tw1ddle / blind-crossword-3d Goto Github PK

In the clues table, they are represented as unicode replacement characters.

http://www.fileformat.info/info/unicode/char/fffd/index.htm

Presumably a wrong character encoding is being used somewhere down the line.

Would make ruling out misspellings or misplacement of punctuation easier when the text to speech engine appears to make a mistake

.xwc files are 2D crossword files from Spoonbill Software's BG Crossword Puzzle program.

Blind Crossword 3D can load these but does it always load/save them correctly?

As part of this, the "Day" column of the table needs hiding (and so does the sort-by-day and voicing) when non-xwc3d crosswords are loaded

Emails fail to open on OSX Mountain Lion, using QDesktopServices to do it and with only the default mail client installed

The spec for the .xwc3d format needs writing up.

Some users will want to use familiar controls e.g. Spoonbill's BG Crossword type controls. Make this the default, then allow switching to others

In crossword puzzle clues, bits of punctuation like dashes and exclamation points can be very important. Make sure they are spelled out by the voice, even when speaking "naturally".

On Windows:

"I cannot decompress the program file in a directory with a funny name with spaces and "(" (like Program Files(x86))"

If SAPI can be used without ATL, which it probably can, then remove the dependency

Could be useful for catching errors where full stops or punctuation is missed.

Currently guesses cannot be sorted by length - as the length and guess column contents are sorted as strings. Make it possible to sort them by length by clicking the columns/hitting the sorting key

Another font size incongruity

The "file", "options", "about"... items on the main menu bar may need to be resized in order for partially sighted users to see them.

If a puzzle has several grid highlights that consist of multi-word or separate answers, it is necessary for the program to be able to read the contents of the highlights one by one, rather than as a single string.

Play it asynchronously and keep it short (couple sentences).

Accessing local web pages should use "file:" protocol, not http

Solvers could bring up a list of suggested words which they could try to match to the partial solutions to words.

Chambers Standard Reference Dictionary would be the most appropriate for this list.

I am pretty sure I am missing the idea here, as all the dynamic casts of the view's model to a QSortFilterProxyModel don't look like they don't belong there.

Eric suggested that the currently selected (and highlighted) item in the graphical grid should be displayed in large text to minimize the amount of zooming and panning a partially sighted person might have to do.

So, add an enlarged view of currently selected guess in an appropriate place (possibly at the top of the graphics view, if it doesn't get in the way)

Make the grid items into wedge-shapes to make them look nicer. Also, only draw the centres of discs once to avoid anything like z-fighting or simply blank grid items obscuring letters in the centre

What happens if someone runs the program on a computer with non-English default language?

Since, for the forseeable future, crosswords will be all English, maybe just set the default voice language to English.

In the event that you have a word like this:

..someletters.....

It would be read "two dots someletters two dots two dots dot". Otherwise it's fine, but the regex used doesn't cover all cases.

With a fresh install of Windows 8, emails fail to open.

Is this because there's no default email client set or something else unavoidable, or is there a reason? Need to check it out.

Use accessibility API, these posts may be useful:

http://juliankniephoff.wordpress.com/2011/03/27/use-objective-c-classes-in-cpp-opaquely-readably-and-safely/

http://juliankniephoff.wordpress.com/2012/09/22/revisiting-using-objective-c-objects-in-c-classes-in-the-age-of-arc/

It would be nice to zoom in or out of the clues table in order for a sighted user to get a comfortable view of the table.

You cannot pan vertically on the graphical grid view. There should probably be some extra flexibility as it would allow users to arrange the crossword grids as they liked.

Add an option to set the printing size for answers. Or perhaps add it as a config file option. What's important is that the answers are large enough for a partially sighted person to read.

The voice gets interrupted when the dialog opens and stops speaking the line about the print dialog opening, but continues speaking normally from thereon

Enable a solver to find and solve the clues that intersect another one they just solved.

Multi-word solutions have spaces or dashes between individual words. These need to be taken into account when reading out guesses, as otherwise the voice will mispronounce them

Icons show up left of the filenames in the load file dialog box, but there is no reason to have them as the files have no special icon of their own. They also look pretty ugly on Windows XP.

This would give blind solvers a real feel for what a 3D crossword grid is like.

Zooming in the graphical view takes you further to the left when your mouse is to the left of the centre of the view, than to the right.

The main menu font is too small and partially sighted users might have difficulty with it. Is this controlled by the system or can these font sizes be changed?

With a goal of making it easier for nontechnical users to get started. This would be part of adding:

• File associations
• Accepting drag-and-drop files to load,
• Add desktop/bar shortcuts

For instance, the quit dialog has small text whereas 20pt Lucida Console is specified. Maybe OSX doesn't have it by default. What is a font that Win/Mac/Linux all use and isn't ugly?

Doxygen documentation needs putting up on the gh-pages, and more extensive commenting needs adding

There is a new puzzle type coming out in next year's calendar that is somewhat like a jigsaw.

Wait for more information about this, will probably require some implementation for loading/saving this type.

Idea for the future, would be cool if speech recog is good enough for recognising keywords needed for getting around the program e.g. navigating the table and reading stuff... Without needing (much/any) individual training.

Speaking the guesses would likely be way too inaccurate, but worth a look too.

Probably using festival/flite

At the minute users have to switch between using the program and reading help/tutorials. Having a tutorial video to talk through/explain steps would help people get to grips faster, especially partially sighted/fully sighted solvers

Font sizes are hardcoded in a few different places. Let's move those together and add an option to increase/decrease font sizes throughout the app to suit the needs of the different classes of partially sighted user.

A tool within the program that helps solvers find solutions to clues whose answers are anagrams.

If a user misses the last thing spoken or takes a break, then it is possible for them to get disoriented. Let's enable users to listen the last thing that was said to help avoid this.

Reduces the number of controls user needs to remember

Some syntactical changes might be needed to compile the program using GCC, and different library versions (Qt sure)... will SAPI work?

Would make it easier to go about getting it working on OSX and Linux

Some puzzles have background images. It would be nice to be able to do something to maintain the aspect ratio on these images. Such as, by keeping the image centered in the view, instead of stretching them when the graphical grid view is resized.

JAWS can't read the file/options/about main menus.