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
Attempt to guess the label for unlabelled form fields on web pages #84
Comments
Comment 1 by jteh on 2008-05-03 23:28 There is no reliable way to determine a label for a field when the page author has not specified a label correctly. Some other screen readers use the previous piece of text, but this is not always correct. Indeed, I have observed cases where such incorrect reporting is actually very confusing. I am leaving this open for discussion as to whether "guessing" the name for such fields is something we want to implement. While the fault is really with web pages, I understand that there are unfortunately far too many pages that don't correctly label form fields. However, I would argue that potentially incorrect reporting is perhaps worse. |
Comment 2 by chris1988 (in reply to comment 1) on 2008-06-10 07:23
I really cant think of anything or any way to improve on this. do you want me to shut the ticket? Does anybody else have any ideas? |
Comment 3 by jteh on 2008-12-05 03:06 |
Comment 4 by kennydog (in reply to comment 3) on 2009-01-14 22:55 |
Comment 5 by jteh (in reply to comment 4) on 2009-01-14 23:16
I think you're missing a key point here. There is no field name for these fields - we have to guess the field name. (We already read fields that have definite labels.) First of all, I'm not sure how we can make this guess. Second, it may be inaccurate/incorrect, and as I noted above, that can be highly confusing. Note that if a name is not present, you know to check it manually. If a name is present, though, you don't necessarily know to check it manually, even if it is an incorrect guess.
There's no specific method to query a field name. You just read the text around the field on the page. Where a field is unlabelled, that's all you can do. The idea is that generally, the label will be close to the field. |
Comment 6 by jteh on 2009-07-05 11:14 |
Comment 8 by k_kolev1985 on 2013-04-14 11:00 |
Comment 9 by briang1 on 2013-04-14 14:34 Maybe in the survey Monkey case its that the person who writes the form is able to make bad choices from the access point of view, in which case maybe the vendors of such online systems need to look at this as well. |
@jcsteh: It seems you started working on this years ago, is there still a desire to implement this? |
I think we still want this, though there's some controversy over whether it should be disabled by default. While I think this feature is important for usability, I also think users really need to know when the information they are receiving is possibly inaccurate. Mislabelling fields could be super problematic in a finance system, for example. Having said that, I'm not working on this at present. |
Many users say that in most of the cases label detection would be useful.
From: James Teh [mailto:notifications@github.com]
Sent: Wednesday, June 14, 2017 4:22 AM
To: nvaccess/nvda <nvda@noreply.github.com>
Cc: Subscribed <subscribed@noreply.github.com>
Subject: Re: [nvaccess/nvda] Attempt to guess the label for unlabelled form fields on web pages (#84)
I think we still want this, though there's some controversy over whether it should be disabled by default. While I think this feature is important for usability, I also think users really need to know when the information they are receiving is possibly inaccurate. Mislabelling fields could be super problematic in a finance system, for example.
Having said that, I'm not working on this at present.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub <#84 (comment)> , or mute the thread <https://github.com/notifications/unsubscribe-auth/AE08vzfpxNaunrOnVjk3oChMTuXUn0jSks5sDxKpgaJpZM4N49Jo> . <https://github.com/notifications/beacon/AE08v_t1AL8RwLre_bOvMqlncq1Ypu7lks5sDxKpgaJpZM4N49Jo.gif>
|
@LeonarddeR the gmail form is properly recognized by NVDA including all labels. Thus, I suggest closing this issue in favor of #8335. |
The GMail example is not reproducible here, see my comment above. Thus, closing this issue in favor of the referenced one above. |
Reported by chris1988 on 2008-05-03 19:18
is it at all posible to have nvda read the titles of a form field. when going to fill out a form such as on:
https://www.google.com/accounts/NewAccount?service=mail&continue=http%3A%2F%2Fmail.google.com%2Fmail%2Fe-11-10ca38dcb38a7e30861186cae6d0522b-d070ccae47d4bd7d2b6600c9e6c69e02aa32349b&type=2
is it at all posible to have it read "first name" last name etc? without having to turn vitural buffers off and back on.
Blocking #3153
The text was updated successfully, but these errors were encountered: