Re: [Rails-devel] Versioning

[brett l. <bre...@gm...>]

On Sun, Mar 28, 2010 at 10:38 AM, Jim Black <ji...@ko...> wrote:
>> If the number of safe files that we need to migrate forward to the new
>> format grows substantially, obviously we'll need to change this
>> strategy or develop tools to allow users to migrate save files. For
>> now, I'd just like to see bug fixes and new games/features take a
>> higher priority than save file compatibility.
>
> As a rails/pbem user, host, n00b trainer, and dedicated evangelist- most all outstanding Rails issues don't even /rank/, by comparison, with versioning, and ease-of install/configuration/upgrade/launch.
>
> I can understand that this is hard from the development side of the fence, really.
>
> Nonetheless- I strongly suspect that almost the /only/ users able to keep to one version for a whole game, successfully, are the ones on this list, that are running Rails directly from their shared dropbox folder, and then- only playing a title that was stabilized many months ago.  (eg 1830 18al 18eu).  We've all agreed this is an insecure approach- it's an expert/developer/insider solution, not a community solution.
>
> So, if you think most of your users /are/- or even, /can/- use the same rails version to complete a rails game, as they started with- it's simply not remotely close to the actual reality, in my experience.  It would really help the discussion- and rails' evolution- if you took this into honest account.
>
>  - jim

Jim -

I think you're missing the point.

Each of us volunteers their time to the development of this project.
The hours we dedicate are finite and limited.

So, it's simply a question of how we spend that time.

We certainly could spend that time making slow, deliberate, careful
changes that don't break save file compatibility and doing the
time-intensive regression testing to ensure that's the case.

But... why would we? How does it help us achieve our goal of being
able to support playing as many 18xx games electronically as we can?

If we do things that way, it would take us literally years to
implement any new feature because we'd need to painstakingly test
every nook and cranny of the code to make sure those changes didn't
break anything.

Now, I'm not saying we shouldn't ever test anything. Stefan has done a
great job improving our ability to test saved games.

My point is, our time is limited and we must choose where to spend that time.

The whole point of the e-mail you're quoting is that my recommendation
on how to spend our limited time is to spend it on developing new
features, implementing new games, and generally getting Rails into the
state we'd like it to be in. The price we pay for focusing more on new
development is that we will not be able to spend as much time on
minimizing breakage.

It's frustrating to hit a bug and have it tank your game. I understand
this, perhaps more intimately than you're giving me credit. However,
we can't be all things to everyone. Trade-offs and sacrifices are
necessary if we want to get anything done.

---Brett.