On 10/31/06, goran@krampe.se goran@krampe.se wrote:
Hi!
"Laurence Rozier" laurence.rozier@gmail.com wrote:
I don't support what Juan's proposal as stated. I'd like to hear why
you,
Juan or anyone else don't agree with what Jecel/Guy proposed
Ok, let me quote so that you don't need to skip back, Jecel wrote:
This is a plan that is practical and which I fully support: if we can unload eToys but not load it back, then let's just include eToys in the full image that we distribute and allow everyone the one way option of removing it. I don't mind at all eliminating the Flash logo, the welcome windows and other stuff every time I begin working with a newly downloaded image. The fact that I can't get any of these easily back if I want has never been a problem since I could just start over from a clean image.
Sure, a reloadable eToys would be even better but I doubt it will happen.
My problem with that proposal is that it means that Morphic is actually not cleaned up - it would just persist as it is today.
... I can see your point but what I had in mind was cleaning up Morphic and being able to "revert to the old Morphic with eToys" ... this is btw also where Spoon comes in(more below) and makes me wonder if there isn't a middle ground that would automatically detect whether a project being loaded is using old Morphic and then load that project as an isolated project along with old Morphic?
It also means that any cleanup that such a removal-script would do in
itself (and Juan has already said that such a delta script is NOT easy to maintain IIRC) would actually fork Morphic itself - people writing code on top of Morphic in the future would have to wonder if it is being run in the image-with-a-clean-morphic-without-etoys or in the image-with-etoys.
In other words, the above sounds like either "status quo" or a fork of
Morphic.
I dislike both. :)
Note that eToys is not a clean "addon" on top of Morphic - it is in fact
an intertwined mess with Morphic methods including lots of logic specifically for eToys. So if you clean out eToys you end up with a different Morphic.
Agreed. However, eToys was originally presented as a Morphic add-on and that is the impression many(most?) new comers will continue to have.
I think it would be much better to lift the burden of eToys, let Juan do
his magic and give us an improved Morphic in the baseline - let anyone step up and retrofit eToys on top of that as a cleaner addon, or not.
I understand the difficulties, but since it's taking away something that currently works, the onus is on the remover in my view.
In
either case - Squeakland is the place to go for eToys and has always been.
Oh if only it were that simple. Now if etoys had started out in a forked image that would be simple, but from the beginning etoys was in the only squeak image. IIRC the early browser plugin images were based on the current Squeak image. The Squeakland website has and continues in many places to make reference to Squeak without distinction and Google has recorded tons of this. No amount of assertion here will change that fact. We can decide that it doesn't matter, but it will be a long time before the close association between squeak.org and etoys is gone. Until then people who are attracted to Squeak because of etoys will continue to have bad impressions. This video http://video.google.com/videoplay?docid=2933987255545037916makes it clear though it is somewhat painful for Squeakers regardless of what distribution they use to watch because it's not only about the etoy experience. Squeak and Smalltalk have done this over and over - we can't blame
if we can unload eToys but not load it back, then let's just include
eToys
in the
full image that we distribute and allow everyone the one way option of removing it.
which mirrors what I said in "Smalltalk Reloaded". More specifically, do folks supporting Juan's proposal disagree with my statement:
All other considerations aside, if e-toys is unloaded from the main
distribution and cannot be easily reloaded by a new Squeaker,there will
be
confusion and for many disappointment and/or some other non-positive experience.
I am not sure I agree with this.
Thing is, eToys in the "main distro"
(squeak.org) has been relatively unmaintained for some time I think (others with better knowledge might disagree). I personally would not trust eToys in official Squeak to work as one would expect it to work relative to the Squeakland image.
I understand but most people coming to the Squeak site don't have your perspective.
And this goes especially for 3.9 and beyond, since Squeakland is 3.8
based.
So to rephrase myself - if we aim to minimize the risk of a non-positive experience of eToys I think we should clearly and distinctly direct people towards the Squeakland distro directly from the main page of Squeak.org for those people looking specifically for eToys.
That makes it less non-positive but it's still not an improvement IMO until the new morphic is used for something ... well new.
For me it's not that I'm opposed to a cleaner Morphic, I just don't want to
add stumbling blocks for wider acceptance at a time when eToy images are getting more and more visibility. Sure it would be nice to see the
default
image with a cleaner Morphic but not at any price - especially since
Spoon
is coming.
Mmmm, I don't see how Spoon relates technically to a cleanup of Morphic. I consider those two efforts quite separate.
The etoy/morphic entanglement will eventually yield to imprinting and with a dynamically configurable image, we'll get the best of both worlds.
It's not clear to me whether the upside of a cleaner Morphic wil=
l be that great or long lasting because I believe we're in the early
stages o=
f a big paradigm shift in which Croquet UI<
http://www.meshverse.com/2006/10/18/the-64-billion-dollar-question-reloa=
ded/>(oh where is Wicket?) is where the action is. In this new paradigm, I don't
kno=
w whether a cleaner Morphic is going to have advantages over Tweak.
It is not so much about which is "best". Tweak is radically different, and many of those ideas are great (AFAICT). I have on the other hand heard people moaning about it being hard to debug and as we all know - it is not "ready" for Squeak.org yet. Again AFAIK.
I don't think we should sit around and wait for Tweak to save the day - we have already done so to a certain extent and I think we should stop doing that. Morphic is nice and it can be small and lean too. Juan offers to give us that and I think that is great.
It is and I wouldn't mind at all seeing it but not at the price that's required for benefits that are not proven. If developers take the new etoyless, improved Morphic and start doing great things with it then a strong case can be made to break with the old. But that's not known yet, we could end up with a core Squeak that can't use etoys AND doesn't have much new to show because more people decide to go with Tweak or wxWidgets or ...
Also - yes, the Squeaklanders did spend time to update eToys in 3.8 with
the changes that were done in the Squeakland image. But will they keep that up? I am very unsure about that, would be neat to hear more from Michael Rueger (or someone) on that issue.
If they indeed have an intention of maintaining eToys in 3.9 and beyond (which I doubt) then that would indeed make the question more complicated.
Cheers,
Laurence
regards, Göran
PS. I don't claim the almighty truth - these are just my opinions based on my perceptions and hunches. :)
I feel the same way. I also am looking at the history of Smalltalker's being our own worst enemy because we didn't consider how big an impact we could have if united in one interoperable ecosystem.
Cheers,
Laurence