Welcome to this new smalltalk that was not really advertised :)
http://www.ambrai.com/smalltalk/
Stef
On 24 Jun 2004, at 07:13, stéphane ducasse wrote:
Welcome to this new smalltalk that was not really advertised :) http://www.ambrai.com/smalltalk/
FYI...
Ambrai: '52761747 bytecodes/sec; 3742798 sends/sec' Squeak: '102236421 bytecodes/sec; 4102275 sends/sec'
Ian
Not bad at all for a first beta! Also because it uses native UI elements it feels FAR faster than Squeak. Obviously because it uses native UI elements you can not change them.
Michael
On Jun 24, 2004, at 10:55 PM, Ian Piumarta wrote:
On 24 Jun 2004, at 07:13, stéphane ducasse wrote:
Welcome to this new smalltalk that was not really advertised :) http://www.ambrai.com/smalltalk/
FYI...
Ambrai: '52761747 bytecodes/sec; 3742798 sends/sec' Squeak: '102236421 bytecodes/sec; 4102275 sends/sec'
Ian
On Jun 25, 2004, at 8:33 AM, Michael Latta wrote:
Not bad at all for a first beta! Also because it uses native UI elements it feels FAR faster than Squeak. Obviously because it uses native UI elements you can not change them.
Yes, although I think we can do much better at native OS X UI in Squeak with the Objective C bridge, because we have access to the full AppKit class library rather than going through the C API as Ambrai does. There are lots of nice things about Ambrai, but I'm bewildered that they chose to use Carbon instead of Cocoa for the UI.
For those of us who care about such things, I think there's potentially a very nice model there for commercial use of Squeak: develop web apps on OS X using a native UI environment, and then deploy headless on linux or any other platform you care for.
Avi
Yes, although I think we can do much better at native OS X UI in Squeak with the Objective C bridge, because we have access to the full AppKit class library rather than going through the C API as Ambrai does. There are lots of nice things about Ambrai, but I'm bewildered that they chose to use Carbon instead of Cocoa for the UI.
For those of us who care about such things, I think there's potentially a very nice model there for commercial use of Squeak: develop web apps on OS X using a native UI environment, and then deploy headless on linux or any other platform you care for.
I try to understand what you imply :). You mean that some people would be interested to have an environment on mac OS X that looks like a mac application and not squeak :) and that they would deploy applications I imagine with Seaside running on other platforms. Do I imply correctly that this would be targeted at non squeakers? Is it correct?
Stef
On Jun 25, 2004, at 1:35 PM, stéphane ducasse wrote:
For those of us who care about such things, I think there's potentially a very nice model there for commercial use of Squeak: develop web apps on OS X using a native UI environment, and then deploy headless on linux or any other platform you care for.
I try to understand what you imply :). You mean that some people would be interested to have an environment on mac OS X that looks like a mac application and not squeak :) and that they would deploy applications I imagine with Seaside running on other platforms.
Right - it wouldn't look like Squeak in that it wouldn't look like Morphic, although it would still be Squeak in every other respect. It would be useless as an educational or simulation environment, but would be great for writing network code.
Do I imply correctly that this would be targeted at non squeakers?
I'd rather say, would expand the group of potential Squeakers. Many Smalltalkers are unwilling to use Squeak because they don't like Morphic. It's also a hard sell in corporate environments, whereas VisualWorks and Dolphin are easier - not *just* because of UI, but it's a factor. This is silly and shortsighted of them, but it doesn't help matters to tell them so.
This isn't about making Squeak a good platform for developing native UI desktop apps - that's a huge amount of work, which I'm happy to leave to Ambrai and Dolphin. It's certainly not about doing native UI portably. It's just about putting together enough native development tools so that developers on some platforms (or even just one) get an improved look and feel for browsers, inspectors, debuggers, etc, with Morphic as a fallback for anything else.
Avi
ok I understand :)
For those of us who care about such things, I think there's potentially a very nice model there for commercial use of Squeak: develop web apps on OS X using a native UI environment, and then deploy headless on linux or any other platform you care for.
I try to understand what you imply :). You mean that some people would be interested to have an environment on mac OS X that looks like a mac application and not squeak :) and that they would deploy applications I imagine with Seaside running on other platforms.
Right - it wouldn't look like Squeak in that it wouldn't look like Morphic, although it would still be Squeak in every other respect. It would be useless as an educational or simulation environment, but would be great for writing network code.
Do I imply correctly that this would be targeted at non squeakers?
I'd rather say, would expand the group of potential Squeakers. Many Smalltalkers are unwilling to use Squeak because they don't like Morphic. It's also a hard sell in corporate environments, whereas VisualWorks and Dolphin are easier - not *just* because of UI, but it's a factor. This is silly and shortsighted of them, but it doesn't help matters to tell them so.
This isn't about making Squeak a good platform for developing native UI desktop apps - that's a huge amount of work, which I'm happy to leave to Ambrai and Dolphin. It's certainly not about doing native UI portably. It's just about putting together enough native development tools so that developers on some platforms (or even just one) get an improved look and feel for browsers, inspectors, debuggers, etc, with Morphic as a fallback for anything else.
Avi
On Jun 25, 2004, at 2:46 PM, Avi Bryant wrote:
This isn't about making Squeak a good platform for developing native UI desktop apps.
Speak for yourself. :-)
I'd like to be able to do Cocoa apps in Smalltalk and ship them. Cocoa is a really good api and being able to use tools like IB is very useful. The problem I have with Ambrai is that its not Cocoa. I will hazard the opinion that this will hinder its adoption among developers who already work on the Mac. Ambrai's UI library adds a completely unnecessary learning curve where a simple bridge would pay in spades given Objective C's similarity to Smalltalk wrt message sends.
Just my $0.02
FWIW, I'm hoping soon (next couple months) to kick off another assault on the bridge.
-Todd Blanchard
squeak-dev@lists.squeakfoundation.org