[uf-new] hAudio Issue D4: 2008-01-10 rel-enclosure does not allowforlinks to streaming files

[Martin McEvoy]

Hello

Tantek Celik wrote:
> This is an excellent follow-up Toby - perhaps you can capture your implementation experience on the wiki on a page like "streaming-implementations".
>
> Toby wrote:
>   
>> However, the percentage of links that actually *include* a 'type'  
>>     
> attribute is woefully small. Performing HTTP HEAD requests for each  
> file would be too slow.
>
> While I'm not claiming that you are using this as an argument for a new format/value, I will point out that similar forms of argument have been made in the past (and in other forums) for creating new things.
>
> The brief point to keep in mind:
> If existing content publishers are not making use of an *existing* format feature as they should, why would anyone think that such publishers would make any more use of a *new* format feature?
>
> And in a proactive way: rather than inventing a new feature for which there is already an existing feature (which may or may not be widely adopted), provide better documentation and how-tos (perhaps with benefits) for how to use that existing feature - as you have recommended in your email.
>
> Your solution is good. In addition, we may wish to consider making use of the "type" attribute on a rel="enclosure" links in hAudio a SHOULD to communicate the importance of doing so.
>   

Ahh I wish I had read this email before I sent My last one...


+1 to

"Your solution is good. In addition, we may wish to consider making use 
of the "type" attribute on a rel="enclosure" links in hAudio a SHOULD to 
communicate the importance of doing so."


Thanks

Martin McEvoy
> Tantek
>
>
> -----Original Message-----
> From: Toby A Inkster <mail at tobyinkster.co.uk>
>
> Date: Wed, 20 Aug 2008 16:18:41 
> To: <microformats-new at microformats.org>
> Subject: [uf-new] hAudio Issue D4: 2008-01-10 rel-enclosure does not	allow
> 	forlinks to streaming files
>
>
> Tantek wrote:
>   
>> Toby wrote:
>>     
>
>   
>> differentiating between download and streaming URLs
>>     
>>> in the case where both are available.
>>>       
>> That distinction is already made via different MIME types on the  
>> enclosure, reflected in the type attribute (on a or link tags). No  
>> need to invent something new.
>>     
>
> I've written an hAudio -> M3U converter (M3U is a common playlist  
> format supported by several media players including Winamp, XMMS and  
> iTunes) and do indeed use the 'type' attribute to differentiate  
> between playable downloads (e.g. MP3s, OGGs, etc) and downloads which  
> are not directly playable (e.g. BitTorrent files, ZIP files, etc).  
> This distinction needs to be made in order to avoid placing non- 
> playable files into the playlist.
>
> However, the percentage of links that actually *include* a 'type'  
> attribute is woefully small. Performing HTTP HEAD requests for each  
> file would be too slow.
>
> One solution of course might be for hAudio to strongly encourage the  
> 'type' attribute to be set on any rel=enclosure links. It would help  
> if all the examples of the Wiki included the attribute, and a list of  
> commonly used MIME types could be provided to help newcomers. (Some  
> are not immediately obvious - e.g. application/ogg instead of audio/ 
> ogg.)
>
>