Consider a user who decides to write their document with a default font size that is smaller, or larger, than the 12pt the default template offers; let's say it's 11pt. Our user creates the new document, edits the Default Paragraph Style and changes the font size. This should be enough for the user's document to be nicely rendered relative to a basic font size of 11pt: * The footnotes and endnotes should be smaller than usual * Captions should be smaller than usual * Headings should be slightly smaller than usual etc. etc. in actuality - most of these things happen. Or rather, this happens for those styles which are _exactly_ the same size as Default Paragraph Font, and none of the others. So it's not even consistent over everything. Our user needs to manually and carefully set all the styles' font sizes to an appropriate value relative to their new default. That is annoying and time-wasting - and is not what a user would expect. (I mean, ok, we're used to this being the case so we already expect to have to do everything manually, but that's a bad thing.)
Font size can be styled as % of the 'Inherited from' style instead to set up a fixed size. But that is a designer election. Maybe could be developed a script to change those style with a fixed size different from default and substitute with the percent that it is from default.
(In reply to m_a_riosv from comment #1) > But that is a designer election. Well, we are the "designers" of the default document template (and a bunch of other template). Why did we choose to fix the sizes instead of making them relative?
Percentage has the drawback of non-integer values and headings use absolute values now, see bug 142423, reviewed in bug 151258. Plus, these values are hard to understand and users struggle to change it back (we could improve this with an explicit control, see bug 72662). There are also the duplicate bug 121011 and bug 121008. My take: NAB/DUP The actual solution is to define and use a template with the "advanced" settings.
I think some character styles also need to be made relatively-sized to the default, e.g. footnote reference (correct me if I'm wrong). Bug 127702 is about making the possible.
(In reply to Heiko Tietze from comment #3) > Percentage has the drawback of non-integer values We could support percentage-sizes only manifesting as values rounded to points, or half-points. Let me check if that bug exists, otherwise I'll file it. > and headings use absolute values now, see bug 142423, > reviewed in bug 151258. Would be nicely resolved by percentage-with-rounding. > Plus, these values are > hard to understand and users struggle to change it back (we could improve > this with an explicit control, see bug 72662). > There are also the duplicate bug 121011 and bug 121008. You mean, duplicates of each other? Anyway, those are about how the user sets font sizes in the dialogs; this is about the user's experience with the document overall and without the user making any changes themselves. > My take: NAB/DUP You have brought up technical difficulties which do not contradict the fact that it's a bug. Users expect to be able to scale all sizes with a single change; and specifically, expect font sizes in styles to have a logical relationship with the default paragraph font size. They do not expect nor want to manually change all the font sizes, and we shouldn't make them. You could argue this is a WONTFIX because the fix is problematic (which I would disagree with); but certainly not NAB. Also, this is not a dupe of any of the bugs you mentioned.
(In reply to Eyal Rozenberg from comment #5) > We could support percentage-sizes **only** manifesting as values rounded to > points, or half-points. Let me check if that bug exists, otherwise I'll file > it. (emphasis mine) Making it **only** support pt/half-pt would mean limiting what's possible for our native file format, to please a foreign format. It may be reasonable to make that an option. Of course, that whole bug 142423 resolution by replacing percentage with fixed values was silly, and made already poorly visible feature totally hidden (even by existing example). Well, sigh...
(In reply to Heiko Tietze from comment #3) > Percentage has the drawback of non-integer values and headings use absolute > values now, see bug 142423 which is a loss, exactly for the reason that Eyal wrote this bug. > reviewed in bug 151258. To me that does not look like a review. (In reply to Eyal Rozenberg from comment #2) > Well, we are the "designers" of the default document template (and a bunch > of other template). Why did we choose to fix the sizes instead of making > them relative? Supportive for that view: - the default template is behaving much better; - the inheritance that is shown in the Styles pane, is respected; - someone changing a style, noticing that the font size is given in a percentage, can easily come to the idea to use a fixed pt size.