a few features like dual brush, Photoshop's brush engine is a subset of GIMP's brush engine and would map to existing features nicely. OTOH, we'd need to backport Mixbrush from gimp-painter to get an even more complete feature map, but this is about tools.

As for the rest, GIMP# illustrates that a lot is doable in terms of Ps features support once you want doing it. Besides, those extra blending modes are useful (darktable has those).

So you are saying that we should never again blame the proprietary software for faults in the OpenSource software? That's new to me, it will cost me much assimilate it.
What about LibreOffice? Do they have full-time developers, right? I still read a lot of complaints about poor compatibility with MS Office formats.

@GnuTux, capitalism is not a problem for me. But the State always controlling and preventing abuses. Yep, Social Democracy is my thing

@YAFU

I can't, for the love of god and all that is holy, imagine how you arrived at this conclusion.

As for LibreOffice, I'm not overly familiar with how that project is organized.