reserved web space for names URIs?

[Sanjiva Weerawarana <sa...@opensource.lk>]

Hi Guys,

Can we agree to reserve a Web space for NS URIs for Apache WS projects?
How about 
	http;//ws.apache.org/namespaces/*

So if Axis2 wants to create something we might ask for say
	http://ws.apache.org/namespaces/axis2/<whatever>

Likewise we should agree that all WS projects will (try to) put
namespaces under this pattern.

I'd like to push to put RDDL [1][2] documents at the end of the NS URI
but that's a secondary question. 

Thoughts?

Sanjiva.

[1] http://www.rddl.org
[2] http://xml.coverpages.org/rddl.html

Re: reserved web space for names URIs?

[Sanjiva Weerawarana <sa...@opensource.lk>]

Hi Anne,

> Hmmm ... 
> 
> A namespace URI is a name. Is "Sanjiva" evil? It seems pretty flat to
> me, unstructured, and unresolvable wthout global knowledge. But it
> seems to work pretty well. Pretty much everyone working in this space
> can recognize and resolve the name "Sanjiva". The reason I like URNs
> is that they aren't resolvable, and therefore it's clear that they
> are, in fact, names.

Ah I see your reasoning. OTOH you, with your built in semantic (Web ;-))
processor (aka your brain) know how to associate semantics/information
to names intelligently. The advantage of making namespace names
resolvable is to provide the really brain-dead, as dumb, browser-like
software to deliver some (potentially semantic) information about the
namespace name to the user. Without that ability, its just a flat opaque
string and that's it.

> But that's just my opinion. Who am I to argue with the likes of TimBL
> and Tim Bray?

:-) I didn't follow all the religious debate on this topic that happened
on the TAG mailing list but I know it was far from a decision that was
arrived at quickly or easily. I'm pretty should Tim Bray was for it
(after all he's the/a driver of RDDL) but I'm certain you're in
distinguished company.

Sanjiva.

Re: reserved web space for names URIs?

[Anne Thomas Manes <at...@gmail.com>]

Hmmm ... 

A namespace URI is a name. Is "Sanjiva" evil? It seems pretty flat to
me, unstructured, and unresolvable wthout global knowledge. But it
seems to work pretty well. Pretty much everyone working in this space
can recognize and resolve the name "Sanjiva". The reason I like URNs
is that they aren't resolvable, and therefore it's clear that they
are, in fact, names.

But that's just my opinion. Who am I to argue with the likes of TimBL
and Tim Bray?

Anne

On 5/4/05, Sanjiva Weerawarana <sa...@opensource.lk> wrote:
> Hi Anne!
> 
> > I suggest using the urn: scheme rather than the http: scheme for URIs.
> 
> Hmmm. The W3C Tag has "found" a while back that NS URIs be
> dereferencible and that they result in some "meaningful description of
> the namespace" - hence the need for something like RDDL. See
>         http://www.w3.org/DesignIssues/Architecture#Namespaces
> 
> Also, I've bought into the mantra that URNs are evil .. they're flat,
> unstructured and unresolvable without global knowledge (and only if
> people registered!). In fact many people even write them down wrong -
> the syntax is spsed to have urn:foo:bar but must people don't include
> the 2nd part. (I'm the one who used urn:foo for the first sample that we
> shipped with IBM SOAP (which became Apache SOAP) .. and I've seen many
> follow that incorrect example!)
> 
> Sanjiva.
> 
>

Re: reserved web space for names URIs?

[Sanjiva Weerawarana <sa...@opensource.lk>]

Hi Anne!

> I suggest using the urn: scheme rather than the http: scheme for URIs. 

Hmmm. The W3C Tag has "found" a while back that NS URIs be
dereferencible and that they result in some "meaningful description of
the namespace" - hence the need for something like RDDL. See
	http://www.w3.org/DesignIssues/Architecture#Namespaces

Also, I've bought into the mantra that URNs are evil .. they're flat,
unstructured and unresolvable without global knowledge (and only if
people registered!). In fact many people even write them down wrong -
the syntax is spsed to have urn:foo:bar but must people don't include
the 2nd part. (I'm the one who used urn:foo for the first sample that we
shipped with IBM SOAP (which became Apache SOAP) .. and I've seen many
follow that incorrect example!)

Sanjiva.

Re: reserved web space for names URIs?

[Anne Thomas Manes <at...@gmail.com>]

I suggest using the urn: scheme rather than the http: scheme for URIs. 

On 5/4/05, Sanjiva Weerawarana <sa...@opensource.lk> wrote:
> On Wed, 2005-05-04 at 14:02 +0200, Jochen Wiedmann wrote:
> > JaxMe does so since day one of incubation.
> 
> Oh you mean JaxMe is already using this pattern? Outstanding! So I take
> it that you guys haven't been putting anything at the end of the NS URI
> so far? I guess that's part of what I'm asking for .. ideally a RDDL
> doc.
> 
> When we start writing them for Axis2 we'll be happy to share them with
> you guys; its pretty simple anyway.
> 
> Sanjiva.
> 
>

Re: reserved web space for names URIs?

[Sanjiva Weerawarana <sa...@opensource.lk>]

On Wed, 2005-05-04 at 14:02 +0200, Jochen Wiedmann wrote:
> JaxMe does so since day one of incubation.

Oh you mean JaxMe is already using this pattern? Outstanding! So I take
it that you guys haven't been putting anything at the end of the NS URI
so far? I guess that's part of what I'm asking for .. ideally a RDDL
doc. 

When we start writing them for Axis2 we'll be happy to share them with
you guys; its pretty simple anyway.

Sanjiva.

Re: reserved web space for names URIs?

[Jochen Wiedmann <jo...@gmail.com>]

On 5/4/05, Sanjiva Weerawarana <sa...@opensource.lk> wrote:

> Can we agree to reserve a Web space for NS URIs for Apache WS projects?
> How about
>         http;//ws.apache.org/namespaces/*
> 
> So if Axis2 wants to create something we might ask for say
>         http://ws.apache.org/namespaces/axis2/<whatever>
> 
> Likewise we should agree that all WS projects will (try to) put
> namespaces under this pattern.

JaxMe does so since day one of incubation.


-- 
Outside of a dog, a book is man's best friend.
Inside of a dog, its too dark to read.
(Groucho Marx)