[vote] focusing on internals

[Stefano Mazzocchi <st...@apache.org>]

This vote is about something that I think we are doing wrong: we are
skinning too soon.

Ok, we all love Cocoon, XSLT and the ability to change the entire look
of a site by simply setting a new variable someplace. But is this a good
thing to do at this very moment?

Looking in the forrest CVS we have:

 - avalon-site
 - basic
 - bert
 - forrest-site
 - jakarta-site
 - xml-apache-site

Now tell me: if I make a structural change in the forrest sitemap, am I
supposed to go into every skin and update it as well?

forget it.

I propose a vote to remove *all* skins from forrest but one until we
consider Forrest stable enough to make a release *and* state to skin
authors that the forrest internals are solid enough to create a skin
upon.

For now, we simply can't guarantee that and I don't want any job in the
forrest internals to be slowed down by the fact that all skins must
proceed in parallel. This is just as bad as early optimization. It's
simply not the right way to do things.

I'm not against people using Forrest with their own skins (I'm thinking
about Avalon and POI who happen to live in a forrest-agnostic land), but
I'm concerned about the fact that more skins at this point mean less
speed for forrest in general and I don't like this.

I don't care which skin is kept in CVS and becomes the reference
implementation, but I want just one.

Vote?

-- 
Stefano Mazzocchi      One must still have chaos in oneself to be
                          able to give birth to a dancing star.
<st...@apache.org>                             Friedrich Nietzsche
--------------------------------------------------------------------

Re: [vote] focusing on internals

[Nicola Ken Barozzi <ni...@apache.org>]

From: "Stefano Mazzocchi" <st...@apache.org>

> I propose a vote to remove *all* skins from forrest but one until we
> consider Forrest stable enough to make a release *and* state to skin
> authors that the forrest internals are solid enough to create a skin
> upon.

+1

> I'm not against people using Forrest with their own skins (I'm thinking
> about Avalon and POI who happen to live in a forrest-agnostic land), but
> I'm concerned about the fact that more skins at this point mean less
> speed for forrest in general and I don't like this.

I was thinking of the same thing.
I will rename the forrest to cocoon.cent, as it was in the first place.
At Avalon and POI we will use Forrest when it's ready.

> I don't care which skin is kept in CVS and becomes the reference
> implementation, but I want just one.

bert skin right away +1

-- 
Nicola Ken Barozzi                   nicolaken@apache.org
            - verba volant, scripta manent -
   (discussions get forgotten, just code remains)
---------------------------------------------------------------------

RE: [vote] focusing on internals

[Steven Noels <st...@outerthought.org>]

> From: Stefano Mazzocchi [mailto:stefano@apache.org]

> This vote is about something that I think we are doing wrong: we are
> skinning too soon.

'Soon' isn't the right word, IMO, I'd rather opt for 'much'.

> Now tell me: if I make a structural change in the forrest
> sitemap, am I
> supposed to go into every skin and update it as well?

No, of course not. The work-around of having to add tab2menu.xsl to all
the other skins to get the default build working again was a nice
example of what is happening right now: we have one or two skins that
people are still maintaining, others are not looked after anymore. So
I'm +1 on reducing the number of skins right away, although I'd rather
keep two instead of only one, to avoid the fact of not showcasing the
skinnability anymore. If we have two skins to maintain, people will be
more aware about change management and XML interface definition.
Standardizing or describing the different outputs of the basic pipelines
that are aggregated has been long post due on my to-do list, and having
only one skin wouldn't make this necesary anymore.

> I propose a vote to remove *all* skins from forrest but one until we
> consider Forrest stable enough to make a release *and* state to skin
> authors that the forrest internals are solid enough to create a skin
> upon.

We cannot ensure this solidity without proving it ourselves.

> I don't care which skin is kept in CVS and becomes the reference
> implementation, but I want just one.
>
> Vote?

Let's vote on the two remaining skins?

The bert-skin should be made the default, for sure.

</Steven>