Hi,
On Thu, Feb 2, 2017 at 1:20 AM, Máirín Duffy <duffy@...43...> wrote:
It's not that they are stupid or don't care. Their goal / focus is simply elsewhere and they don't want to have to make a decision like this - they fear them. I think it is a good practice to have a very obvious "choose this if you dont know" option, because at the very least, it is the least destructive option of any of the 3 we have to offer, it does the least amount of change to the file, and if it ended up being the wrong decision, they can manually adjust the file (I'm assuming!) later on.
I'm not saying that the tactic of having a default choice is absolutely the right one in this scenario - we can take a look at how the screen design reads with the icons, for instance, and see if it makes it clearer which option to choose - but I'm just saying with a lot of experience with this sort of problem (diff context but same problem, think storage and system management UI... ) having a 'default' is the option that users are best able to navigate and requires less cognitive load.
The tension here is the users having to learn something when they're not necessarily in the mood / have the time to learn vs. us taking on the guilt of pushing them towards something that might not be the perfect decision for them and could impact their files negatively. I do think if they choose the default option we push here, that the negative impact isn't going to be significantly high (they print it, it doesn't fit the page correctly, they adjust manually. Costs them a piece of paper, some ink, and time. Ok, so that print could be a plotter print, but still :) )
Unfortunately I think it is safe to assume that a lot of users are going to just leave it at the default safe setting, even when it is not correct for them. That is kind of the (only) idea with a default setting after all, that the lazy user can just ignore the question and click OK? Then when they print they might very well have forgotten all about that question, and even if they remember it not realise that was the problem. And even if they do remember and want to go back, that requires that they still have the backup of the file where they can find it and that they do not mind re-doing all work that has been done on the file between the time they imported it to 0.92 and first time it was printed.
Even if the end-goal is to print something, that can be far into the future. I do not think it is uncommon that you work on a file for several months (and I know, sometimes years) before first printing it. Maybe some pros are clever enough to make sure to test-print things every now and then, and carefully measure the resulting output looking for errors, but I would not count on the majority of users to do that.
So I still think, even if people using Inkscape for printed works are in a minority, they should not be mislead by a default option that can result in a lot of unneeded extra work (and some users abandoning Inkscape). Forcing every user to stop for 2 seconds to think about clicking one of two hopefully obvious choices (most people will know if they want to make art to print or just view on a screen?) does not sound like a high price to avoid that. Of course I am rather biased. :)
The first option (digital artwork) is the do nothing option, which is why it's also being pushed heavily as the default choice.
Would it not make sense to make sure the viewBox is updated so that user-units are correct even when working for screen-output? I know there are some current issues so that user-units are not respected as much as they should, but for the effects that are now working around the broken unittouu things can get ugly if the viewBox is not as expected. If someone wants to keep working with pixels as pixels it should be easy to update the viewBox to match that (and perhaps it is already being done?). I have seen some old documents with pretty weird viewBox-settings, so resetting them to say that pixels (px) really are pixels could only be good even for screen-output only users (I think?). (It looks as if Inkscape currently is cheating a bit and pretend that a px is a pixel instead of a user-unit, so having a broken viewBox might work quite ok, but I consider that a bug rather than a feature.)