GNOME Bugzilla – Bug 548167
Font dialog proposal for Gtk+
Last modified: 2010-07-10 03:43:35 UTC
The current font dialog remains the same for some time, and as a frequently used dialog, should get more usability focus. Attached there's more information about a proposal, together with a functional prototype on how the dialog can be improved for a better layout, faster font selecting and providing a better preview. Full story with details and screenshots: http://hcalves.com/blog/2008/08/17/font-dialog-proposal-gtk/ Functional prototype in Python/Glade: http://hcalves.com/media/files/fontselector.tar.gz It's being discussed on gtk-devel-list too. Thanks! Other information: The other day, while reading planet.gnome.org, I saw a post from Alberto Ruiz about a proposal for a new Gtk+ font dialog. I really liked the idea, but tough Alberto's proposal had some major drawbacks, then I gave it a shot focusing on usability aspects. Original Planet Gnome discussion: http://aruiz.typepad.com/siliconisland/2008/08/font-selection.html
A second try, based on feedback gathered on gtk-devel-list: http://hcalves.com/media/files/fontselector2.tar.gz Full history, new details and screenshots: http://hcalves.com/blog/2008/08/17/font-dialog-proposal-gtk/
I think the font dialog needs more radical changes than what you have done. Look at http://eyegene.ophthy.med.umich.edu/unicode/fontdialog/ for a font-nerds view on these issues...
It was discussed on gtk-devel-list, and some ideas raised: * Being able to filter fonts by aspect (Fixed-Width, Variable-Width) - already done on 2nd prototype; * Filter fonts by language (e.g., filter fonts able to render different glyphs) - Pango is already capable of this; * Using distinct, i18n preview texts, depending on the context (e.g., not using Latin text "Lorem ipsum" if the font selected is for arabic script, so we avoid fallback of preview text to sans, or unicode's hexcode blocks); The thread: http://mail.gnome.org/archives/gtk-devel-list/2008-August/msg00144.html PS: That link (http://eyegene.ophthy.med.umich.edu/unicode/fontdialog/) was mentioned on the thread too.
(In reply to comment #2) > I think the font dialog needs more radical changes than what you have done. > Look at > http://eyegene.ophthy.med.umich.edu/unicode/fontdialog/ > for a font-nerds view on these issues... I very much like these aspects of the font nerd's proposed solution: - It aims to render every font in its native script. - It describes with much detail useful example strings, including eastern scripts that for instance don't have pangrams but different ways to pick representative characters, particularly Chinese kanji. If pango can do that, that'd be nice, I don't really know if pango's examples cover this topic. - It suggests symbols that emphasize on features of non-latin scripts such as the ability to render particular Arabic characters that not all fonts support. - It elaborates on the concept of grouping fonts, similar to how one can have important folders as bookmarks in the file chooser. It would be nice if these points could be considered in the proposed changes.
(In reply to comment #4) > (In reply to comment #2) > > I think the font dialog needs more radical changes than what you have done. > > Look at > > http://eyegene.ophthy.med.umich.edu/unicode/fontdialog/ > > for a font-nerds view on these issues... > > I very much like these aspects of the font nerd's proposed solution: > > - It aims to render every font in its native script. > > - It describes with much detail useful example strings, including eastern > scripts that for instance don't have pangrams but different ways to pick > representative characters, particularly Chinese kanji. If pango can do that, > that'd be nice, I don't really know if pango's examples cover this topic. > > - It suggests symbols that emphasize on features of non-latin scripts such as > the ability to render particular Arabic characters that not all fonts support. > > - It elaborates on the concept of grouping fonts, similar to how one can have > important folders as bookmarks in the file chooser. > > It would be nice if these points could be considered in the proposed changes. See what I committed to Pango last night: http://svn.gnome.org/viewvc/pango/trunk/pango/pango-language-sample-table.h?view=markup That's what pango_language_sample_string() uses now.
*** This bug has been marked as a duplicate of bug 407885 ***