On Oct 31, 2005, at 4:22 PM, mario ruggier wrote:

> On Oct 21, 2005, at 1:04 PM, David Binger wrote:
>
>> QP uses an htmltext-like class that is a subclass of unicode, and it
>> organizes the publisher and request and response instances a little
>> differently from Quixote.
>>
>
> For unicode applications (done in quixote), is it safe to assume
> that this htmltext-like unicode subclass is the natural replacement
> (evolution) for htmltext?

That might possibly be true, but I would not say it is safe to assume
that
it is true.

> Similarly, qpy seems like the natural replacement (evolution) of ptl?

I don't know.

> Can these be used in quixote now?

I think that would be very hazardous.

> Will a future quixote adopt these instead?

I don't know.

> Given the many similarities between quixote and qp, and given the
> purposes of each (generic object publisher, specialized application
> framework) what are the reasons why qp is a distinct package,
> rather than not being built on top of quixote?

I don't think htmltext and qpy's h8 should coexist in an application.

> Maybe the above, for a clean implementation of qp, would have
> required disruptive changes to quixote. Do you think however that
> (with possibly changes to quixote) qp will grow to use it? (i.e.
> reduce duplication of generic classes for http/request/response etc)

It seems simpler to avoid that dependency.