direction of forrest

[Tim Williams <wi...@gmail.com>]

With limited spare time I've been struggling to figure out where/how
Forrest concerns me anymore.  I thought its relevance to me is waning.
 I thought its relevance in general may also be waning.   I mean, I
realize that Forrest has much more to offer, but still, I look at
projects like Exhibit[2] and wonder if we're on the wrong path.

There's been a number of emails that make me question whether that's
true or not.  We have some interesting code in Forrest2.  We have
Thorsten talking[1] about using PoJo/Spring in dispatcher.  We have
Ross beginning to integrate some Ajax-y code. We obviously discussed
much of this before the release but it was thought best to just put it
on the back-burner until after the release.  Unfortunately, after the
release, we seemed to focus on JIRA to answer "what's next"[3].  Now,
I've got nothing against JIRA, I just think we've got bigger "issues"
than those currently captured in JIRA.

So, I ask what's the future of forrest?  Is Forrest2 or some
derivative the answer?  Status Quo?

Disclaimer:  I've tried to follow most discussions in this area but
I've missed where this has already been addressed, I'll humbly accept
a link.  I also realize the energy drain that accompany these
questions but I just feel like we're dog-paddling around in the pool
right now because we don't have a clear bold direction pointed out.

Thanks,
--tim


[1] - http://www.mail-archive.com/dev@forrest.apache.org/msg12567.html
[2] - http://simile.mit.edu/exhibit/
[3] - http://www.mail-archive.com/dev@forrest.apache.org/msg13006.html

Re: direction of forrest

[Thorsten Scherler <th...@juntadeandalucia.es>]

On Mon, 2007-07-16 at 17:18 -0400, Tim Williams wrote:
> On 7/16/07, Thorsten Scherler <th...@apache.org> wrote:
> > On Mon, 2007-07-16 at 08:16 -0400, Tim Williams wrote:
...
> > The one wonderful think that cocoon still is awesome at is that you are
> > not bind to ONE browser. Exhibit and similar project of the MIT are
> > bound to XUL.
> 
> That's not right, they are cross-browser.  Most decent javascript
> libraries out there are also cross-browser.  As I said above though, I
> wasn't pointing to it as a "solution", more as a different way to
> solve problem, has lighter weight, mashup opportunities.  It has
> limits too, of course.

Ok, I must have mixed it up, but like you said the important thing is
"lighter weight". 

> > I personally starting to like xul but is cocoon really obsolete?
> 
> As I indicate above, those are technically unbound topics, but I think
> the answer to the latter is slowly becoming  - yes, for our purposes.

IMO not only for our but as well in general.

...

> > The think we never really decided is which road to go.
> 
> As you point out above, the options we have aren't necessarily at
> odds.  It's less about which new road to choose I reckon and more
> about us making the decision that a new road is indeed in Forrest's
> future.  No?

Exactly. It seems we do not want to make a decision because we are
scared of the consequences. 

salu2
-- 
Thorsten Scherler                                 thorsten.at.apache.org
Open Source Java                      consulting, training and solutions

Re: direction of forrest

[Tim Williams <wi...@gmail.com>]

On 7/16/07, Thorsten Scherler <th...@apache.org> wrote:
> On Mon, 2007-07-16 at 08:16 -0400, Tim Williams wrote:
> > With limited spare time I've been struggling to figure out where/how
> > Forrest concerns me anymore.  I thought its relevance to me is waning.
> >  I thought its relevance in general may also be waning.   I mean, I
> > realize that Forrest has much more to offer, but still, I look at
> > projects like Exhibit[2] and wonder if we're on the wrong path.
>
> I think I know what you mean but can you explain a wee bit more? I mean
> Exhibit is a firefox/xul application. In the end this means exchange
> cocoon for firefox as underlying framework?

Exhibit is not a firefox application (Maybe you're thinking of
Piggybank?). It's a javascript application and isn't tied to a
specific browser.  It's using standard javascript/dom/ajax that's,
nowadays, fairly cross browser specific.  Anyway, I wasn't pointing to
that as a *better* solution, just a less heavyweight solution to what
is a relatively simple problem.  They are mashing up google
spreadsheets and other source formats and providing multiple
"presentations" of them - not so different from our ultimate goal.

> The one wonderful think that cocoon still is awesome at is that you are
> not bind to ONE browser. Exhibit and similar project of the MIT are
> bound to XUL.

That's not right, they are cross-browser.  Most decent javascript
libraries out there are also cross-browser.  As I said above though, I
wasn't pointing to it as a "solution", more as a different way to
solve problem, has lighter weight, mashup opportunities.  It has
limits too, of course.

> I personally starting to like xul but is cocoon really obsolete?

As I indicate above, those are technically unbound topics, but I think
the answer to the latter is slowly becoming  - yes, for our purposes.

> > There's been a number of emails that make me question whether that's
> > true or not.  We have some interesting code in Forrest2.  We have
> > Thorsten talking[1] about using PoJo/Spring in dispatcher.  We have
> > Ross beginning to integrate some Ajax-y code.
>
> Actually the good things of both approach is that it will easy to
> integrate, meaning using the standalone dispatcher as presentation
> framework and forrest2 as multi-channel publisher.

Yeah, I'd like to see Forrest2 take even *more* advantage of Spring
personally.  That was the one thing I didn't care for in that
code-base - the implicit piping with unique protocols.  I'd rather we
make locationmap standard beans (e.g. inject the configuration) and
pull them into Spring as well - I can explain later.  Define the whole
thing inside Spring context and take advantage of Spring's list
inheritance, overriding capabilities, etc.  Anyway, probably too
technical, I'm just trying to get a feel for what others think the way
ahead is and if forrest2 was an interesting conversation that's going
nowhere or not.

> > We obviously discussed
> > much of this before the release but it was thought best to just put it
> > on the back-burner until after the release.
>
> After the release we fall in a comma (like it seems).

Indeed.

> > Unfortunately, after the
> > release, we seemed to focus on JIRA to answer "what's next"[3].
>
> We tried to brainstorm but the discussion died through general
> statements like the jira review you point out.

Exactly.  Maybe we should have this discussion and capture it as the
roadmap in JIRA to take that off the table?  There's lots of "real"
issues in JIRA right now, but none that I see that provide an
interesting story for our roadmap.

> >  Now,
> > I've got nothing against JIRA, I just think we've got bigger "issues"
> > than those currently captured in JIRA.
>
> The think we never really decided is which road to go.

As you point out above, the options we have aren't necessarily at
odds.  It's less about which new road to choose I reckon and more
about us making the decision that a new road is indeed in Forrest's
future.  No?

> >
> > So, I ask what's the future of forrest?  Is Forrest2 or some
> > derivative the answer?  Status Quo?
> >
>
> I for one tend more in forrest2 direction then status quo. IMO cocoon is
> suffering and will more in the future. Meaning projects based on it like
> forrest and lenya are suffering as well.
>
> I believe the future of forrest lies in "bringing back cocoon 1" and
> make it plugable. I am not sure whether exchanging cocoon for xul will
> solve anything. I do not like program in javascript, I prefer java.

That wasn't what I was suggesting,  as I said above, Exhibit isn't
XUL, but I wasn't suggesting it anyway.  Forrest2 is good enough for
me at this point.

> However the current community is IMO in deep sleep. Our code is quite
> stable which for me indicates the next release should be 1.0 with:
>
> - build on ivy dependency management
> - make core independent of theming engine (re-factory regarding
> skins-plugin, all code of the core regarding skinning have to go to a
> plugin)
> -- core build target
> -- core sitemaps
> - internal format based on xhtml2
>
> This brings us a stable branch and the next thing for me would be blueprint planning.
>
> I mean if we manage to write a "documentation" before we write any code
> we can split task for coding.
>
> Anyway ATM I am concentrating my spare time on lenya (we have a moment) but like said
> above lenya is facing in the near future the same situation regarding cocoon.
>
> IMO I guess future development of lenya could very much use something like forrest2.

Thanks for you thoughts.


-- 
--tim

http://williamstw.blogspot.com

Re: direction of forrest

[Thorsten Scherler <th...@apache.org>]

On Mon, 2007-07-16 at 08:16 -0400, Tim Williams wrote:
> With limited spare time I've been struggling to figure out where/how
> Forrest concerns me anymore.  I thought its relevance to me is waning.
>  I thought its relevance in general may also be waning.   I mean, I
> realize that Forrest has much more to offer, but still, I look at
> projects like Exhibit[2] and wonder if we're on the wrong path.

I think I know what you mean but can you explain a wee bit more? I mean
Exhibit is a firefox/xul application. In the end this means exchange
cocoon for firefox as underlying framework? 

The one wonderful think that cocoon still is awesome at is that you are
not bind to ONE browser. Exhibit and similar project of the MIT are
bound to XUL. 

I personally starting to like xul but is cocoon really obsolete?

> There's been a number of emails that make me question whether that's
> true or not.  We have some interesting code in Forrest2.  We have
> Thorsten talking[1] about using PoJo/Spring in dispatcher.  We have
> Ross beginning to integrate some Ajax-y code. 

Actually the good things of both approach is that it will easy to
integrate, meaning using the standalone dispatcher as presentation
framework and forrest2 as multi-channel publisher. 

> We obviously discussed
> much of this before the release but it was thought best to just put it
> on the back-burner until after the release.  

After the release we fall in a comma (like it seems).

> Unfortunately, after the
> release, we seemed to focus on JIRA to answer "what's next"[3]. 

We tried to brainstorm but the discussion died through general
statements like the jira review you point out. 

>  Now,
> I've got nothing against JIRA, I just think we've got bigger "issues"
> than those currently captured in JIRA.

The think we never really decided is which road to go.

> 
> So, I ask what's the future of forrest?  Is Forrest2 or some
> derivative the answer?  Status Quo?
> 

I for one tend more in forrest2 direction then status quo. IMO cocoon is
suffering and will more in the future. Meaning projects based on it like
forrest and lenya are suffering as well. 

I believe the future of forrest lies in "bringing back cocoon 1" and
make it plugable. I am not sure whether exchanging cocoon for xul will
solve anything. I do not like program in javascript, I prefer java. 

However the current community is IMO in deep sleep. Our code is quite
stable which for me indicates the next release should be 1.0 with:

- build on ivy dependency management
- make core independent of theming engine (re-factory regarding
skins-plugin, all code of the core regarding skinning have to go to a
plugin)
-- core build target 
-- core sitemaps
- internal format based on xhtml2

This brings us a stable branch and the next thing for me would be blueprint planning.

I mean if we manage to write a "documentation" before we write any code 
we can split task for coding.

Anyway ATM I am concentrating my spare time on lenya (we have a moment) but like said
above lenya is facing in the near future the same situation regarding cocoon. 

IMO I guess future development of lenya could very much use something like forrest2.

salu2

> [1] - http://www.mail-archive.com/dev@forrest.apache.org/msg12567.html
> [2] - http://simile.mit.edu/exhibit/
> [3] - http://www.mail-archive.com/dev@forrest.apache.org/msg13006.html
-- 
Thorsten Scherler                                 thorsten.at.apache.org
Open Source Java                      consulting, training and solutions

Re: direction of forrest

[Ross Gardler <rg...@apache.org>]

On 16/07/07, Tim Williams <wi...@gmail.com> wrote:
> So, I ask what's the future of forrest?  Is Forrest2 or some
> derivative the answer?  Status Quo?

Thanks for raising this Tim. I don't have time for a full answer right
now, and given I am travelling at present I doubt I will have time for
one until later in the week, or maybe even next week. So, my very
quick answer for now is...

Forrest is dead, long live Forrest2.

When I say Forrest2 I do not necessarily mean the code in whiteboard
(although I think it is pretty cool), I mean the concept of Forrest 2.
For those who have not read my original Forrest2 direction see the
archives that Tim linked to in his post, the summary is...

Forrest is an XML publishing framework, it is not a web framework.

Our dependence on Cocoon is unnecessary and is restricting our
continued development since Cocoon is moving in different directions
these days, i.e. no CLI in Cocoon 2.2 so no way to move to true
blocks, so no way to become more usable.

The value in Forrest is our transformations.

Forrest 2 aims to leverage that value whilst adding flexibility to our
users to be able to use it in whatever environment they need (which
may be Cocoon, but may not be).

BUT...

Who will do this? I'd love to say I will, but I am only one person and
right now I don't have the time or, to be honest, the inclination.
However, if we gather community support around whatever the concept of
Forrest2 becomes I could well be swayed.

I look forward to catching up on this discussion and hopefully helping
to build on whatever develops.

Ross