On 25/04/2009, António Mota <amsmota_at_gmail.com> wrote:
> What a interesting thread I miss :)
>
> Although the "operational" effects of the solution described are effective,
> and I don't doubt that a app build like that will work, I'll say however
> that it is *not* a restfull solution.
>
> Basically, you are confusing the concept of "connectedness" with the concept
> of "hipermedia as the engine of application state", or HATEOAS for short.
>
> Basically, hateoas doesen't mean that there should be "links" in the
> resource representations, means more that the server should drive the
> transitions between application states, and should do this in a way
> independent of the client, without the client having out-of-band information
> about it, namelly the "structure" of URI's.
>
> So, like Christmas, URI should be to any day a man wnats it to be. The
> server should be capable of changing it's all structure of URI's without the
> client knowing it. This is key for scabality.
Though one of the powerful features of the web is linking - letting
other services link to you, or letting your client use bookmarks.
Changing URIs is usually bad - even if a 100% pure HATEOS client would
not be affected - there are plenty of non-pure clients around, such as
the browser I'm using to type this :-)
>
> On another point, Rest is about resources, not data, so comparing it to CRUD
> is a dangerous and limiting thing to say...
>
> _______________________________________________
>
> Melhores cumprimentos / Beir beannacht / Best regards
>
> António Manuel dos Santos Mota
>
> _______________________________________________
>
>
>
>
> 2009/1/29 James Strachan <james.strachan_at_gmail.com>
>
>> 2009/1/29 Paul Sandoz <Paul.Sandoz_at_sun.com>:
>> > On Jan 28, 2009, at 3:37 PM, jstrachan wrote:
>> >
>> >> Or to say that another way - a URI is for life not for christmas;
>> >
>> >
>> > Ho ho ho, nice quote :-)
>>
>> :)
>>
>>
>> > I should stress that storing absolute URIs is a bad idea. Store relative
>> > URIs (be they keys or not) which are then resolved to absolute URIs that
>> > with the host, port, base path and may be a contextual path.
>>
>> Yeah; in my previous ramble I was only really considering relative
>> URIs which a view layer might morph into an absolute URI with
>> host/port/base path etc.
>>
>>
>> > I also point you to Bret's further work in this area:
>> >
>> >
>> http://markmail.org/search/?q=list%3Anet.java.dev.jersey.users+brett.dargan%40gmail.com#query:list%3Anet.java.dev.jersey.users%20brett.dargan%40gmail.com+page:1+mid:xjrp52mn7ywl4pdq+state:results
>> >
>> > Bret, sorry i have not had time to look at this in detail, but i really
>> like
>> > your solution.
>>
>> Yeah, it looks interesting.
>>
>> --
>> James
>> -------
>> http://macstrac.blogspot.com/
>>
>> Open Source Integration
>> http://fusesource.com/
>> - Show quoted text -
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_jersey.dev.java.net
>> For additional commands, e-mail: users-help_at_jersey.dev.java.net
>>
>>
>
--
James
-------
http://macstrac.blogspot.com/
Open Source Integration
http://fusesource.com/