Chris Mellon wrote: > wx does (in large part), but most likely the problem is that the "rich > text" control used in most editors is not the win32 rich text control, > but instead Scintilla, which is designed for source editing and is > much easier to use. Very few editors, of any kind, use the native > win32 text control for source highlighting. > > wx does have (some) support for the accessibility features in win32, > you might post a feature request on the wx bug tracker to add them to > the wx platform bindings for Scintilla. > > The main reason editors don't use the standard control is for syntax > highlighting and perhaps folding and margins, though, which I'm not > sure are especially valuable to you. What kind of features makes a > Python editor 'smart' for someone who's coding with a screen reader?
what you said makes wonderful scents. Thank you for explaining. Now, if you read the application note, after the list of rich text objects they expect, they described a fuzzy in between state where if you tweak the configuration, you get Select-and-Say but you know a whole bunch of things about the objects etc. would you be so kind as to read over that part of the application note and tell me if it applies to any of the Scintilla objects for text display? I'm glad to hear you have some of the extensibility features. Some are better than none. There is an event coming up in the next few weeks that will trigger a need for accessibility interfaces on the Linux side. Well, I'm not sure about a screen reader but I'm using speech to text. My apologies if I wasn't clear (reading back, I see a couple dropouts that I didn't catch and change the meaning significantly. As a brief aside, one of the classic problems is can versus can't. Which leaves you very much in doubt if someone writes "I can go to bed with you." Is it a misrecognition or an expression of desire? only your natural language processing will know for sure:-) there's a whole hierarchy of needs. The voice coder project has done some nice work in that domain. For me, the fundamental level is the ability to correct and replace without error. I should be up to select a phrase or set of words and have that region highlighted properly so I can operate on it. I should be able to delete the last utterance and not have it count wrong. the next level (that the voice coder project handles) is the ability to dictate a word and depending on the context or its knowledge of symbols, generate a code symbol be a bumpy cap or joined with underscores or some combination thereof. Additionally, David Fox, primary developer, did some absolutely gorgeous worked with correction mechanisms. The next level up would be contextual awareness of variables within your scope so that when you dictate something, you don't have to have a big collection of static symbols. You create them dynamically based on where you are located in your code, which file, and what modules you include. This last one is going to be difficult. May not show up for a few years. But all I'm asking for right now is simple Select-and-Say as outlined in that application note from nuance. I can say "_" with English words on either side and the code will work. -- http://mail.python.org/mailman/listinfo/python-list
