Re: [hybi] Extensibility mechanisms?

[Maciej Stachowiak <[e-mail]>]

On Apr 18, 2010, at 12:20 AM, Greg Wilkins wrote:

>
> If you want you concern of enabling amateur programmers
> to implement the protocol included, then please
> propose it as a requirement and see if others
> agree.
>
> Similarly, I'm concerned about websockets transiting
> intermediaries and I too will be proposing requirements
> to that effect, and they too will need community
> support.

I agree that everyone should advocate for what they feel to be  
requirements. That seems like a sensible way to drive the requirements  
document.

But in fairness, in the message Ian was responding to you said:

On Sat, 17 Apr 2010, Greg Wilkins wrote:
>
> I'm sorry but despite your original aims, enabling amateur programmers
> to write protocols that avoid the need for vendor supplied
> infrastructure is not the prime reason this WG has come into  
> existence.

...which seemed dismissive of Ian's goals.


We should try to be respectful of each other's proposed requirements,  
even when we disagree. The IETF operates by rough consensus, so it's  
totally acceptable for different people to be participating in this WG  
with different goals.


My own view: I think we need to support both lightweight "amateur"  
deployments and large-scale deployments involving complex  
infrastructure. Another way to put it is: the protocol should be  
compatible with advanced back-end infrastructure, but should not  
*require* complex infrastructure. Yet another way to put it is: the  
protocol should scale from very small experimental deployments, to  
very large deployments with worldwide reach.

Why do I take this view? A large part of the Web's success has come  
from the "long tail" of content that exists only because the  
underlying formats and protocols make it easy to do simple things.  
This has led to the explosive growth of the Web. At the same time,  
it's clearly important to consider the needs of larger organizations  
as well. A solution that works for Joe Hobbyist Programmer at home,  
but doesn't meet the needs of, say, Facebook or Yahoo, is not a  
complete solution. Large sites are clearly a critical part of the  
Web's success as well.


I think ultimately these two goals are not fundamentally at odds. It  
may be more challenging to satisfy both sets of needs, but we have a  
lot of smart people here so we should be able to figure it out.


Regards,
Maciej