forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ross Gardler <>
Subject Re: forrest:views and xhtml2
Date Fri, 09 Sep 2005 10:41:31 GMT
Thorsten Scherler wrote:
> On Fri, 2005-09-09 at 09:31 +0100, Ross Gardler wrote:
>>Thorsten Scherler wrote:
>>>On Fri, 2005-09-09 at 08:56 +0100, Ross Gardler wrote:


>>>>Views are to go into core as soon as they are mature enough.
>>>If they are not mature enough, why are they going then into this plugin?
>>>What are missing to make them mature enough?
>>I didn't say they were not mature enough *you* did. I proposed moving 
>>them into core immediately after the 0.7 release. I proposed it again 
>>about a week ago.
> Hmm, who said they have to go into the whiteboard instead of let them in
> the core plugins in the first place? 

That was during the 0.7-dev phase, now we are in the 0.8-dev phase. 
Things move on.

> If you think they should move does
> it mean that we e.g. agree on the naming? Every second mail your are
> complaining they are to complicated and need changing, now I want to
> discuss this but ...(change my name with yours):
> "Honestly Thorsten, sometimes I find it impossible to please you." 

OK, so discuss it. I have made my opinion clear on the naming, I'm with 
Davids suggestion of using structurer and I have started using it in my 
mails and *experimental* whiteboard work. I've been using forest:views 
as a collective term for the whole lot (structurer, contracts, 
templates, themes etc.).

I'm comfortable with that.

>>>>XHTML2 work requires a rewrite of a large proposrtion of core.
>>>agree, but should we not just focus on the document part first?
>>The document part is done, now for the next stage.
> Hmm, really? 

Yes. If you think otherwise please actually say why.

The goal of Forrest Tuesday was to make Forrest core work with XHTML2 
*not* to rewrite the core sitemaps or the forrest:views implementation.

The current *experimental* whiteboard plugin works, in a limited way, 
with XHTML2 at its core. It does not rewrite any of the sitemaps. what 
it does to is provide a base on which we can convert the existing 
contracts to work with XHTML2.

>>>>Skins are to be deprecated when views go into core.
>>>Should we not state this officially first?
>>I suppose we have never called a vote on this buyt I think it is pretty 
>>clear to anyone who reads the dev list.
> That it is obvious not makes it official.

What is the problem with conducting some *experimental* work in the 
whiteboard ahead of the vote being taken?

>>>>Doing the XHTML2 work in skins would be wasted effort.
>>>Agree partly, you and me having the same thought of getting away from
>>>tab, menu, doc and site processing, but are we *all* agree on that? 
>>I think so, all the discussions about forrest:views have been about 
>>doing this and the TR states it.
> *You* stated that we need to modify the TR if we agree on the
> terminology of views. What happened with that?

I suggested modifying the title of one section in order to remove 
confusion between forrest:views and forrest:themes. I did *not* suggest 
changing the meaning of the document.

What is the problem here?

>>However, we never had a vote, so it is not *the* way of doing it yet.
> Exactly. I would like to see that is official.

OK, if you think the community has discussed it/explored it enough to 
call the vote then do so. Personally I think we need to experiment a 
little more first. The XHTML2 plugin is one part of that experimentation.

>>>>Therefore we use views.
>>>Actually I have problems to follow this logic, see the question above.
>>I find it ironic that you now want to talk about things rather than just 
>>get on with them. 
> Why? I never stated, one do not have to talk. It is about finding the
> balance. Going ahead and trying to rewrite the core of forrest because
> one person stated in a sentence is not what I meant. Even more if one or
> two are trying to do this by themselves. 

We had a +1 from you to use views I figured that meant you supported the 
idea, why do we need to discuss further? The rest is implementation detail.

Here's how I see it working.

We now have an XHTML2 plugin that gives us a framework for converting 
the contracts to XHTML2. We can therefore get on with the important task 
of converting part of the future Forrest implementation (contracts) to 
work with XHTML2. In the meantime you and others can continue your 
refactoring work on the other part of the future Forrest implementation 

When the refactoring is complete we can bring it all together and we are 
  ready to test (OK it won't be that simple, but we need to move 
forwards somehow).

> It is like comparing apples with peaches (sorry if that not makes sense
> in English it is a literal translation).

We (UK) say Apples and Oranges so it works for me.

>>I believe this is the right approach, on Forrest 
>>Tuesday I did a load of work towards this approach (thanks to the 
>>groundwork put in place by others earlier in the day).
> I think I did some comments on that in the commits. 

Yes, and my response is still the same. We needed to get up and running 
quickly, we are now up and running in an experimental environment. We 
now have a framework for converting the contracts to XHTML2

> I do not think the
> implementation of views is done well (honestly I think actually is done
> really bad). Stating I can fix that is not the way I will go, I do not
> feel like cleaning up that code because it would be faster for me to do
> it from the ground up.

Great - please go ahead and do it. Nobody is saying you have to work on 
the code in the XHTML2 plugin, you keep working on the existing views 
plugins using XDocs. When the refactoring is complete we can merge the 

> "If you want to change something give an example that is working and we
> can discuss this."
> I am trying to discuss the approach you have chosen right now, or not? I
> thing with this approach we doing *too* much extra work and I am not
> willing to take that burden, sorry.

All the XHTML2 plugin does is provide a *version* of the *current* core 
that works with XHTML2. That was the *goal* of Forrest Tuesday.

It is a long way from complete, but it is a long way further on than it 
was on Monday. Clearly it is not the approach that you wanted to take, 
sorry, but we worked hard on that and you should recognise that and draw 
some lessons from it rather than just complaining.

>>So lets have specific reasons why this is not the right approach so that 
>>we can get on with improving what we have.
> See my and your comments on the commits:
> On Wed, 2005-09-07 at 13:15 +0100, Ross Gardler wrote:
>>Because, as I said, I wanted to get on with some work and didn't want
>>have to debug code that was not working. Tim said he was going to fix 
>>it, but we got onto other things.
> We should concentrate doing one step at a time.

Which task do you want to concentrate on?

The task for Forrest Tuesday was to get XHTML2 working not to refactor 
the sitemaps or views. That is what the XHTML2 plugin does (yes, it 
provides a framework in which we can do the wider refactoring, but it 
does not do that yet)

It is not possible to go further forward without doing some refactoring 
of views. So now we need to start doing that. I fail to see what the 
problem is, lets stop complaining about the work that we *have* done and 
proceed with making a plan for the next stage.

>>Because I wanted to work on XHTML2 which was the goal. I thought this 
>>was needed since it was in the sitemaps. I don't understand what you 
>>have done in the LM fallback resolver, you weren't present and I
>>to move forward, at that point I was on my own - go ahead and make
>>necessary changes - you now have code to work with.
> If you are not understand something then *please* ask. Like stated above
> cleaning this code cost more time then to write it again.

You were not present to ask and those present did not have the 
background knowledge to start from scratch. Therefore, with some 
reluctance (check the logs), chose to copy across the whole lot in order 
to get started. We recognised (check the logs, code comments and commit 
messages) that this was not the best approach but was the fastest way to 
get working.

If *you* want to start from scratch and help us in an asynchronous 
fashion that is fantastic. We simply need to rip out structure.xmap and 
start work. I assume this will be your approach since you say that it 
will be quicker to start from scratch than to clean the existing code. 
Surely, you can understand that it was all put into a single file to 
make it *easy* to do this subsequent refactoring, i.e. stip out the old 
versions, write the new.

So far nobody has said no when you have expressed a desire to do this, I 
can't see why you have a problem with these experiments and our 
*support* of your desire to refactor views for this work.

Please stop complaining about the *considerable* efforts of all those 
who were present on Forrest Tuesday and recognise the contribution those 
people have made to these experiments in moving over to XHTML2.

As a result of this work you will now find that more devs understand how 
views work, even more now that Tim took even more time time to document 
his experiments.


View raw message