cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Turner <j...@socialchange.net.au>
Subject Re: Standardizing XSP Namespaces
Date Fri, 09 Feb 2001 04:27:47 GMT
On Fri,  9 Feb 2001 14:33, Donald Ball wrote:
> On Fri, 9 Feb 2001, Jeff Turner wrote:
> > While everyone's talking about taglib namespaces, is there any consensus
> > on what non-"official" taglibs should use?
> >
> > Possibilities:
> >
> > 1) http://<somecompany>/xsp/<taglib>/<version>
> > 2) http://apache.org/xsp/<taglib>/<version>
> >
> > The second is effectively Apache donating a bit of "name space" to taglib
> > authors.
> >
> > I'd prefer 2), because taglibs are the products of individuals, not
> > companies. What if I'm not affiliated with a company, or I change from
> > <somecompany> to <anothercompany>? No doubt, one could use URNs or
> > something, but that's my other reason: "rememberability" ;) It's much
> > easier to remember taglib namespaces if they all start with
> > "http://apache.org/xsp/" and all follow a regular pattern.
>
> sorry, but no, i don't think i could stand for that. we can't open up the
> apache.org namespace to all and sundry who happen to write an xsp
> logicsheet. 

Name space is cheap :) It's very unlikely that there'd ever be case where a 
namespace conflict can't be resolved amicably. In return for this risk, one 
gets the benefit of uniform naming across all taglibs, which taglib users can 
benefit from every day.

> by all means, i welcome development of xsp logicsheets by all
> and sundry, but under their own namespaces. if we all vote to include a
> 3rd party logicsheet in the core xsp namespace, then they can (and should)
> be in the apache.org namespace, but not otherwise.

Another reason to "lease" http://apache.org/xsp/ name space :) You're saying 
that in the normal evolution of a taglib, *at least* one namespace change 
must take place. That seems a high price for users to pay for conceptual 
purity.

> > "RDDL offers an answer to "what might a namespace URI mean?" by
> > providing: a text description of some class of resources and of other
> > resources related to that class. It also contains a directory of links to
> > these related resources. An example of a class of resources is that
> > defined by an XML Namespace. Examples of such related resources include
> > schemas, stylesheets, and executable code. A Resource Directory
> > Description is designed to be suitable for service as the body of a
> > resource returned by deferencing a URI serving as an XML Namespace name."
> >
> >  -- http://xmlhack.com/read.php?item=994
>
> that's a good idea. if it pans out as described here, then it's
> practically mandatory that we choose (1) above - if only so that
> logicsheet authors can create and maintain its RDDL documentation.

Ah bugger.. caught in my own web ;)

OTOH, how many taglib authors can convince their company to let them create 
something at http://somecompany/xsp/ in the first place? If Apache did serve 
RDDL docs from http://apache.org/xsp/, there could be a submission process 
for taglib authors. This would a) give company-unaffiliated taglib authors a 
place to put their RDDL doc, b) be much easier for company-affiliated taglib 
authors to use than navigating a minefield of company politics.

Pragmatism vs. purity.

--Jeff
(who has a LDAP taglib poaching apache name space right now ;) and will 
change when this gets resolved)


> - donald

Mime
View raw message