[Seaside] backwards compatibility vs. progress

Jim Benson seaside@lists.squeakfoundation.org
Wed, 26 Jun 2002 13:35:03 -0700


Burn away!!!

I'll give my view on this. I haven't seen a whole lot of traffic on the list
about people complaining about maintenance issues, so I would assume that
folks aren't having very many problems with the existing (0.93) code base. I
would consider that one a done deal, move on.

As for the rest of it, like release 0.94, I think that part of the deal is
experimenting with the underlying concepts and ideas. Once you get a good
handle on what's involved, and have an overview of what would be a better
way to handle the design, I say go for it !!! That's why you went through
the experiments in the first place. My view is that if people are relying on
the older code base, they need to support themselves with it anyway,
especially if they have invested in it. 0.94 isn't released, it doesn't
matter if it is or not., assuming what is coming is better.  Personally, I
want the best thing I can get going forward, don't keep the weak stuff for
nostalgias sake.

I would encourage you to build something that you are happy with, and not
worry too much about what other people are doing at this point.


> Just a general philosophical question (maintainer angst, if you will).
> I'm struggling with the mythical 0.94 release, trying to shoehorn various
> agendas into it: my desire to separate the template system from the rest
> of the core and make programmatic html generation feasible, the need to
> simplify the bindings system, features that have been requested like
> dynamic/polymorphic subcomponents, and so on.  Since each of these changes
> tends to break existing code anyway, the temptation is becoming stronger
> to just "burn the diskpacks" and start fresh on a new implementation of
> Seaside that takes what I've learned and reshapes it to better accomodate
> these and other features.  What I don't know is how this would affect
> everyone else: some of you are deploying applications, some of you are
> working on ports to other dialects, etc.  How important is a smooth
> migration from version to version, versus a clean, fresh design?
> Obviously the first step is to just release 0.94; the question is where to
> go after that.