Re: [hybi] WebSocket and metadata (Was: BWTP Proposal - Bidirection Web Transfer Protocol)

[Greg Wilkins <gregw at webtide.com>]

Ian Hickson wrote:
> On Mon, 29 Jun 2009, Greg Wilkins wrote:
>> Ian Hickson wrote:
>>
>>> Minimal framing isn't a requirement. Being able to implement the 
>>> server-side component easily is:
>>>
>>> | - It must be possible to implement a fully-conforming server-side 
>>> |   component for this in a few dozen lines of scripting code, in the case 
>>> |   where sharing the connection with an HTTP server isn't required.
>> Can you expand on why you think this is a requirement?
> 
> If we make it difficult to write a compliant implementation, then we won't 
> get many compliant implementations. This will either be because we don't 
> get many implementations at all, or because we get many poor quality 
> implementations.
> 
> Having few implementations is bad because it leads to a monoculture where 
> a single vulnerability means the entire ecosystem is vulnerable.
> 
> Having many poor quality implementations is bad because interoperability 
> will be difficult to achieve, making it less likely that users will have a 
> good experience.



Ian,

I think lines of code and software quality are orthogonal.

There are many many programmers who are able to produce
very poor quality code in just 1 line.  Even with simple CGI,
I have had many experiences where poor quality client "implementations"
have had unintended consequences.

I believe the path to a good user experience is to give them an
environment that abstracts away the details of the protocol.
Websocket does this on the client side, so I don't see why this
should not be done server side.

But I do agree with you that the protocol should not be difficult to
implement and that we do not want a monoculture.   But it is overly
simplistic to say a protocol that needs more than a few dozen
lines of code is going to be rarely implemented or of poor
quality.


cheers