cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Russell <pauljamesruss...@gmail.com>
Subject Re: [RT] The Silkworm Experiment
Date Wed, 23 Feb 2005 16:24:03 GMT
Hi Stefano,

On Wed, 23 Feb 2005 17:01:41 +0100, Stefano Mazzocchi
<stefano@apache.org> wrote:
> Paul Russell wrote:
> > So, let me know what you think! Am I mad? Is this a bad/good idea?
> > What have I missed? Is this something we should take further, or just
> > a distraction?
> Paul,
> 
> first of all, I thank you deeply for having taken the time to write
> this, there is a lot of thinking and you can tell, there is a lot of
> good thinking and elegant design.. there is only one problem:
>  YAGNI

Thanks for the positives :) Fair enough - had a feeling that might be
how you'd feel! There were some reasons for the ideas, related to some
of the feelings I've had for a long time about Cocoon, but you're
right in a way -- a lot of them are about tidying up compromises we've
made en route. Sometimes I think that's a necessary step though, if a
product is going to survive.

> You are proposing design abstractions to solve problems I don't feel we
> have. Branching pipelines? you gotta be kidding. block++? we can't even
> get the real one implemented!

Just one thing: By Block++, I meant 'I like the idea of blocks', not
that I was proposing something over and above blocks. I don't think
I've said anything there which is more capable than the currently
proposed 'real blocks' infrastructure, have I? Maybe I've
misunderstood -- still suffering from a bit of lack of time to keep up
with the mailing list, despite GMail's help ;)

> We some problems on the table, real, painful... and you propose
> solutions to things that nobody really think they are a problem. Not me
> at least. I think you will add a huge burden of abstraction for very
> little functional benefit. I might entirely wrong, but it feels like a
> 2.0 syndrome: there is no incrementality and no plan to get there, no
> requirements, just a whiteboard cleanup of all the little compromises
> that have been made over time... and the goal is elegance, not simplicity.

You're right. The goal is elegance. Are you saying that you think
doing a 'fresh job' of Cocoon 2 was a mistake? I'm surprised to hear
that -- I still believe it was the only way to do it at the time.

You're absolutely right about the lack of migration path. What I'm
spelling out here is a vision -- after all, it is a random thought.
There are some things we could do to make a migration path, but most
of these are for us (e.g. providing Adaptors to wrap up existing
transformers etc.), not the users. The reason for this is simple: What
I'm discussing here is a fundamentally different way of doing things
-- the sitemap tooling approach for a start is 'just different'. It's
difficult to see how to make that kind of change incrementally.

> Been there, done that. I'm sorry, it doesn't work, don't count me in,
> I'm interested in solving real problems in the simplest possible way,
> even if they end up feeling hacky at times.
> The last thing that cocoon needs is a redesign.

Yep, I respect that, and totally understand what you mean about a
re-design -- by definition, it would take resource from the existing
Cocoon work, which is clearly important. Anyway, thanks for the
feedback!

Paul

Mime
View raw message