On Wed, 2012-05-02 at 17:45 -0400, Jonathan Kamens wrote: > On 05/02/2012 05:36 PM, Adam Williamson wrote: > > On Sun, 2012-04-29 at 09:48 -0400, Jonathan Kamens wrote: > > > I agree. The save vs. export thing in the new version of gimp is mind-numbingly stupid. > > Makes perfect sense to me. It's the same logic used in Audacity. You are > > always losing information when 'saving' to an image format - the record > > of your edits to the file. You're exporting for production, you're not > > saving your work. > It doesn't make sense for most people, because most people don't give > a flying fig about "lost information," they just want to edit an image > and be done with it. Optimizing an interface for the minority of > people who use it is not wise, especially when the people who do care > about it are knowledgeable enough to know how to make the program do > what they want even without changing the interface. GIMP is a professional image editing suite, just like Photoshop. It's not a basic touch-up tool. It doesn't really make sense to criticise it on the basis that you want to use it for something it's not really designed to do. > In any case, changing the behavior without making it configurable by > the user, was just totally gratuitous and stupid. Don't get me started, but I'll give you the short version: it's very easy to say "well just give me an option to make it do whatever it is I wanted!", but it's very often not a very good idea for the developers to actually do so. This is the cause of featuritis, which is in turn the cause of impossible-to-use interfaces and almost all bugs (which are, at root, usually the result of code complexity). When your app has a hundred interface configuration options and consequently, say, a few billion codepaths, it's almost impossible to anticipate all the consequences of _any_ code change. Keeping the number of codepaths to the minimum possible is vitally important to producing stable and reliable code; implementing any old UI option anyone asks for is a great way to wind up with an excessive amount of codepaths. > The smart way to accomplish what they were trying to accomplish would > have been to add a File | Import... command, and a corresponding > "--import" command-line option, and a corresponding "Import into GIMP" > context menu item, which takes an image in whatever format is being > imported but then defaults to saving in XCF format. Did you check and see if they considered it and saw some drawbacks to it, perhaps? Did you try simply suggesting this, to the GIMP authors rather than the Fedora test list (I'm still not sure why this thread is here), without the disapproving editorial? As someone else replied to this thread, it's vital to present criticism with an attitude that makes the person you're giving it to want to act on it, rather than get defensive. Saying what they are doing is 'ridiculous', implying that it's dumb (by saying another way would be 'the smart way'), making unsubstantiated assertions about 'most people' (have you done a scientifically supportable poll?), and calling their software 'mind-numbingly stupid' are all excellent ways to make developers more inclined to be defensive and less inclined to believe that you have something constructive and useful to suggest. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net -- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/test
