W3C home > Mailing lists > Public > www-dom@w3.org > July to September 2008

RE: RFC4288 script media types (was: Using DOM to replace media attribute...)

From: David Perrell <davidp@hpaa.com>
Date: Tue, 29 Jul 2008 14:31:03 -0700
To: "Patrick Garies" <pgaries@fastmail.us>
Cc: <www-dom@w3.org>
Message-ID: <HIEDLECHAFDEPLGKECDDGELDAFAB.davidp@hpaa.com>

Patrick Garies:
| I also know of no cases where this is strictly required for “functional 
| compliance” (whatever that is). I simply said that the MIME type is 
| deprecated (i.e., discouraged).

I don't agree the media types should be referred to as "deprecated." The meaning of "obsolete" in this context is in RFC4288 (which obsoleted 2048): '...media types that are no longer believed appropriate for use can be declared OBSOLETE by a change to their "intended use" field'. Obsolete can mean "outmoded," which is reasonable. RFC4329 is not an 'Internet standard', and - unlike HTML - doesn't need an interim period of deprecation (when UAs SHOULD continue to support something) prior to obsolescence.

'Functional compliance': where functionality depends on compliance (e.g. at some point security or versioning benefits might depend on full compliance). You're striving for syntactical compliance; there currently appears to be no functional difference.

| There are functional differences for application/ecmascript assuming 
| that the script is in its own file, at least (e.g., the version 
| parameter). I haven’t tested whether current browsers that support the 
| new application/* types actually adhere to the RFC  in this regard though.

Current FF and Opera versions process version info for ecmascript as if it were javascript, i.e. FF accepts 1-1.8, Opera 1-1.5, and type="application/ecmascript;version=1.5" is equivalent to type="text/javascript;version=1.5". Unsupported versions cause script to be ignored, enabling NOSCRIPT in Opera but not in FF. 

| Again, the MIME type is discouraged, not banned. Use of text/javascript 
| affects compliance; i.e., you are “conditionally compliant” instead of 
| “unconditionally compliant”. What value you place on that is, of course, 
| up to you.

Yes, by stating that application/javascript and application/ecmascript SHOULD be used instead, RFC4329 makes use of text/javascript 'conditionally compliant' with the RFC - the 'condition' IMO being that all UAs to be supported support those types. 

We agree that it's a good thing to standardize javascript & ecmascript media types and bring them into compliance with the recommendations regarding text/- and application/- types as defined in RFC2046 (surely if PostScript is an application type, JavaScript should be also). We hope that Microsoft also agrees and makes IE 8 fully compliant.

We disagree whether or not it's currently advisable to use conditional comments for no other reason than to present a registered-but-not-fully-supported media type to all UAs except MSIE. No harm there. Thanks for the discussion and CC examples. 

David Perrell
Received on Tuesday, 29 July 2008 21:32:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 June 2012 06:13:59 GMT