Wow. :( "I'm confused. Why would you (a blind user) want to know what the poster frame is? How does it affect you?"
That is pretty bad.
I think they need a group of people with disabilities to review all of the specs. There have to be many more issues like this lurking (that those with disabilities could probably already point me to).
I think it probably sounded that way due to the editorializing of the headline. If instead it had read "Editor of HTML5 tries hard to gain a better understanding of potential use cases facing people with disabilities" you might have read it differently.
A non-sighted user hasn't managed to get to a video without somehow navigating their way there. The people replying to the tracker seem to be under the assumption that a non-sighted user will just mystically be presented with a video and not know what it is all about. I can't imagine a case where this is likely.
Navigating their way there gives them context to what the video will be (in the use case that the video is on a blank page). If the page contains textual content then the user will get context from the text. If the video is completely unrelated to location context or to page-level context then the content will be of little to no interest, so providing an alt description adds no real-world value.
Also note that unless the video has audio captioning then the entire video is of little to no interest. Imagine watching something like Lost in Translation with no visual input and no descriptive captioning.
Note that this is not a question of providing an alt text for e.g. an image of a movie poster on IMDB.
The 'poster' in this case is specifically a single image preview for a video shown before the video starts playing. According to the spec "typically one of the first non-blank frames".
It can be compared with e.g. a low resolution preview of a high-resolution image, as was used in the early days of the web. It is definitely not obvious that it would make sense to have an alternative text specifically for the lo-res preview as distinct from the alt text for the image itself.
Note that the video element already support an alternative text description - its just the single-frame preview which doesn't support a separate text.
I think Ians questioning of this feature is completely valid.
The only objection I can think of is that the poster can be intended to convey meaning. If this meaning is entirely lost without an alt text, then the lack of an alt text for the poster is an oversight.
But then again the use of colors in an image conveys meaning. That does not mean that there should be alt texts for each color channel. The alternative content should be provided at a level that makes sense if you don't have access to the specific media.
The HTML5 spec specifically addresses accessibility for video like this:
To make video content accessible to the blind, deaf, and those with other physical or cognitive disabilities, authors are expected to provide alternative media streams and/or to embed accessibility aids (such as caption or subtitle tracks, audio description tracks, or sign-language overlays) into their media streams. (http://www.whatwg.org/specs/web-apps/current-work/multipage/...)
This is much more sensible that providing an alt text for the first frame in the video.
The standard should support an accessible version of what is staticly and dynamically displayed in the browser to a non-disabled person. If a non-disabled person will always see that frame on load, that frame needs an accessible way of making sense of that frame.
On the one hand, according to the current spec it's supposed to be a representative frame of the video. On the other hand, it's likely to be used to convey additional information about the video by content producers. (Title of the video/performers/location - think about how many people have inserted a single frame in a YouTube video in order for it to appear as the poster)
I think the question is whether or not other attributes on the video tag can convey that information, making it redundant, or will the poster likely be used to convey different information.
If you have a <video> element, you can control what frame is displayed before the video is played, by inserting a <poster> element into it. If the page author doesn't explicitly specify a poster frame, the browser is free to choose a frame to display.
The person who filed the bug is (effectively) saying "if the page author specifies a poster frame, it should be possible to add alt text. If the page author uses the default poster frame, it should be impossible to add alt text."
I believe the question Hixie was getting at was "why would you (a blind user) care about an explicitly-set poster frame but not a default poster frame?" which, I think, is reasonable.
I'm not surprised that he has trouble understanding that. The 'meaning' of the poster frame isn't clear from that conversation. Does he want a description of what's being displayed in just the poster frame? How does this differ from the title of the video?
For example:
Title: Cat jumping on bed
Frame: Image of cat jumping towards bed
Video: Cat jumps toward bed and lands on it.
The vast majority of the time, the poster frame isn't going to be different than the title of the video.
Now, I agree that when it is different, there should be a way to specify an alt text for it. But to require it when it's the same as the title seems silly to me.
To use a non-hypothetical example, right now I'm watching a series:
Title: Dexter.
Description: A likable Miami police forensics expert moonlights as a serial killer of criminals who he believes have escaped justice.
The exact episode also has those attributes:
Title: My Bad.
Description: In the aftermath of last season's chilling finale, Rita's death has left Dexter feeling responsible, possibly even...guilty. Unable to deal with the trauma, Dexter makes a drastic decision that will impact everyone around him. Quinn stirs up trouble at the station when he notes that Rita's murder, which is being pinned on the Trinity Killer, doesn't fit his normal murder profile. Things get even more heated when Quinn suggests to LaGuerta that they look at the husband: Dexter Morgan. Meanwhile, Quinn offers Debra much needed support, causing her to see Quinn in a new light.
----
Simply put, titles and descriptions are not the same thing. It seems silly to me that somebody would advocate that they should be treated as the same thing just because there is overlap when the content is a 15 second video of a cat.
>But to require it when it's the same as the title seems silly to me.
Isn't the problem there that one doesn't know it is the same. Requiring a description of the poster image, which can be blank as for alt-attributes of non-content images, means that it clearer when the author has simply chosen not to bother attempting to be accessible as opposed to the poster having no useful content beyond the title.
Isn't he simply questioning an assumption? (assumption being that every bit of image always needs an alternative that is not in normal flow of page)
Ian's often repeated naive-sounding "Why would you do that? What's the use case?" has shot down many feature requests that seemed good in theory, but weren't that useful in practice.
Perhaps description of the video makes description of poster redundant? Maybe one can use ARIA? Maybe media framework can extract description from the video file?
Even if there is a way to describe the poster, will authors actually care to do it? Will they ignore it (naively assuming blind users don't watch videos) or fill it with boilerplate?
> Perhaps description of the video makes description of poster redundant? Maybe one can use ARIA? Maybe media framework can extract description from the video file?
Or may it's an honest question. May he's just skeptical(I am) and really wants to know. After-all there are no stupid questions, only stupid people, right?
The OP of the issue decides on which movies to watch based on the poster image's alt description. Movies can also be watched online.
Poster images for videos are also part of journalistic publications, where the poster also works as a still image and conveys meaning or illustrates a point of the story.
You're describing a real movie poster, which is a different thing from HTML5 poster frame:
> The image given by the poster attribute, the poster frame, is intended to be a representative frame of the video (typically one of the first non-blank frames) that gives the user an idea of what the video is like.
(http://www.whatwg.org/specs/web-apps/current-work/multipage/...)
There is a significant difference between the title and what's in the poster image. One commenter on the thread pointed to the Casablanca Wikipedia page.
The "poster" image for Casablanca has an alt text: "Black-and-white film screenshot with the title of the film in fancy font. Below it is the text "A Warner Bros. - First National Picture". In the background is a crowded nightclub filled with many people." If you imagine a movie being embedded in a webpage, that might be the alt text for the poster image whereas the title might be "Casablanca".
OTOH, it kind of goes to Hickson's point that describing frames of movies to the blind is not very useful when the frame is only there to give a hint of what the movie looks like. Do I really care that Casablanca contains a shot with "a man standing next to another man looking at two other men with people in the background" or some such?
Poster frames of videos on the web usually aren't the hollywood kind of posters. They're usually a random frame and/or logo of the site.
Also, time spent on describing poster could be spent on describing the video itself, which might give better indication whether video is worth watching.
But that has nothing to do with the 'poster' attribute on the 'video' element! Despite its name, this attribute is not intended for picture of a poster for the movie. It is intended for a preview frame to show in the player before the movie starts playing - the spec suggests one of the first non-blank frames of the video.
It's not like they suddenly want to disallow alt text on movie posters on the web.
Nobody wants to deny users anything. It is not really a question of pro or con accessibility despite the editorializing in the HN headline. Its a question of content versus presentation.
The first-frame preview is an artifact of the visual presentation of the video player and not relevant for non-sighted users, just as the colors used in css is not relevant for non-sighted users.
The actual content is the video, and if you want to make that accessible you should follow the recommendations in the spec and "provide alternative media streams and/or to embed accessibility aids (such as caption or subtitle tracks, audio description tracks, or sign-language overlays) into their media streams".
Watching poster frame is not a goal, it's only one of the ways to recognize whether video is worth watching.
There may be better ways for users with impaired vision to decide whether they want to watch a video.
For sighted users poster is useful, because visual recognition is faster than reading of descriptions.
However, when user has to read some description, then reading of description of one of the frames won't necessarily be faster or more informative than reading title or description of the video itself.
That's not necessary - we don't ourselves know the characters in the movie upon chancing on the page, and there's also readable text below explaining the actors (and who they're playing).
Flagged for misleading title. How about "Editor of HTML5 questions need for alt text in movie poster frames." This would 1) be less editorializing; 2) make this a submission of a technical issue as opposed to a submission questioning Ian's competence.
The submission title, "Editor of HTML5 (Ian Hickson) shows poor understanding of accessibility", is unnecessarily personalizing, editorializing, and polarizing.
It's likely Hickson's questions, and reluctance to accept a specific proposal without more data, are a matter of seeking understanding, and working through different ideas about what the proper significance and use of a 'poster' image for video might be.
(A better title might be: "Debating fine points of HTML5 accessibility: video 'poster' images and alt taxt".)
This type of editorializing is absolutely out of line. Ian is obviously trying to understand the other side, which is more than many developers ever try to do. As a low vision user of the web since its beginnings, I have encountered site after site designed for low vision and blind users based on advice and recommendations from people with good eyesight.
You can't understand what a person needs unless you ask them directly. That's what Mr. Hickson is doing here.
And lets not forget the more important matter - videos need transcripts so that the blind and deaf can follow along.
A few days ago I asked for feedback on how I might better explain accessibility to developers over at http://news.ycombinator.com/item?id=1716463 - issues like this are exactly the kind of stuff I was looking for. Feel free to leave comments there if you have them.
48 comments
[ 4.2 ms ] story [ 97.3 ms ] threadThat is pretty bad.
I think they need a group of people with disabilities to review all of the specs. There have to be many more issues like this lurking (that those with disabilities could probably already point me to).
Navigating their way there gives them context to what the video will be (in the use case that the video is on a blank page). If the page contains textual content then the user will get context from the text. If the video is completely unrelated to location context or to page-level context then the content will be of little to no interest, so providing an alt description adds no real-world value.
Also note that unless the video has audio captioning then the entire video is of little to no interest. Imagine watching something like Lost in Translation with no visual input and no descriptive captioning.
The 'poster' in this case is specifically a single image preview for a video shown before the video starts playing. According to the spec "typically one of the first non-blank frames".
It can be compared with e.g. a low resolution preview of a high-resolution image, as was used in the early days of the web. It is definitely not obvious that it would make sense to have an alternative text specifically for the lo-res preview as distinct from the alt text for the image itself.
Note that the video element already support an alternative text description - its just the single-frame preview which doesn't support a separate text.
I think Ians questioning of this feature is completely valid.
The HTML5 spec specifically addresses accessibility for video like this:
To make video content accessible to the blind, deaf, and those with other physical or cognitive disabilities, authors are expected to provide alternative media streams and/or to embed accessibility aids (such as caption or subtitle tracks, audio description tracks, or sign-language overlays) into their media streams. (http://www.whatwg.org/specs/web-apps/current-work/multipage/...)
This is much more sensible that providing an alt text for the first frame in the video.
I think the question is whether or not other attributes on the video tag can convey that information, making it redundant, or will the poster likely be used to convey different information.
The person who filed the bug is (effectively) saying "if the page author specifies a poster frame, it should be possible to add alt text. If the page author uses the default poster frame, it should be impossible to add alt text."
I believe the question Hixie was getting at was "why would you (a blind user) care about an explicitly-set poster frame but not a default poster frame?" which, I think, is reasonable.
For example:
Title: Cat jumping on bed
Frame: Image of cat jumping towards bed
Video: Cat jumps toward bed and lands on it.
The vast majority of the time, the poster frame isn't going to be different than the title of the video.
Now, I agree that when it is different, there should be a way to specify an alt text for it. But to require it when it's the same as the title seems silly to me.
Title: Dexter. Description: A likable Miami police forensics expert moonlights as a serial killer of criminals who he believes have escaped justice.
The exact episode also has those attributes:
Title: My Bad. Description: In the aftermath of last season's chilling finale, Rita's death has left Dexter feeling responsible, possibly even...guilty. Unable to deal with the trauma, Dexter makes a drastic decision that will impact everyone around him. Quinn stirs up trouble at the station when he notes that Rita's murder, which is being pinned on the Trinity Killer, doesn't fit his normal murder profile. Things get even more heated when Quinn suggests to LaGuerta that they look at the husband: Dexter Morgan. Meanwhile, Quinn offers Debra much needed support, causing her to see Quinn in a new light.
----
Simply put, titles and descriptions are not the same thing. It seems silly to me that somebody would advocate that they should be treated as the same thing just because there is overlap when the content is a 15 second video of a cat.
The issue is only about describing poster frame of the video (a placeholder picture that is displayed before user hits 'play').
Title: Dexter - My Bad.
Description: Dexter stands in a dank, dirty, dimly lit public bathroom, gripping a fishing anchor, a malevolent look in his eyes.
Isn't the problem there that one doesn't know it is the same. Requiring a description of the poster image, which can be blank as for alt-attributes of non-content images, means that it clearer when the author has simply chosen not to bother attempting to be accessible as opposed to the poster having no useful content beyond the title.
Ian's often repeated naive-sounding "Why would you do that? What's the use case?" has shot down many feature requests that seemed good in theory, but weren't that useful in practice.
Perhaps description of the video makes description of poster redundant? Maybe one can use ARIA? Maybe media framework can extract description from the video file?
Even if there is a way to describe the poster, will authors actually care to do it? Will they ignore it (naively assuming blind users don't watch videos) or fill it with boilerplate?
The plan seems to be that <track kind="descriptions" src="desc.en.srt"> be used as a child of <video> to provide descriptions in sync with the audio. http://www.whatwg.org/specs/web-apps/current-work/multipage/...
That is NOT questioning an assumption. He seems to imply that blind users have no use for an information that is interesting to other people.
Would you be interested in description of image representing video if you could get description of the actual video?
— I watched a funny video!
— Cool! What was on the poster image?
Poster images for videos are also part of journalistic publications, where the poster also works as a still image and conveys meaning or illustrates a point of the story.
> The image given by the poster attribute, the poster frame, is intended to be a representative frame of the video (typically one of the first non-blank frames) that gives the user an idea of what the video is like. (http://www.whatwg.org/specs/web-apps/current-work/multipage/...)
The "poster" image for Casablanca has an alt text: "Black-and-white film screenshot with the title of the film in fancy font. Below it is the text "A Warner Bros. - First National Picture". In the background is a crowded nightclub filled with many people." If you imagine a movie being embedded in a webpage, that might be the alt text for the poster image whereas the title might be "Casablanca".
http://en.wikipedia.org/wiki/Casablanca_%28film%29
I had to check this out. The alt text on the images really are excellent.
Also, time spent on describing poster could be spent on describing the video itself, which might give better indication whether video is worth watching.
It's not like they suddenly want to disallow alt text on movie posters on the web.
Why should a user with impaired vision be denied the same opportunity?
The first-frame preview is an artifact of the visual presentation of the video player and not relevant for non-sighted users, just as the colors used in css is not relevant for non-sighted users. The actual content is the video, and if you want to make that accessible you should follow the recommendations in the spec and "provide alternative media streams and/or to embed accessibility aids (such as caption or subtitle tracks, audio description tracks, or sign-language overlays) into their media streams".
There may be better ways for users with impaired vision to decide whether they want to watch a video.
For sighted users poster is useful, because visual recognition is faster than reading of descriptions.
However, when user has to read some description, then reading of description of one of the frames won't necessarily be faster or more informative than reading title or description of the video itself.
It's likely Hickson's questions, and reluctance to accept a specific proposal without more data, are a matter of seeking understanding, and working through different ideas about what the proper significance and use of a 'poster' image for video might be.
(A better title might be: "Debating fine points of HTML5 accessibility: video 'poster' images and alt taxt".)
You can't understand what a person needs unless you ask them directly. That's what Mr. Hickson is doing here.
And lets not forget the more important matter - videos need transcripts so that the blind and deaf can follow along.
A few days ago I asked for feedback on how I might better explain accessibility to developers over at http://news.ycombinator.com/item?id=1716463 - issues like this are exactly the kind of stuff I was looking for. Feel free to leave comments there if you have them.