On 10/5/05, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
> You offered up Rose::URI for Dave to use independently of the
> Rose module suite.
>
> I said that if Rose::URI is *meant* to serve such a purpose, then
> its name is bad.
But what if it "can be" used to serve such a purpose? That fact that p
* John Siracusa <[EMAIL PROTECTED]> [2005-10-05 21:10]:
> Well, Rose::URI isn't a factory for URI objects (neither "isa"
> URI nor ref $uri eq 'URI') so there'd have to be a different
> name, at the very least.
That part was missing from the discussion so far. (Or at least I
had missed it.)
(And
* John Siracusa <[EMAIL PROTECTED]> [2005-10-05 18:15]:
> It's better precisely because it'd be so hard to come up with
> meaningful ::FooPurpose name components for the individual
> modules. In reality, the ::FooPurpose is at least partially
> "to be part of the Rose module suite", which Rose:: re
Maybe leave the name alone but change the abstract?
"A URI object built for easy and efficient manipulation" may be
true but it doesn't tell us: manipulation of what, by what, nor
that it is compatible with the URI in many ways. Perhaps
"standalone module for constructing URIs by building them
On 10/5/05, John Siracusa <[EMAIL PROTECTED]> wrote:
>
> As for names like URI::Rose, I don't like to see branded namespace pieces
> wandering off and polluting existing hierarchies.
But that's just the point: there is nothing branded or rose-ish about
a generic URI factory.
> I also think name
On 10/5/05, Dave Rolsky <[EMAIL PROTECTED]> wrote:
> On Wed, 5 Oct 2005, John Siracusa wrote:
>> If Dave searched for "URI" or "URI from hash" he'd have seen Rose::URI on
>> the first results page:
>
> I did do such a search, but I ignore Rose::URI cause of the name ;) [...]
> "Rose::URI" says to
On 10/5/05, John Siracusa <[EMAIL PROTECTED]> wrote:
>
> If I had to come up with separate, non-branded names for everything under
> Rose::* they'd be all over the map. Since they are all interdependent in
> one way or another (via inheritance and also intended usage scenarios) I
> think it makes
On 10/5/05, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
> Yes the combined ::Object + ::DateTime package would have no
> dependencies on any other Rose::* modules, so could be useful
> outside those.
I think it'd be pretty odd to combine DateTime utility class with an object
base class.
> If you wa
* John Siracusa <[EMAIL PROTECTED]> [2005-10-05 16:30]:
> On 10/5/05, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
> >I think if I keep this up I¹ll end up proposing actually
> >descriptive names all over the map for where Rose::* stuff
> >could have been placed. I think I¹ll instead stop here and
> >co
On Wed, 5 Oct 2005, John Siracusa wrote:
If Dave searched for "URI" or "URI from hash" he'd have seen Rose::URI on
the first results page:
I did do such a search, but I ignore Rose::URI cause of the name ;)
Here's the thing. "Rose::URI" says to me "this module integrates into
John's Rose fr
On 10/5/05, A. Pagaltzis <[EMAIL PROTECTED]> wrote:
> I think if I keep this up I¹ll end up proposing actually
> descriptive names all over the map for where Rose::* stuff could
> have been placed. I think I¹ll instead stop here and continue to
> consider Rose::* an ignorable walled garden whose us
* John Siracusa <[EMAIL PROTECTED]> [2005-10-05 14:55]:
> It's not "cutesy", it's branded! :)
And the difference is…? :)
> Since they are all interdependent in one way or another (via
> inheritance and also intended usage scenarios) I think it makes
> perfect sense to have them all under a brande
On 10/5/05 8:40 AM, A. Pagaltzis wrote:
> * John Siracusa <[EMAIL PROTECTED]> [2005-10-05 14:31]:
>> This is pretty close to what you want already:
>>
>> http://search.cpan.org/dist/Rose-URI/lib/Rose/URI.pm
>
> Which demonstrates why cutesy names are a bad, *bad* idea. I¹d
> never have thought to
* John Siracusa <[EMAIL PROTECTED]> [2005-10-05 14:31]:
> This is pretty close to what you want already:
>
> http://search.cpan.org/dist/Rose-URI/lib/Rose/URI.pm
Which demonstrates why cutesy names are a bad, *bad* idea. I’d
never have thought to look whether something called “Rose::URI”
scratche
On Oct 5, 2005, at 6:57 AM, A. Pagaltzis wrote:
I thought about this precise point some more, and then I realized
there is an established term for an interface which lets you
create (and only create) objects of another class.
How about URI::Factory?
That one I like.
-Ken
On 10/5/05 3:21 AM, Dave Rolsky wrote:
> On Wed, 5 Oct 2005, Xavier Noria wrote:
>> Wouldn't you expect such a constructor in URI.pm itself, what about a patch?
>> Maybe it is not there already by design though.
>
> A couple things:
>
> - This module probably won't support _all_ types of URIs, at
* Dave Rolsky <[EMAIL PROTECTED]> [2005-10-05 09:45]:
> Excellent point! Um, how about ...
>
> URI::NamedParams
> URI::Constructor
> URI::Creator
I was going to suggest URI::Builder, but that’s more for
interfaces where you accumulate the end result over multiple
calls.
Of the ones you list, UR
On Wed, 5 Oct 2005 02:44:33 -0500 (CDT), Dave Rolsky wrote:
> On Wed, 5 Oct 2005, Orton, Yves wrote:
> > FromHash seems inappropriate, as you aren't actually using a hash there.
>
> Excellent point! Um, how about ...
>
> URI::NamedParams
> URI::Constructor
> URI::Creator
>
> The first is cleare
Title: RE: Name for module to construct URI from named params?
> > FromHash seems inappropriate, as you aren't actually using
> a hash there.
>
> Excellent point! Um, how about ...
>
> URI::NamedParams
> URI::Constructor
> URI::Creator
>
> The first
On Wed, 5 Oct 2005, Orton, Yves wrote:
I'm planning to extract some code from MasonX::WebApp and release it
separately. All it does is take a set of named params and
return a new
URI object from it:
my $uri = URI::FromHash::uri( scheme => 'http', domain => ... );
It'll probably just have t
Title: RE: Name for module to construct URI from named params?
> I'm planning to extract some code from MasonX::WebApp and release it
> separately. All it does is take a set of named params and
> return a new
> URI object from it:
>
> my $uri = URI::FromHas
On Wed, 5 Oct 2005, Xavier Noria wrote:
On Oct 5, 2005, at 7:39, Dave Rolsky wrote:
I'm planning to extract some code from MasonX::WebApp and release it
separately. All it does is take a set of named params and return a new URI
object from it:
my $uri = URI::FromHash::uri( scheme => 'http
On Oct 5, 2005, at 7:39, Dave Rolsky wrote:
I'm planning to extract some code from MasonX::WebApp and release
it separately. All it does is take a set of named params and
return a new URI object from it:
my $uri = URI::FromHash::uri( scheme => 'http', domain => ... );
It'll probably jus
23 matches
Mail list logo