Thanks guys.
My keyboard is UK, it does have a key between Z and shift; thanks for the advice to ignore it. Yes, I see that ≠ and ≥ are already available on the Alt/Option key which is good. I just need to choose another modifier key for the set of dead keys (without using multiple modifier keys which is a bit cumbersome).
Can I use Control 'ctrl'? or Function 'fn'?
Did I understand correctly that (unlike on a PC) right Alt is not separate from left Alt?
Could I achieve my aim without using a modifier key, eg using the key to the left of 1 has a first level dead key, then a letter as a second level dead key to select the set of symbols then a key for the actual symbol I'm after?
I realise that Ukelele is running in real time, and that if such functionality exists it will be by Ukelele specifying these key combinations to functionality in apple's keylayouts. Some symbols are made up of multiple elements, for instance one symbol is two horizontal lines (longer than =), and another is three horizontal lines. Does Ukelele have an ability to turn multiple normal keystrokes into a single character, ie if I type two underscores, can Ukelele translate this into the symbol that has two horizontal lines, and translate three underscores into the triple horizontal line symbol?
On 06 Dec 2015, at 13:29, Connor <elg...@tackleanything.co.uk> wrote:Geke, what advantage does LibreOffice have over OpenOffice?I have another problem that is not solvable by Ukelele. Zeta-calculus uses grids, an extension of simple grids (tables) used in z-notation see http://www.w3.org/TR/wsdl20/wsdl20-z.html (scroll down a bit to see some grids). Grids are special purpose tables that have a particular shape (only some of the lines are drawn in, which lines are are which are not define which type of grid it is). Making these manually from tables is a bore. It would be really great if an expert in one of those open source word processors would write an extension to support grids. I have begun documenting the requirements, but which one of those word processors would be easier to add that functionality I don't know, but that will really determine which word processor we use.
Yes, MS Word does have a character picker, but it only covers a selection of unicode (instead of showing all the characters in the font), which is a bit rubbish. MS word does allow unicode characters to be copied in. By using Ukelele as an input method I could use Word, but since going to Office 365 its reliability is too poor; I've stopped using Excel for the same reason.
Connor.
On Saturday, December 5, 2015 at 10:05:52 PM UTC, Geke wrote:O, I forgot:
Do some research, e.g. on Wikipedia: I think it’s better to switch to LibreOffice from your OpenOffice.
Both work OK with Ukelele’s keyboard layouts.
MS Word should have a character picker, I think: "Insert symbol..." or some such command. Otherwise use the Mac’s built in Character Viewer (or however it’s called in newer OSes) from the Keyboard Layout menu.
Is there anyone who has already mastered Ukulele who can put together updated documentation to help new users? I'm sure it would be much appreciated.
Regards,
Connor.
Oh dear. I've been really struggling trying to use Ukelele (3.0.0). There is a lot that is not intuitive, where I don't know what practical effect my actions are having, or what I need to do to have the desired effect. I think the application is fine, and I take my hat off to John for all his hard work in making it. There are two pieces of documentation, a PDF which cover some basics, and html which is out of date (because the user interface has changed so much, the old html simply does not correspond with todays Ukelele). The software does a really good job, but trying to work out how to use it from outdated and disjoined (its not obvious in which order to read them) documentation is frustrating; the application deserves better.
Is there anyone who has already mastered Ukulele who can put together updated documentation to help new users? I'm sure it would be much appreciated.
Oh dear. I've been really struggling trying to use Ukelele (3.0.0). There is a lot that is not intuitive, where I don't know what practical effect my actions are having, or what I need to do to have the desired effect. I think the application is fine, and I take my hat off to John for all his hard work in making it. There are two pieces of documentation, a PDF which cover some basics, and html which is out of date (because the user interface has changed so much, the old html simply does not correspond with todays Ukelele). The software does a really good job, but trying to work out how to use it from outdated and disjoined (its not obvious in which order to read them) documentation is frustrating; the application deserves better.
Is there anyone who has already mastered Ukulele who can put together updated documentation to help new users? I'm sure it would be much appreciated.
John,
Thanks for the practical advice.
"I'm sorry that things are not intuitive. I welcome any suggestions in improving the interface. There are many complex interactions that need to be handled, and I have been experimenting with different methods, none of which I'm entirely satisfied with."
Yes, creating an application does throw up a number of design issues. First I’m mention some small frustrations such as:
"New from Current Input source” - it delivered an error message telling me that the key layout file did not exist; this may be because there is no UK keyboard in the set of unicode layouts but there is one in the roman set.
It is not made clear why someone should select a layout from Roman verses Unicode (I presumed I should use Unicode, so based my layout on US key layout).
When choosing a name for my new keyboard, I thought that a new layout was being created; what actually happened is that the original US keyboard was renamed and my sandbox playing was ‘damaging’ the original file. What I ought to have done was duplicate and rename the original file, as well as renaming within Ukelele (or Ukelele should copy and rename the file in addition to what it normally does).
There is a keylayout ID, a negative number, but there was not indication how this is used, or what problem there may be if the number chosen is the same as an existing keyboard. Would there be any value in Ukelele identifying what ID numbers are already in use and suggesting a number that does not conflict? From a UI point of view, asking the user for a value without the user knowing what it is for is not good.
When I was playing, I made a dead key. When I realised I had changed the original file rather than my own file I wanted to remove the dead key, but could not see how this could be done.
These are my thoughts on the design side...
One big problem with user interfaces is modes. The more modes there are the mode difficult/confusing software tends to be. With Ukelele the keyboard window has a different mode for each combination of modifier keys, and a different mode for each dead key. An alternative approach may be better.
I wonder whether editing a list for each key would be easier, ie the keyboard window does not change mode, but, when a key is double clicked, instead of displaying just one input field for the character for that mode, it displays a list of fields, each field for one combination of modifier keys and other key states. Eg first field for the desired character from the plain keystroke, next field for the shifted keystroke…. with dead key states added at the lower end of the list. That way the user only has to define new keystates in a separate place (which is conceptually easy), then desired characters can be added to all keystates in a consistent manner, ie double click a key icon, and it shows the list of all possible states for the key and the user can drag and drop etc. That would avoid the more awkward mode changes.
I would recommend that the add new keystates panel should either not use a representation of a keyboard, or be visually different so that there is a clear difference between defining new keystates and placing character codes on the keys in each state. The add new keystates panel could be a table with one row for each key state, with 2 fields for keystate name, and the key sequence to enter that state (the sequence being entered by the user typing the keys on the physical keyboard while the cursor is in that field). I suspect that would be more intuitive. I hope that makes sense. I suspect it would work well if the add new keystates panel and the keyboard drag and drop panel are displayed one above the other in the same window, that way a user on opening the application will twig immediately the relationship between keystates, and keys, and will not need a manual.
If the double click could be replaced by either a single click, or a hover-over, the user doesn't need to be told that a double click is necessary.
Typinator sounds like a good product, especially because some symbols I wish to use may need different length key sequences, eg to get an upward pointing arrow, I want to type <some dead key> A (for arrow symbol set), N (for North), but I also want: <some dead key> A (for arrows), N E (for North East) to give an arrow that points diagonally up and to the right.
OpenOffice (like other word-processors) has an auto correct feature that replaces one string with another which would work well for me, however it has a couple of drawbacks. Entering new substitution values through the screen dialogue is time consuming for large numbers of symbols. Instead of this information being stored in a simple flat file that is easy to edit and easy to pass to other people, it is stored in an XML file which is then compressed, with other files, into a .dat file. I think that was a design mistake. Although that would work well for work in my specific field, it would of-course, not allow me to access those symbols when typing emails etc, which is why 'Typinator' or 'aText' may be my best option.
I hope that these thoughts are useful, and I wish you well.
Regards,
Connor
Yes, creating an application does throw up a number of design issues. First I’m mention some small frustrations such as:
"New from Current Input source” - it delivered an error message telling me that the key layout file did not exist; this may be because there is no UK keyboard in the set of unicode layouts but there is one in the roman set.
It is not made clear why someone should select a layout from Roman verses Unicode (I presumed I should use Unicode, so based my layout on US key layout).
When choosing a name for my new keyboard, I thought that a new layout was being created; what actually happened is that the original US keyboard was renamed and my sandbox playing was ‘damaging’ the original file. What I ought to have done was duplicate and rename the original file, as well as renaming within Ukelele (or Ukelele should copy and rename the file in addition to what it normally does).
There is a keylayout ID, a negative number, but there was not indication how this is used, or what problem there may be if the number chosen is the same as an existing keyboard. Would there be any value in Ukelele identifying what ID numbers are already in use and suggesting a number that does not conflict? From a UI point of view, asking the user for a value without the user knowing what it is for is not good.
When I was playing, I made a dead key. When I realised I had changed the original file rather than my own file I wanted to remove the dead key, but could not see how this could be done.
These are my thoughts on the design side...
One big problem with user interfaces is modes. The more modes there are the mode difficult/confusing software tends to be. With Ukelele the keyboard window has a different mode for each combination of modifier keys, and a different mode for each dead key. An alternative approach may be better.
I wonder whether editing a list for each key would be easier, ie the keyboard window does not change mode, but, when a key is double clicked, instead of displaying just one input field for the character for that mode, it displays a list of fields, each field for one combination of modifier keys and other key states. Eg first field for the desired character from the plain keystroke, next field for the shifted keystroke…. with dead key states added at the lower end of the list. That way the user only has to define new keystates in a separate place (which is conceptually easy), then desired characters can be added to all keystates in a consistent manner, ie double click a key icon, and it shows the list of all possible states for the key and the user can drag and drop etc. That would avoid the more awkward mode changes.
I would recommend that the add new keystates panel should either not use a representation of a keyboard, or be visually different so that there is a clear difference between defining new keystates and placing character codes on the keys in each state. The add new keystates panel could be a table with one row for each key state, with 2 fields for keystate name, and the key sequence to enter that state (the sequence being entered by the user typing the keys on the physical keyboard while the cursor is in that field). I suspect that would be more intuitive. I hope that makes sense. I suspect it would work well if the add new keystates panel and the keyboard drag and drop panel are displayed one above the other in the same window, that way a user on opening the application will twig immediately the relationship between keystates, and keys, and will not need a manual.
If the double click could be replaced by either a single click, or a hover-over, the user doesn't need to be told that a double click is necessary.