Improve message in pre-0.92 file dpi upgrade
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Inkscape |
Fix Released
|
Low
|
Mc | ||
0.92.x |
Fix Released
|
Low
|
Mc |
Bug Description
The current dialog has several UX drawbacks : non-technical people might find it too technical, it gives almost no hint on which option to prefer, and "Ignore" might be an ignored option, while it would in fact be the better option for anyone not wanting to send his/her drawing to print.
Ideally, we should rework on it to perform some autodetections and context (maybe based on http://
================
Relevant IRC parts from tonight:
22:35 < Mc> I might want to land a string change
22:38 < Mc> to add something like "<b>What to do now?</b> If your document is intended to have a size in pixels, choose Ignore. If you wanted it to keep some size in real-life units, choose to set viewbox."
22:53 < Mc> I'm not for removing the technical explanation
22:54 < Mc> just add some rule of thumb for people not knowing what it implies for them
22:53 < Mc> I'm not for removing the technical explanation
22:54 < Mc> just add some rule of thumb for people not knowing what it implies for them
23:20 < deevad> Mc: su_v ; maybe http://
23:21 < Mc> well, in your case, none of the options work :p
23:21 < Mc> you want what's currently "ignore"
23:21 < deevad> "auto-convert" might sound a bit dangerous ; "auto-adapt"
23:22 < deevad> ignore sounds dangerous too ; as if I'm missing to do something important
23:27 < deevad> http://
01:48 < Mc> maybe http://
01:49 < Mc> (for now "Advanced" contains Tav's text)
01:51 < Mc> (also, I think it could be a good idea to maybe add a link (in "Advanced") to a page on our website where we would have leisure to add more information about this change)
01:53 < Mc> (current diff : https:/
02:02 < Mc> http://
02:07 < bryce> does this sound at all accurate?
02:07 < bryce> Scaling behavior has changed in this version of Inkscape.
02:07 < bryce> How would you like to update this old file to current?
02:07 < bryce> 1. Resize the printed page to fit drawing
02:07 < bryce> 2. Resize the drawing to fit the printed page
02:07 < bryce> 3. This is for screen display only, so disable scaling
02:14 < su_v> bryce: your first item is not what actually happens - the page is not changed. the viewBox scales the content to match the intended physical size
02:14 < su_v> AFAIU
02:15 < su_v> only if the file was entirely defined in px (e.g. based on default template from 0.48), the width, height are scaled (and the viewBox set to the old values)
02:17 < su_v> if the width, height are e.g. in 'mm' (A4 template from 0.48), then it only needs a viewBox
02:17 < bryce> if you have a more info... link then there is no need to make mention of CSS compliance; that can be moved to the full explanation page
02:22 < Mc> bryce: would it be a faq item ?
02:22 < su_v> Mc: basically, the first and the second option both scale the drawing, just using different methods (viewBox, or scale each drawing element) - the intent is that a page defined in physical units keeps the exact size (and
the content is scaled to match its relation to the page size)
02:23 < Mc> it's scaled so that 96 user units = 1 external inch
02:23 < Mc> right ?
02:27 < bryce> 1. Shrink or enlarge drawing elements in relation to the page, keeping physical dimensions constant; or 2. Keep drawing elements constant with respect to the page, changing dimensions as needed; or 3. Do not apply
02:27 < su_v> but no prompt.
02:27 < bryce> more accurate?
02:28 -!- bastianilso [~<email address hidden>] has quit [Quit: bastianilso]
02:28 < su_v> 1 and 2 are the same
02:28 -!- Ede_123 [~Miranda@
02:29 < su_v> (the methods of the two options differ internally - viewBox for scaling, or scale each object)
02:29 < su_v> AFAIU
02:31 < su_v> each method can have drawbacks (can expose bugs in inkscape; might not be supported well by what might post-process SVG file)
02:33 < su_v> scaling via viewBox is more stable I think (except that some extensions might produce incorrect results in such files); scaling elements of complex drawings (clones, masks, clips, etc.) doesn't always keep appearance (bugs)
02:34 < Mc> (hence recommendation to use "fit page")
02:34 < su_v> scaling elements is often preferred by anyone who does not want transforms or other attributes like viewBox, but simply wants the values in element attributes to represent the actual position or length
02:35 < su_v> Mc: yes, I agree that it is "recommended" (current "set 'viewBox'"), if scaling is needed.
02:36 < su_v> I do understand though why (some) users would prefer to scale each element instead.
02:38 < su_v> I just have the impression that the difference that now is made (on that proposal) does not exist: both options aim to preserve the physical size of the page area, and the original relation of the content to that page area.
02:42 < su_v> if you have an A4 page defined in px (based on 90 px per inch), and draw a rect which is 10.5 cm wide; the result of either option should be that the page area still is 21cm wide (defined in px based on 96 px per inch),
and the rect measures 10.5cm in width (half the page)
=================
My current diff:
Changed in inkscape: | |
milestone: | none → 0.92.1 |
status: | New → In Progress |
importance: | Undecided → Low |
tags: | added: ui |
Changed in inkscape: | |
milestone: | 0.92.1 → 0.93 |
Changed in inkscape: | |
status: | Fix Committed → Fix Released |
Patch with hyperlinks activated.