On Sat, 2002-10-19 at 06:33, Sam Ruby wrote:
> John McNally wrote:
> > -1.  
> > Jakarta already has two webapp frameworks and I do not see any reason to
> > add another.  
> 
> It is a non-goal of Jakarta to have only one webapp framework, or to 
> limit itself to two.
> 
> - Sam Ruby
> 

If a project is proposed that overlaps a (or a few) current project, I
just think the bar needs to be a bit higher for approval.  If someone
proposed another java regex package, I think many people would want to
see distinguishing features and even if a few existed, it should be
clear that it would be extremely difficult to add the functionality to
one of the current projects or an attempt has been made to work with one
of the current projects and the communities are incompatible.

With database connection pools, I think there were 4 implementations
floating around the jakarta projects.  When I started to look at
upgrading turbine's version or dropping it for one with the features I
was after, I was unable to find a replacement.  This included a survey
outside jakarta where I investigated PoolMan.  Unfortunately I did not
look into avalon's pool which may have met my requirements, but
misconceptions led me to overlook it.  So I set about to create the cp
that implemented the current api's as specified by jdbc.  I still did
not want to do this within turbine, so I engaged the connection pool
project in the commons.  Now it turns out the developer of PoolMan
wanted to stop development and it was proposed that it be brought to
apache.  

I would have said the same thing: jakarta already has a couple database
connection pools, why do we need this one.  And in addition the ones
that are here already implement the latest specifications, while this
proposed one does not.  But PoolMan has name recognition, so it is able
to overcome my resistance to add YetAnotherDBCP.  And it has a member of
Apache who is pushing its adoption, which helps to alay my concerns
about lack of a developer community, though not completely.

I think one of the goals of jakarta is to create high quality
implementations of recognized standards and another is to try to create
standards where they do not formally exist by developing a high quality
technology that is able to become a defacto standard.

As much as I hate it, JSP is the recognized standard for webapp
development.  Jakarta's development of a general purpose java templating
technology, Velocity, is a valid alternative and is not even in direct
conflict with JSP.  But it is a simple, powerful alternative to JSP as
well. Does tapestry give us another alternate template system that is
only usable within the framework?

Granted I could try to investigate Tapestry in depth to answer all my
reservations, but I'm busy and on the surface the project seems to
overlap several existing projects.  My -1 is not a statement that
Turbine (or Struts, Velocity, Avalon) should not have any competitors
within Jakarta.  I would prefer that Tapestry make the case that it
offers something that these projects do not and I don't think the
original proposal makes the case forcefully enough.

john mcnally
 


--
To unsubscribe, e-mail:   <mailto:general-unsubscribe@;jakarta.apache.org>
For additional commands, e-mail: <mailto:general-help@;jakarta.apache.org>

Reply via email to