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

Re: Using DOM to replace media attribute in the link tag on page load

From: Patrick Garies <pgaries@fastmail.us>
Date: Thu, 24 Jul 2008 01:31:40 -0500
Message-ID: <488821CC.3020202@fastmail.us>
To: David Perrell <davidp@hpaa.com>
CC: www-dom@w3.org

David Perrell wrote:
>  IE > 6 won't run on pre-XP Windows and there are clearly many who
>  feel no need to upgrade their OpSys or change its default UA.

The primary point of the CC‐based approach is to work around this issue.

David Perrell wrote:
>  Meanwhile, there are millions of web pages with script typed as
>  "text/javascript". UA acceptance of "text/javascript" won't be going
>  away any time soon.

I didn’t dispute this. I simply noted that the mentioned MIME type is 
deprecated.

David Perrell wrote:
>  IMHO, conditional mime types are a bit premature.

I’m curious as to what criteria should be used to determine when the 
time is “mature” in your view. The older MIME types text/javascript and 
text/ecmascript are essentially deprecated per RFC4329 (or “obsolete” as 
that RFC, unfortunately, puts it); RFC4329 is already on the IANA MIME 
Media Types list. The question is: “Can you start using the new MIME 
types here and now without significantly hurting anything?” The answer 
may vary depending upon your goals; in my case and, probably, most 
people’s cases (where they have enough control over their code to use 
CCs), the answer is “Yes.”.

David Perrell wrote:
>  If you're presenting plain HTML then it should be safe to use the
>  <script> tag without a type declaration.

I agree. However, it is “safe” to do quite a number of things that are 
not technically correct per specification(s), not conforming, invalid, 
or poor practice.

In this case, omission of the |type| attribute makes the document 
invalid (and non‐conforming?) and keeping it doesn’t hurt anything. I 
prefer to write documents that are valid and conforming since I believe 
that doing so is good form and, naturally, suggest that others do the 
same (as I did here).

Anyway, I don’t see how this even matters unless you’re suggesting that 
nimblehost should drop the attribute entirely.

David Perrell wrote:
>  However obsolete, "text/javascript" is currently the most-supported
>  content type for javascript and was probably a wise choice.

I’ve recognized the support levels for the various MIME types and my 
CC‐based approach takes this into account. (*/javascript can be 
substituted for */ecmascript in the example code that I previously 
presented, if preferred.) If you need to support browsers that don’t 
support the new application/* MIME types and/or the CC‐based solution 
doesn’t work for you (e.g., the problematic browsers are not WIE), then, 
by all means, use the text/javascript MIME type.

As a side note, I disagree with the term “obsolete” which means 
“useless” (clearly, text/javascript has not entered the realm of being 
useless); I prefer use of the term “deprecated” (“disfavored”).

David Perrell wrote:
>  I read an IE8 blog complaint: IE8b1 doesn't support the new mime
>  types. Perhaps that will change before final release.

That’s both interesting and unfortunate.

— Patrick Garies
Received on Thursday, 24 July 2008 06:32:22 GMT

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