[Feature Request] Customizable Character Descriptions

[school510587]

For languages not based on spelling alphabets, we need a character description to distinguish characters having the same pronunciation. To improve the efficiency of reading/writing, the description of every character should be designed to be recognizable via hearing, and this is the aim of l10n teams. However, due to different background and experience such as the process of learning in elementary school and/or working environment, it is impossible to find one description giving everyone the best working efficiency. Especially, the most recognizable description may be entirely different for two visually disabled users having poor hearing ability. Thus, I think NVDA needs a way to let users customize the character descriptions. Would you develop this feature? Many thanks!

[josephsl]

Hello, can you provide use cases and examples, including languages that could benefit from this and how this will work for beginners? Thanks.

From: school510587 [mailto:notifications@github.com] Sent: Thursday, January 26, 2017 2:30 AM To: nvaccess/nvda nvda@noreply.github.com Cc: Subscribed subscribed@noreply.github.com Subject: [nvaccess/nvda] [Feature Request] Customizable Character Descriptions (#6811)

For languages not based on spelling alphabets, we need a character description to distinguish characters having the same pronunciation. To improve the efficiency of reading/writing, the description of every character should be designed to be recognizable via hearing, and this is the aim of l10n teams. However, due to different background and experience such as the process of learning in elementary school and/or working environment, it is impossible to find one description giving everyone the best working efficiency. Especially, the most recognizable description may be entirely different for two visually disabled users having poor hearing ability. Thus, I think NVDA needs a way to let users customize the character descriptions. Would you develop this feature? Many thanks!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nvaccess/nvda/issues/6811 , or mute the thread https://github.com/notifications/unsubscribe-auth/AHgLkNJbBylqptrIfikjgzASxSXMJ4sIks5rWHWZgaJpZM4LufVC .

[school510587]

Sorry, I only speak Chinese and English, and Chinese is one important example of language. Chinese characters were made mainly based on the shapes of things, so there are many characters sharing the same pronunciation. Visually disabled Chinese users can only (quickly) recognize one character through phrases that are often corresponds to it. However, criteria of good phrases are very subjective. When one user thinks some phrase isn't recognizable enough and most of other users don't think so, the l10n team always asks he/she to modify characterDescriptions.dic manually, which will be rewritten on NVDA upgrade. I think there are 3 operations on characterDescriptions.dic tables, addition, removal and reordering. Actually, users dispute about the order of phrases in most situations.

[josephsl]

Hi, recently I and another Korean translator went through a similar process. I think the best solution would be to tell Chinese translators to update character descriptions to account for variations. Thanks.

From: school510587 [mailto:notifications@github.com] Sent: Thursday, January 26, 2017 7:42 AM To: nvaccess/nvda nvda@noreply.github.com Cc: Joseph Lee joseph.lee22590@gmail.com; Comment comment@noreply.github.com Subject: Re: [nvaccess/nvda] [Feature Request] Customizable Character Descriptions (#6811)

Sorry, I only speak Chinese and English, and Chinese is one important example of language. Chinese characters were made mainly based on the shapes of things, so there are many characters sharing the same pronunciation. Visually disabled Chinese users can only (quickly) recognize one character through phrases that are often corresponds to it. However, criteria of good phrases are very subjective. When one user thinks some phrase isn't recognizable enough and most of other users don't think so, the l10n team always asks he/she to modify characterDescriptions.dic manually, which will be rewritten on NVDA upgrade. I think there are 3 operations on characterDescriptions.dic tables, addition, removal and reordering. Actually, users dispute about the order of phrases in most situations.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nvaccess/nvda/issues/6811#issuecomment-275421259 , or mute the thread https://github.com/notifications/unsubscribe-auth/AHgLkN1UJIRpCB5zesoYmU7xiak2rzVVks5rWL7cgaJpZM4LufVC .

[school510587]

Hi, Joseph Lee, do you mean that the l10n team maintains different versions of characterDescriptions.dic for each domain of demands? If no, could you describe it more in detail? Thanks. What Chinese users care about is the first phrase of each character, which is presented when the character candidate menu is dropped down. It is very common that two users want different first phrase. L10n team can only obey the majority of users. However, if many versions of characterDescriptions exist, I also think it's not a good way to replace the characterDescriptions of the system.

[josephsl]

Hi, I think I know what could be happening. If I’m understanding you right, each user would hear different pronunciation for same phrases, or want to hear certain characters shown before others. If so, this is a bit beyond the domain of this issue tracker – this falls into localization. Can you ask NVDA Chinese translators and users (such as @VGJH2005) so a possible solution could be developed? Thanks.

From: school510587 [mailto:notifications@github.com] Sent: Thursday, January 26, 2017 8:18 AM To: nvaccess/nvda nvda@noreply.github.com Cc: Joseph Lee joseph.lee22590@gmail.com; Comment comment@noreply.github.com Subject: Re: [nvaccess/nvda] [Feature Request] Customizable Character Descriptions (#6811)

Hi, Joseph Lee, do you mean that the l10n team maintains different versions of characterDescriptions.dic for each domain of demands? If no, could you describe it more in detail? Thanks. What Chinese users care about is the first phrase of each character, which is presented when the character candidate menu is dropped down. It is very common that two users want different first phrase. L10n team can only obey the majority of users. However, if many versions of characterDescriptions exist, I also think it's not a good way to replace the characterDescriptions of the system.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nvaccess/nvda/issues/6811#issuecomment-275431902 , or mute the thread https://github.com/notifications/unsubscribe-auth/AHgLkOf8qzfdgjTMW2fFswv9V_B4TKo4ks5rWMcngaJpZM4LufVC .

[school510587]

Oh sorry. It isn't what I meant. Every user hears the same pronunciation for the same phrase, assuming that issues about the voice synthesizers are ignored in this discussion. However, different users may map the same pronunciation into different characters. You may think this is impossible. I have pointed out that some users also have limitted hearing ability, and they can only guess what they just heard if the phrase sounds like some other phrase. (In fact, I am also one of such users...) Of course, I can ask the l10n team to consider my (and other users with poor hearing ability) situation. But, the majority of users have good hearing ability, and they will be influenced if the first phrases are changed. This can't be decided only by one l10n team, because it can only take care of demands agreed by most of users. So I proposed this feature request. characterDescriptions.dic needs a mechanism for customizing like symbols.dic, the presentation of special symbols.

[josephsl]

Hi, I see. @VGJH2005, @Nishimotz, @MichaelDCurran, any thoughts?

From: school510587 [mailto:notifications@github.com] Sent: Thursday, January 26, 2017 9:04 AM To: nvaccess/nvda nvda@noreply.github.com Cc: Joseph Lee joseph.lee22590@gmail.com; Comment comment@noreply.github.com Subject: Re: [nvaccess/nvda] [Feature Request] Customizable Character Descriptions (#6811)

Oh sorry. It isn't what I meant. Every user hears the same pronunciation for the same phrase, assuming that issues about the voice synthesizers are ignored in this discussion. However, different users may map the same pronunciation into different characters. You may think this is impossible. I have pointed out that some users also have limitted hearing ability, and they can only guess what they just heard if the phrase sounds like some other phrase. (In fact, I am also one of such users...) Of course, I can ask the l10n team to consider my (and other users with poor hearing ability) situation. But, the majority of users have good hearing ability, and they will be influenced if the first phrases are changed. This can't be decided only by one l10n team, because it can only take care of demands agreed by most of users. So I proposed this feature request. characterDescriptions.dic needs a mechanism for customizing like symbols.dic, the presentation of special symbols.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nvaccess/nvda/issues/6811#issuecomment-275445693 , or mute the thread https://github.com/notifications/unsubscribe-auth/AHgLkCACTwPl6FssLQzKvSMo7Yc--tn3ks5rWNHxgaJpZM4LufVC .

[surfer0627]

@josephsl

Hi, recently I and another Korean translator went through a similar process. I think the best solution would be to tell Chinese translators to update character descriptions to account for variations.

Hi, Due to different background and experience, some users customize character descriptions by editing file "characterDescriptions.dic"". While updating to a new versionm, the file would we overwrite. This is one of the reasons why users request this feature. As we know, speech dictionaries files are stored in a special NVDA directory located in your Windows user profile. While updating, users still have these files saved.

[school510587]

By the way, there is another way to solve this problem. As what I just mentioned above, the (relative) minority of users with poor hearing ability need this functionality to adjust characterDescriptions according to individual condition without influencing other users. We can also assume these users read messages via Braille displays, and NVDA is now showing all character descriptions after each character when the focus is on the candidate list of an input method. With all the descriptions shown, one can determine what the character is with high confidence. However, NVDA doesn't change what displayed on the Braille display synchronously when the user moves the focus among those candidates. Precisely, NVDA always shows the first character followed by its all descriptions. Users can't read the list only via a Braille display. It is another issue and is beyond the scope of this discussion, but it is an alternative solution. Compared to the alternative solution, customizing characterDescriptions benefits more users. Each user can maintain his/her own small changes of characterDescriptions when his/her individual hearing habit is different from the majority of users for any possible reason. Sorry, I don't know whether someone mentioned a related issue about the alternative solution previously. But if NVDA developers think that fixing this bug is more important, I also think this issue can be closed after the bug becomes fixed. Thanks. (If no other user points out that this feature request has some functionality which can't be satisfied after the bug fix.)

[vgjh2005]

Hi guide: I have read all your comment written in github. But I cannot sure your problem.

1. The character file has been written, and I think it is enough mature to use to seperate most characters that use same pronounciation. If you have any quesstion, please send me an e-mail directly in Chinese to describe your topic.
2. If you are talking about IME and TSF support issues, I think we need help from @MichaelDCurran. You can do some search to get relative issues. If you have any thought, please comment below. Thanks!

2017-01-27 3:27 GMT+08:00, school510587 notifications@github.com:

By the way, there is another way to solve this problem. As what I just mentioned above, the (relative) minority of users with poor hearing ability need this functionality to adjust characterDescriptions according to individual condition without influencing other users. We can also assume these users read messages via Braille displays, and NVDA is now showing all character descriptions after each character when the focus is on the candidate list of an input method. With all the descriptions shown, one can determine what the character is with high confidence. However, NVDA doesn't change what displayed on the Braille display synchronously when the user moves the focus among those candidates. Precisely, NVDA always shows the first character followed by its all descriptions. Users can't read the list only via a Braille display. It is another issue and is beyond the scope of this discussion, but it is an alternative solution. Compared to the alternative solution, customizing characterDescriptions benefits more users. Each user can maintain his/her own small changes of characterDescriptions when his/her individual hearing habit is different from the majority of users for any possible reason. Sorry, I don't know whether someone mentioned a related issue about the alternative solution previously. But if NVDA developers think that fixing this bug is more important, I also think this issue can be closed after the bug becomes fixed. Thanks. (If no other user points out that this feature request has some functionality which can't be satisfied after the bug fix.)

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/nvaccess/nvda/issues/6811#issuecomment-275486810

[school510587]

Hi, @vgjh2005

I have sent a mail describing my original problem to vgjh2005@gmail.com in traditional Chinese. Thanks for your reply! If there are other possible solutions, please tell me.

[school510587]

@vgjh2005 gave me some idea about this feature that I wanted. The idea is a simpler way to implement. The problem is that self-defined characterDescriptions.dic will be rewritten on every update, so there can be checkboxes (default to checked) during update process to select whether an user wants to update l10n data. (3 boxes for nvda.mo, characterDescriptions.dic and symbols.dic)

Any point of view? (@vgjh2005, please correct me if any of my statements is wrong or not complete enough.) Thanks!

[vgjh2005]

Hi: There are a bit difference. I think only add an option to decide whether update language folder when installing NVDA. But I don't know which one is the best option. Thanks!

2017-02-12 18:09 GMT+08:00, school510587 notifications@github.com:

@vgjh2005 gave me some idea about this feature that I wanted. The idea is a simpler way to implement. The problem is that self-defined characterDescriptions.dic will be rewritten on every update, so there can be checkboxes (default to checked) during update process to select whether an user wants to update l10n data. (3 boxes for nvda.mo, characterDescriptions.dic and symbols.dic)

Any point of view? (@vgjh2005, please correct me if any of my statements is wrong or not complete enough.) Thanks!

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/nvaccess/nvda/issues/6811#issuecomment-279208350

[jcsteh]

We definitely don't want to add options which relate to what gets updated in the program directory. The program directory is never supposed to be touched by the user. If language communities really can't come to a consensus as to the best character descriptions for the majority, I guess we'll need to make them user configurable as you originally suggested.

[school510587]

Thank @jcsteh for the response. Please let me present another use case of this feature, which I have described in Chinese in letters sent to @VGJH2005. Input methods (IMs) depending on pronunciations of characters are intuitive to learn, but they are not efficient enough for blind users whose careers need high level of correctness in writing. These users have to learn IMs based on shapes of characters. However, it's hard for blind users to remember the shape of each character, so they must maintain a version of characterDescriptions to remind them of character shapes at any time. It is not the demand of everyone, and descriptions for shapes may be very tedious. Thus, the l10n team must refuse to modify characterDescriptions, because most of users require that characterDescriptions be short and simple. The conclusion is that this feature is still be demanded. Thanks!

[bhavyashah]

@jcsteh Do we have an accurate picture of the UX of this feature request, or is more discussion necessary before the same can be ascertained properly?

[ehollig]

This issue has other use cases as well. As stated in #4083:

There are more generic use cases where one may wish to alter one's character descriptions. Say, for instance, a user has learning difficulties and would prefer F to be defined as "Fred" instead of "Foxtrot". This should absolutely be done at the user level, not the factory level.