Re: [apps-discuss] draft-ietf-weirds-bootstrap-00 and our lawn -- feedback?

John,

Your client is a three line shell script because someone took the time
to write programs like wget that implement a standard protocol.  It
would surprise me if someone hasn't already written a URI template
processing utility that you can call from your script.  Your 3 lines
would likely end up being 4 lines. Pseudocode would look something
like,

get homedoc from homeUri
parse desired service uri template from homedoc
resolve URI template with parameters
get informationdoc from resolved URI

However, we seem to be addressing this whole issue from the wrong
perspective.  It is quite probable that the people who are
implementing the RDAP servers will have no issue handling the URIs.
It is quite likely that the clients will be able to successfully
create the necessary URIs.  It is possible that you will not need to
change the RDAP URI space once the spec has been baked.

However, you are highly visible, publicly accessible service that has
chosen to use HTTP.  IETF best practices recommend that you don't do
what you are doing.  Will it work for you, sure, probably, because of
your context.  The problem is that other developers developing HTTP
based services are going to look to what you have done as an example
and say, look, at how RDAP works.  If it works for them, then why
won't it work for us.  But their context may not be the same as yours.
 It is in the web's best interest if high profile services like RDAP
are a model for others to follow.

The question should not be, why should RDAP follow best practices when
they are not absolutely necessary? The question should be, what
justification exists for not following that guidance. Justifications
might include,
- there are no URITemplate processing libraries available on the
target client platforms.
- caching infrastructure does not exist on client platforms to
minimize the impact of the extra round-trip to get the home doc.

I don't know if these are valid objections, but I think that would be
a better discussion to have than the current one.

Regards,

Darrel

On Sat, Feb 15, 2014 at 10:55 PM, John Levine <johnl@taugh.com> wrote:
>>One of the major advantages of using URI templates, beyond freeing the
>>server from conforming to conventions, is that it makes client code
>>really simple.
>
> People keep saying this.  My client is a three line shell script that
> uses wget and grep (really.)  Could you explain how that works with
> templates?
>
> While you're at it, it'd also be helpful to identify some web servers
> that people might plausibly use to implement RDAP in 2014 (as opposed
> to an exercise in retrocomputing) that couldn't handle the syntax in
> rdap-query.  You can skip the .wellknown stuff, which is a holdover
> from a point when we thought we might use something like SRV records
> that don't provide a path in the bootstrap.
>
> R's,
> John
>
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss