This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
I have a comment/concern about the Draft Community Report 8 (April 2015) on webVTT, where it says: "Authors are encouraged to write cues all on one line except when a line break is definitely necessary, and to not manually line-wrap for aesthetic reasons alone." Perhaps I'm misunderstanding the context of this suggestion (and in what situations it is meant to apply), but I think the term "aesthetic" is being employed in a somewhat prejudicial sense here, albeit unwittingly so. Often the "aesthetic" line breaks in caption text are the ones that make captions marginally more readable by the deaf, especially in fast-paced programming, bearing in mind that the deaf often have to rely solely on the text for meaning unlike the hearing who pick up significant meaning clues from the audio while they themselves are reading captions. That's why there has long been a general captioning prohibition (at least according to best practices) against separating nouns from the direct and indirect articles that precede them, which meant that one would ideally not end a caption line with "the" or "an" -- and a still more emphatic rule against separating phrases between lines: so that, for instance, one would avoid separating the phrase "you know" via line break into "you" and then "know." These may be only marginal reading aids in themselves, but collectively they boost the overall caption readability for the deaf, especially in a fast-paced, talkative program. Of course, if the browsers are going to have the final unpredictable say on where line breaks occur, then the captioner might be wasting their time to attend to that level of detail, but this in turn makes me wonder why HTML5 captioners are creating such lengthy captions in the first place: again, the best practices in captioning would suggest limiting captions to two lines of text. Even if the actual number of lines that show in playback may increase at times based on the browser used (and how it might have been resized), I think that the general target in captioning should still be to provide two-line captions, and certainly not 5-or-more-line captioning that one sometimes sees these days in HTML5. Sorry if I sound "picky" about this, but I tend to be a little suspicious whenever I hear the word "aesthetics" used in reference to captioning practices, because in my experience(25 years in the closed-captioning business), it's typically employed as a pejorative term by caption companies whenever they want to cut corners on time-consuming practices. Suddenly the old practices are denounced as "aesthetic," whereas they were originally justified based on actual studies that showed that such practices conduced, so to speak, to improved overall caption readability for the deaf. Thus the companies turn the vice of cutting corners into the supposed virtue of not being sticklers. This pleases the company and its clients (who get lower rates thanks to the shortcut) and the change is often welcome by some of the more lazy new-hires, but the deaf community takes the subtle but real "hit" of receiving somewhat less readable captions overall.
Thanks for the review, Brian. I think we could improve the wording and also point out the use of for such so situations. Could even provide an example.
This is excellent feedback, thank you. It's obvious that captions and subtitles are not simply line balanced sometimes, and it's interesting to learn some of the best practices behind this. A "problem" with WebVTT is that you can't assume anything about the font size, so this advice is written to avoid having manually wrapped two-line captions wrap into two long and two short lines instead of three medium-length lines when the font size is too big. Of course, the reality is that you can author captions with manual line wrapping where each line is so short that the font size would have to be preposterously large before lines would wrap. In those cases, picking where the lines wrap could well improve readability. Not sure how to rephrase this, but it could be more clear. Adding anywhere it would be a bad idea to break a line is of course a clean solution, but I wonder if anyone would follow that advice, sounds like more work...
Given the variables associated with captioning for the Web (potentially resized video screens, various font sizes, and so forth) this "line break" concern of mine might need to be addressed more in caption-creation software and coding rather than in webVTT standards. Even then, I suppose the software could only flag cases where problematic line breaks (from the standpoint of viewer readability) might occur, rather than offering any decisive red flags for the captioner. That's not ideal but it looks like an unavoidable shortcoming for the foreseeable future, given the number of variables involved in caption display in HTML5. I guess the only remaining suggestion from me then would be to remove the word "aesthetic", which has a slightly pejorative connotation to me, given the fact that most manual line-wraps that I can think of would be motivated by (or at least could be justified as) an attempt to improve readability. Perhaps we could merely omit the final phrase that contains the word "aesthetic": Original: "Authors are encouraged to write cues all on one line except when a line break is definitely necessary, and to not manually line-wrap for aesthetic reasons alone." Revised: "Authors are encouraged to write cues all on one line except when a line break is definitely necessary." I think this latter wording still ticks all of the substantive boxes when read in the context of the short paragraph from which it was drawn.
That works for me. The final phrase is accurate if taken literally, but it doesn't really add anything, so: https://github.com/w3c/webvtt/pull/190
Brian, I thanked you as "Brian Quass", let me know if you'd prefer something else.
Pull request has landed. Should we add something about using for non-breaking space character use in automatically wrapped captions or leave that out for authoring guides to describe?
I think it's somewhat unlikely that people would take such advice seriously and do that over manual line wrapping, but if you have something in mind I will obviously review it :)
Brian, what do you think?