Slashdot Mirror
GIMP Not Enough for Linux Users?
nursegirl writes "Novell has been running a survey about apps that people need in order to convert their data centers or desktops to Linux. The online survey has been running since Jan 13, and Adobe Photoshop was at the top of the list as of February 1. Desktoplinux.com has an interesting article about why the existence of the GIMP isn't enough for many professionals."
(dons flame resistant suit of anonymity)
Maybe this is because GIMP has one of the most god-awful GUIs known to man. I mean seriously, it seems to be designed to hide functions and impede work, not t'other way round.
>99% of business desktops don't have Photoshop, let alone whatever a "datacenter" involves. If Photoshop is at the top of Novell's list, all it shows is that if you have an open web survey and ask Teh Community for responses, you get replies from 15-year-olds.
What I'm listening to now on Pandora...
I don't understand why people find this so impossible to understand -- the MacOS APIs (Carbon and Cocoa) do not exist on other platforms. You can can compile vanilla Unix applications on MacOS X, but you can't trivially recompile (or wrap) a Cocoa app on Linux.
I do. I'm a professional, not a hobbyist, and want to do my damn work, not fuck around with the interface.
If Adobe figured out some way to lock down Photoshop so that it couldn't be pirated as commonly as it is currently. I know tons of people who use Photoshop and praise it to the heavens, but not a single one of them actually put the money down on it. I work in a university environment, so there're lots of legal copies of Photoshop around, but a lot of people work with their own hardware, so many copies that get used for preparing images for publication aren't legitimate.
I use the GIMP for the same tasks, and get results that are just as good, though. I think that for most image processing, the GIMP does everything the average user needs it to do, and more. I'm not denying that it doesn't meet the needs of certain professionals. However, if people weren't able to get pirated copies of Photoshop readily, they'd find that the GIMP does the job they need it to do.
Today our lesson will be Chapter 1 of Elementary Necromancy: Proper Use of a Shovel.
The very fact that this question has to be asked says a lot about why Linux (and other OSS) has trouble making it in fields with established software. I presume that the people who wrote GIMP wrote it to meet their own needs, because they certainly haven't taken the time and effort to meet the needs of print graphics professionals. Even if you ignore the interface and a number of other shortcomings, the lack of CMYK support makes it IMPOSSIBLE for it to be used in a graphic arts environment for printed products.
The primary colors of light (and therefore monitors) are red, green and blue (RGB). The primary colors of printing are cyan, magenta, yellow and black (CMYK). A digital image starts out as an RGB and is edited that way, but it must be converted to CMYK before it can be sent to an imagesetter for four-color printing. This isn't a "good thing to have." This is a showstopper not to have. It's like having a car without wheels.
I keep hearing OSS people breezily dismiss criticisms of software such as GIMP or just insist that it IS good enough for professionals. The very fact that some people are arogant enough to try to shove tools onto people that WILL NOT DO THE JOB shows why it's hard to adopt Linux on the desktop. Linux has done well in areas where geeks have written software for other people like themselves. It has not done well in areas where the geeks don't "get" what professionals in other areas must have. A commercial company has a serious incentive to make software that fits the needs of those other people. The people who write OSS tend to just want to write things that are fun and useful to them -- and that severly limits adoption of Linux in non-technical areas. Of course, it also doesn't help that so many Linux people seem to take the attitude that the Linux desktop is fine, but artists and other non-technical types are just too stupid to use it.
David
This really is the key. GIMP will never have more than a marginal user base because they don't understand their users. Their users--nearly all of them--are Photoshop users (or potentially ex-Photoshop users).
Good user interface design means not just creating an inteface that "makes sense," it's also creating an interface that works the way the user expects it to work. If over 90% of your users are used to the way Photoshop does function X, then you sure as hell better implement function X the way Photoshop does. Not because that way is better or makes more sense, but because that's what the user expects you to do, and any deviation from those expections means your app is "broken" in their eyes.
Competing on features in this sort of market is futile. Your program may be able to give me the moon on a stick; but if I can't easily make it work, it might as well do nothing at all. The success stories--those projects that have managed to supplant a deeply-entrenched competitive offering--have always acknowledged this fact and have modified the behavior of their own product to compensate. The failures in this arena (GIMP being the most famous) always refuse to acknowledge the effect on their users' expectations caused by their competitor's dominance. For projects like the GIMP, it seems a matter of pride to not be influenced by such an unworthy competitor.
"With sufficient thrust, pigs fly just fine. However, this is not necessarily a good idea...."
RFC 1925
So perhaps someone should start a project similar to WINE, to add Carbon and Cocoa API compatibility.
09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
*blink*
A web designer shouldn't have to look at HTML source much?
All's true that is mistrusted
I must say, well said. I don't agree with much of it, but you make your point well. I think the failings are in Linux and OSS and not in society. People often better themselves, the problem is that time is a limited resource, and which topic they choose to better themselves with is frequently an exclusive option. Given a 4 hour time block, a typical artist might have a choice... they can dive into one of their projects, add shadows, retouch some photographs... or they can spend it learning a new application. Most people will choose to better themselves by refining their ability to do what they already do well. Maybe using the GIMP would be a marvelous idea that enables them to surpass their wildest creative dreams. But there's really no way to know that before doing it. A person is just as likely to spend hours a day for a few weeks learning a new program only to discover it doesn't offer some core functionality they already had in an existing program.
People aren't stupid. The elitests who believe the average user, and average person, is a gibbering idiot is usually just as dumb when they are confronted with tasks outside their element. A Linux guru might wonder why everyone else is just too dumb to use all the wonderful CLI tools and scripting capabilities, yet when confronted with an automechanical problem, the mechanic is chuckling to himself about how Mr. Linux Guru is too dumb to even perform basic maintenance on his own car.
Like I said, time is a limited resource. Everyone can't spend all their time being an expert at everything.
Well, for Cocoa anyway, there's GNUstep, as, if I'm not mistaken, it's an implementation of the OpenStep specification that was created for NeXT and is still used today for MacOS X as Cocoa. Once GNUstep is reasonably completed, it would in theory be possible to have a certain amount of source-compatibility between any platform with GNUstep and Cocoa. Carbon, now that's a different story...
Qu'on me donne six lignes écrites de la main du plus honnête homme, j'y trouverai de quoi le faire pendre.
It's obviously not made for print due to the lack of CMYK-support
It seems that Adobe and their patents play a role in that, but its true of course that this is a serious limitation for those whoms work is going to be used in print.
and for web production.. well, compare Photoshops "Save for web"-module vs GIMP's "Select a JPEG compression percentage please"-prompt.
If you are doing graphics work professionally, is it too much to ask that you have some idea about how different compression levels work out? This is pretty equivalent to knowing how different kinds of paper work out when you profession is printing.
I am not a graphics artist, but I do run some websites that are used by graphics artists for publication. I had to tell each of them to stop using the bloody 'save for web' module for their pictures because the result of it is crap. Rather, they should be using jpegs in 1280x1024 resolution or better, compressed at 90% quality or better. The website will do recompression when needed. Of course the recompression by the website is why you should feed it high quality sources, but the 'save for web' confuses the hell out of those graphics artists exactly because it explicitly hides what it is doing from the user.
No, it comes down to the fact that the vast majority of graphic designers and artists don't work in a vacuum. Artwork gets sent to customers for approval. It gets sent to publishers and print shops for production. Those people have to be able to read those files with no hassle. They have to maintain color accuracy. They have to work.
If you're billing clients top dollar, and have print runs on the line that can easily cost hundreds of thousands of dollars, then risking that account just to keep from spending $600 on a professional-grade, industry-standard tool is... well... stupid.
Any sect, cult, or religion will legislate its creed into law if it acquires the political power to do so.
They just don't realize it and don't really have (as you said) the time to put into it. Which is still a failing of society.
Oh come on!, a lot of people do not *have* the time just because they DO NOT CARE!. They prefer playing with their Playstation, getting drunk or fixing their car than to get into the computer.
You fail to see that, at the same way you (and I) enjoy hacking the computer, normally people enjoy hacking their cars, stereo system or any other hobby they have. And it does not mean that the society is failing.
We all have our priorities, and although for you, the computer could be a very important tool, there is people who only use it as a comunication tool. Think as the telephone, you do not care how your telephone work... you may not care how is it programmed, you just want to pick up the phone, press the buttons and speak.
We all have our priorities, and the fact that the priorities of other people are not the same as yours does not mean their are doing any wrong.
Although I arrived late to the article, let me state something. This last week, I have been working in some simulations. I made a simulation on the computer wich gave me as results something like 400,00 MB in numbers.
Now, I needed to do statistical analysis on those things, unfortunately, the deadline of the paper is for this wednesday, and I have never used any of those Statistical analysis tools. I didnt need anything too fancy, only std. deviation and averages.
Guess what I used, Excel, it has an OK statistical analysis package. Now, I wont "rant" about the absence of that on OpenOffice, I did everything I needed in MS Office, but to do that I had to import my text files (delimited by a space) to Excel. I did some simple C programs to process my code and then just imported with the File/Open function of excel, it detected it was text file and a wizzard guided me through the import stages.
Now, what does all of this have to do with the "linux still not ready"?, well, after finishing, I thought "how could I do it with OpenOffice" because you know, everybody says OpenOffice is as good as Ms Office (something I do not believe). Well, I tried to open one of those files with the File/Open IN OpenCalc and it just opened a OpenWrite window with the numbers HA!
I looked for an "Import" button, I tried with the "Document Import wizzard" without luck. So I could not even *start* to compare it.
Now there are a number of several details that I *doubt* OpenCalc has, that Excel does besides importing a file or being able to make cross references between worksheets and books but, you must see that the devil of the commercial vs open software is (as in everything else) in the DETAILS. Those small details that people take from granted when using Photoshop, Excel, Word, etc. And the fact that in some of those products you can go from 0 to a complete work in a few minutes (God, this is the first time I do a *real* statistics analysis).
Ubuntu is an African word meaning 'I can't configure Debian'
If you are a freelance web designer, you "must" know the subtile differences between CSS rendering on Firefox and on Internet Explorer. You shouldn't be afraid to open notepad and write the HTML code directly. You must know all the little tricks. If you don't then you are doomed.
But...When you work inside a web agency, then roles are defined. The web designer concentrates on...design. He/she makes the lay-out according to the corporate identity, the marketing stuffs, ergonomy, and so on. His/her role is purely on design. There is another guy, a technician guy who knows everything about techniques. He/she will transform his/her work into a working HTML based lay-out.
He will give all the guarantees that it will work on all major browsers.
Then a web developer will put the lay-out inside the CMS, or as the user interface on a custom built web application.
This is a team.
On large scale project, you've got enough work justify a full-time job on design and another one to make the result HTML compliant.
My company, a small web agency outsources everything related to design. We use traditionnal infographists. We had to "educate" them on basic stuffs, but it in the end, it helps us to concentrate ourselves on the web site features and technical parts.
Most technicians are extremely bad at communication/graphism and so on. Most of us can't understand why we should spend hours to make a stupid paragraph aligned with some tiny parts of the lay-out, nor can we understand that the customer may get mad because font is Arial 10 instead of Arial 12 on the subtitle. We simply can't understand why it matters so much and why the customers cannot understand the beauty of our new CMS with all the new features that let us make multilingual content with simple clicks or this new XML import feature that works automatically with one of their partners.
A lot of talended designer are bad on the technician part. They simply don't care about how it works.
People aren't stupid. The elitests who believe the average user, and average person, is a gibbering idiot is usually just as dumb when they are confronted with tasks outside their element. A Linux guru might wonder why everyone else is just too dumb to use all the wonderful CLI tools and scripting capabilities, yet when confronted with an automechanical problem, the mechanic is chuckling to himself about how Mr. Linux Guru is too dumb to even perform basic maintenance on his own car.
Mmmh, no. You underestimate the stupidity of the average user.
We are talking about people who can't install a program in Windows, who can't guess that if you want to open a file you might want to check "file" menu.
I've seen people using Word to copy files (open & save as) and centering lines using spaces completely ignoring align icons.
What you forget is that User Interfaces are designed to make interaction easy while car engines are not.
Using your analogy an average user wouldn't know how to change gears or which pedal is the brake.
Well, while you are technically correct, you shoot past the who point by miles.
The idea isn't to try to actually view at that color depth. Its already beyond the capabilities of many video output devices, and even possibly the human eye. But again, thats not the point nor in dispute.
The issue is the accumulated filter effects and tranformations applied to a digital image. Each such effect can create subtle artifacts and degradations. When you start with 8bit/color channel (traditional 24bpp) then these can build up fast to become noticably visible in the final image.
But if you apply those effects to a 16bit/color channel (48bpp) image, the artifacts don't become noticable as quickly, if at all, assuming you are using a good quality image manipulation program. Then when all is done, you can convert your final image to 8bit/channel (24bpp) such as jpeg and have a clean image.