Re: UA support for Content-Disposition header (filename parameter)

Maciej Stachowiak wrote:
> 
> Julian,
> 
> On Mar 17, 2008, at 2:33 PM, Julian Reschke wrote:
> 
>> Maciej Stachowiak wrote:
>>> I think we all assumed good faith on Julian's part until it became 
>>> clear he was trying to game the system to prove a point. I'm glad he 
>>> finally
>>
>> I was not.
>>
>> Just because I added an ironic remark about the process doesn't mean 
>> that I did not raise the technical issue with good reasons.
> 
> My impression was that you made your proposal not because you actually 
> wanted it accepted, but because you suspected it would be opposed, in 

I made the proposal because we were discussing I18N on the HTTP WG, and 
I recalled this IE issue I encountered several years ago. As I don't 
think that httpbis can help here, I thought that *this* actually is 
something for the HTML working group, because it mainly affects HTML UAs.

So no, my hope was that this WG actually would be interested in solving 
this problem.

> order to demonstrate a point about other features of HTML5 (such as 
> content sniffing) that touch on HTTP, i.e. that similar arguments apply 
> and they should also be rejected. I was not the only one to get that 
> impression. Since it sounded like that was your intent, I really didn't 
> want to discuss the feature on the merits any further. If that was not 
> your intent then I apologize; but that certainly wasn't clear from your 
> emails.

Apology accepted.

> Now, from the combination of your statements, I'm still unsure what you 
> intend:
> 
> 1) Do you sincerely think HTML5 should require particular handling of 
> the Content-Disposition header, despite in all other cases arguing 
> strenuously against specs "profiling the underlying protocol"? If so, 
> then that deserves to be discussed on the merits, and I apologize for 
> cutting you off. (I happen to think that on balance it's not worth 
> adding to HTML5, but that should be discussed on the list.)

All UAs already implement C-D in some way, they just differ in whether 
they also implement I18N as per RFC2231.

As far as I can tell, FF and Opera do, Webkit may in the future (?), IE 
doesn't. What's worse, there isn't a single format IE will grok that 
works across different client locals (such as Western Europe and Asia).

What I'm interested in is making this (or something better, should it 
exist) interoperable.

So far MS has ignored RFC2231 (and yes, bugs were raised through paid 
support channels by *very* big customers). Thus I think that another RFC 
saying "you really should do this" is unlikely to have any effect.

> 2) Did you raise this proposal to show that the arguments against it 
> also argue for removing other features that are already in HTML5? In 
> that case, please make your point about those features directly instead.
> 
> I assumed these were mutually exclusive options, but I suppose you could 
> have both aims. If so, then mixing them up makes things even more 
> confusing to readers, since they cannot even tell if your proposal is 
> made in good faith.

The proposal was made in good faith, recall I was only making an ironic 
comment once you claimed that this isn't something the WG should care about.

BR, Julian

Received on Monday, 17 March 2008 22:31:14 UTC