The same is going to be true of STYLE, and is true of all cue and region settings.

Why is it less accesible? That it is an English word doesn't mean they need to know English or know the word to use it and know what it does in WebVTT. Also, as Philip said, this is all over the place; other than WebVTT we also have this "issue" in  HTML, SVG, CSS, JavaScript, HTTP...

The VTT language itself is not suppoed to be international or internationalized; suggest 'wontfix'

WONTFIX it is

I disagree vehemently with the statement that the "VTT language is not supposed to be international or internationalized" because I firmly believe you intend for the language to be used globally and for the language to support the needs of users world-wide (that is, it is internationalized). 

But I don't believe this is what you mean by your comment in response. Despite what I'm going to say below, I'm not going to change the status of this bug and I'm going to close the I18N issue in tracker. You choice of adopting an English keyword is a defensible design decision and does not impede the above internationalization goals. However, I do want to make the design point I was making clearer.

Many programming and markup languages use English/ASCII keywords (what Charmod-Norm calls a "vocabulary") and that is a reasonable design choice. But this doesn't necessarily mean that every keyword needs to be an English word--or even a word. Comments, in particular, tend not to be introduced with words and use symbolic markup instead.

In this case, the note keyword appears to participate in the message when you are reading the file--in English:

   NOTE I meant to do this.

And seems less natural when it looks like:

   NOTE Je voulais le faire.

Since there exist a large number of non-word tokens that are well-understood in computing or markup contexts to be introducing a comment, I tend to favor adopting one of them instead.

   // Je voulais le faire.
   # Je voulais le faire.
   <!-- Je voulais le faire -->
... etc.

I see your point, but the existing syntaxes are problematic because they carry assumptions that would not hold in WebVTT:

(In reply to Addison Phillips from comment #5)
>    // Je voulais le faire.
>    # Je voulais le faire.

These are typically line comments, but WebVTT comments are blocks that are terminated at a blank line or at a line containing "-->".

>    <!-- Je voulais le faire -->

This and /* */ are multiline comments where you can comment out an arbitrary number of lines, but this is not possible in WebVTT either.

A non-word syntax I think would work would be to prefix each line with "-->". That makes each line be a block, and it removes the restriction that comments can't contain "-->". It also makes it more obvious how to comment out cues and other blocks.

    WEBVTT

    --> I meant to do this.
    --> Je voulais le faire.

    --> I can use -->!

    --> STYLE
    --> ::cue { ... }

    --> FIXME The commented out cue below is wrong

    --> 00:00:00.000 --> 00:00.10.000
    --> Hello wrold
    This text is still part of the comment block,
    but it's a syntax violation because it's not
    clear where the comment ends, and if it uses
    a --> somewhere, that would start a new block.


If you (or anyone else) think this would be much awesomer than NOTE, I suppose we can reopen.