Module naming and declarations

Brendan Eich brendan at mozilla.com
Tue May 21 00:52:27 PDT 2013


Kevin Smith wrote:
>
>
>     Please! There is no magic pixie dust whereby the Internet solves
>     the configuration problem for us.
>
>
> No pixie dust was involved - just vision.  If you would like to define 
> exactly what you mean by "configuration problem", I would be happy to 
> get specific.

Upthread, Sam's first real reply to Andreas:

3. As a way for two separately developed components to coordinate about which module they mean.


"Coordination problem" would be another phrase for the same thing, but 
the problem is solved in any proposal (including yours) by configuration 
of a registry.

> I'm not entirely sure where the knee-jerk reaction is coming from,

A short reply is not a knee-jerk reply. An appeal to unspecified, 
yet-to-be-created services on "the Internet" as the batteries in a 
batteries-included solution is an oxymoron. That's all.

> but this is what I'd like for anyone to take away from my previous post:
>
> - Any claims that the "logical names" design works out-of-the-box is 
> false advertising.

Where did you demonstrate this?

> - A URL-based system using today's tech can provide superior 
> out-of-the-box usability.

URLs are locations. They're versioned, explicitly so in general (e.g. 
http://ajax.googleapis.com/ajax/libs/jquery/2.0.0/jquery.min.js).

URLs are great when you need them but they're not superior out of the 
box for solving the "coordination problem".

/be


More information about the es-discuss mailing list