- From: Jacob Rossi <Jacob.Rossi@microsoft.com>
- Date: Wed, 25 May 2011 17:07:24 +0000
- To: "www-dom@w3.org" <www-dom@w3.org>
- Message-ID: <D0BC8E77E79D9846B61A2432D1BA4EAE0332A0AB@TK5EX14MBXC286.redmond.corp.microsoft.>
I agree with Glenn. I think adding a convention for prefixing custom key names is appropriate. It could be as simple as a note that vendors should devise a prefix. For example, if Microsoft were to add a custom key, then we might consider prefixing it with "ms". What you don't want is for two vendors to devise colliding non-standardized key names for different keys. Then web apps may potentially respond to the wrong type of input. I think any such note, however, would be non-normative. So we could do this during Last Call. -Jacob From: Glenn Adams <glenn@skynav.com<mailto:glenn@skynav.com?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>> Date: Wed, 25 May 2011 10:34:51 -0600 Message-ID: <BANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg@mail.gmail.com> To: Doug Schepers <schepers@w3.org<mailto:schepers@w3.org?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>> Cc: Giuseppe Pascale <giuseppep@opera.com<mailto:giuseppep@opera.com?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>>, W3C Web and TV <public-web-and-tv@w3.org<mailto:public-web-and-tv@w3.org?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>>, "www-dom@w3.org<mailto:www-dom@w3.org?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>" <www-dom@w3.org<mailto:www-dom@w3.org?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>> It is simply a matter of best practices to define normative conventions for avoiding collisions when arbitrary parties can create names. That is why we have .com, .org, etc in domain names; that is why Unicode has a private use area and doesn't allow users to define character codes outside such area. Java has a convention for ensuring non-collision of package names. Furthermore, you don't want a third party creating a name and then in the future you want to revise the standard to include a standard definition that would use or reuse such a private name. If you don't want to define a convention to permit avoiding collisions between 3rd parties, at least define a convention that all private uses must employ in order to distinguish from standard use. In CSS, property names that start with '-' (HYPHEN) are considered private uses; in MIME types, sub-types that start with "x-" are considered private use. Now is the time to define a convention, and not later when private uses impose constraints on new standard uses. Regards, Glenn On Wed, May 25, 2011 at 10:11 AM, Doug Schepers <schepers@w3.org<mailto:schepers@w3.org?Subject=Re%3A%20Standardize%20Buttons%20For%20TV-Web%20Remotes%3F&In-Reply-To=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E&References=%253CBANLkTinzp5bQUEdTfJffnYgdUrVyTvx8wg%40mail.gmail.com%253E>> wrote: > Hi, Glenn- > > Glenn Adams wrote (on 5/25/11 12:20 AM): > > Thanks, this looks good. My only remaining concern is in regards to >> implementation defined key values. The definition: >> >> [[ >> If there is no appropriate key value in the key values set, then a key >> value must be devised. The key value should be as human friendly as >> possible and must not contain whitespace characters. The value must be >> composed only of characters in the ranges \u0030..\u0039, >> \u0041..\u005A, or \u0061..\u007A, and must begin with a character in >> the range \u0041..\u005A. >> ]] >> >> is not adequate to prevent name collisions between custom definitions. >> Some normative convention is needed to prevent collisions. >> > > What conflicts do you anticipate? > > Even if there are keys on different devices that have the same values, > webapps can still interpret the pressing of those keys in whatever way works > best for that application; this is the intent of the key value feature. > > The main goal of listing key values explicitly is to promote convergence, > not to differentiate between implementations. I am very reluctant to add a > convention that promotes the proliferation, rather than reduction and > coordination, of new key values. While such things are useful for certain > types of extensions (new events, for example), I don't think they will be > useful in this case, because they aren't associated with functionality > exposed to webapps... they are only the indication that a particular button > or key was pressed, and don't trigger any specific webapp feature (other > than what the webapp itself "assigns" to that key value, via script). > > In any case, I believe that this is a comment that could be addressed in > Last Call, because it does not introduce any normative changes to > implementations. Unless you see this as an issue that blocks us moving to > Last Call, I believe the most prudent course of action is for us to do so. > > > Regards- > -Doug Schepers > W3C Team Contact, SVG, WebApps, and Web Events WGs >
Received on Wednesday, 25 May 2011 17:08:02 UTC