New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Certain bullets not reported in Microsoft Word #5267
Comments
Comment 1 by leonarddr on 2015-08-05 09:17 |
Comment 2 by csm120 on 2015-08-05 09:22 |
Attachment Word list test.docx added by leonarddr on 2015-08-05 09:43 |
Comment 3 by leonarddr on 2015-08-05 09:44 |
Comment 5 by leonarddr on 2015-08-05 10:02 |
Comment 6 by vrdhn on 2015-09-13 06:21 However, the o and dash used by word is actually standard ASCII. With symbol level most, the dash is pronounced correctly. With fix to ticket #2446, the level number will be anounced, so user will be able to undertand that there is a list under the caret. However, another idea, worth exploring, is to also replace o and dash by the visibaly-equivalent unicode symbol, but only for Bullet character. Any comments on this ? |
Attachment bullet-types-5267.docx added by vrdhn on 2015-09-13 07:11 |
Attachment bullet-types-5267.2.docx added by vrdhn on 2015-09-13 15:57 |
Comment 8 by vrdhn on 2015-09-13 16:03 As discussed over seperate private mails, the patch will map the most common used character-points from winword's PUA to the corrosponding unicode symbols. A few missing symbols have been added to symbol.dic file. It's worth noting the limitations of this approach here.
Thanks |
@LeonarddeR, do you have your symbol level set to most or all? For me, dash bullets don't get reported at level none or some because dashes don't get reported at those levels. However, they do get reported at most or all. I don't think there's really much we can do about this unless there's some semantically better Unicode symbol for a dash bullet. |
@dkager, @Qchristensen, @michaelDCurran, thoughts on how common this is? Just trying to work out what priority we should give this. |
list items denoted with a dash I would guess are not as common as asterisk / bullet list items, but I think they are still used quite regularly. Particularly as it's an easy conversion from plain text. Another way around it could potentially be to check the style of the text - any list will default to "List paragraph". That way it wouldn't matter what symbol was used. The issues with that would be that both bulleted and numbered lists are identified as "list paragraph", and also it's possible to create custom styles which work as a list. Theoretically it's possible to set text in the middle of a paragraph to be "list paragraph" too but I wouldn't expect that to be very common. As it is now, the list is automatically indented, and assuming NVDA is set to announce indentation, that is announced as you enter a list. Back to the original point, I think the easiest to understand for the user would still be to find a way to announce that a dash list is still a list, even if the current symbol level wouldn't normally announce dashes. |
Sorry, missed this one when searching for an existing issue. Lists with asterisks are quite common. The dashed lists I think already use a plain dash, but the asterisk is a Private Use codepoint that is brailled but not translated by liblouis, e.g. '\xf0b7'. |
I was confused because the asterisk bullet reads fine with speech. However,
it looks like we mapped U+F0B7 to bullet in symbols.dic. That's a total
hack and something we probably shouldn't have done.
|
@jcsteh: Indeed, changing symbol level to most or all does report them, but I'd honestly prefer a fix to have them read for the some level as well. |
At first glance the associated PR looks good, though I wonder if we shouldn't replace the PUA codes with ASCII. Even U+2022 (bullet) isn't in all braille tables. On the other hand it could be argued that those symbols are closer to what the PUA ones actually look like. |
@leonardder commented on 11 Jan. 2017, 5:02 pm AEST:
I don't think that's possible beyond changing dash to level some (which we don't want to do). Even if we said that all punctuation should be read regardless of level for line prefixes (bullets, numbers, etc.), then you'd get the full stop being read for numbered list items, which would annoy users as well. The reality is that the dash is just used for way too many things. @dkager commented on 11 Jan. 2017, 7:35 pm AEST:
I totally accept that the Unicode characters aren't in a lot of tables, but using ASCII feels like a hack to me and is likely to cause us problems in the future. For speech, we'd then have to map those ASCII characters to some words and those words could be wrong or confusing depending on context. |
One final concern: is the PR complete with regards to the possible bullets you can encouonter? And do the replacements visually match? I haven't been able to find a good reference for all standard bullet types. |
The PR is very probably incomplete in terms of available symbols. You can
probably use just about any symbol in Word as a bullet and I imagine that
is a pretty big list. That said, I don't think we need to be "complete"
here; at the very least, we want an extensible framework with the most
common bullets.
|
We should probably try to include all bullets you can type on the keyboard, e.g. dash, asterisk, and the fancier bullets like => and <>. |
Could someone please summarize the limitations of the current implementation and what work still remains for this issue to be fully addressed? |
WRT the braille side of things, an update is in #6778 (comment). |
Reported by csm120 on 2015-08-05 09:05
I have tried this in NVDA 2015.2 and in the latest snapshots.
Steps to reproduce:
The text was updated successfully, but these errors were encountered: