hello list,
I just put up a new version of LambdaMessageSend on SqueakMap (14), and there is now a home page with a bunch of notes attempting to explain as clearly as possible what this thing is: http://www.zogotounga.net/comp/squeak/functionaltalk.htm ... any comment and criticism welcome Stef |
Hi Stéphane,
2007/2/24, Stéphane Rollandin <[hidden email]>: > I just put up a new version of LambdaMessageSend on SqueakMap (14), and > there is now a home page with a bunch of notes attempting to explain as > clearly as possible what this thing is: > > http://www.zogotounga.net/comp/squeak/functionaltalk.htm What you did looks really cool. I will prepare it for next squeak-dev image. Did you already use it in one of your projects ? Can you give us a little background please ? Bye -- Damien Cassou |
Damien Cassou wrote:
> What you did looks really cool. I will prepare it for next squeak-dev > image. Did you already use it in one of your projects ? Can you give > us a little background please ? yes, I use it a lot in µO. it first started from ideous hacks I did on MenuMorph and MenuItemMorph so that I could have highly dynamic behavior from menus. I would set the morphs target as a MessageSend with another MessageSend as receiver and a custom made evaluation trigger as selector, etc. very ugly ! basically what I describe in the first part of the notes: how trying to nest MessageSends naturally leads to LambdaMessageSend. now about all menus in µO rely on it. and again in µO, I also use subclasses of LambdaMessageSend for representing plain mathematical functions. see the NFunctionEditor for example: if you do 'µO objects'->'+ µO tools'->FunctionEditor then click 'help' in the appearing bookmorph, you will see how functions can be defined directly by their equation: this is all based on LambdaMessageSend. the lambda calculus stuff just came along as I tried to figure out the conceptual framework for evaluation, which is a tricky matter. curiously enough I also first started to hack a Lisp interpreter thinking that would be a good way to test the LambdaMessageSend idea, but I eventually came up with Lispkit which does not make use of it at all ! ... still I am now trying to implement a compiler and it seems it could be quite useful at last. so everything came from very practical problems I had with µO (I'm no theorist). trouble is, I could not stop digging and eventually went very astray, since my primary goal is to compose music. I would have been better off learning guitar :) regards, Stef |
Stéphane Rollandin wrote:
> it first started from ideous hacks hideous for sure... Stef |
In reply to this post by Stéphane Rollandin
Thank you very much for your explanation. Why not writing it into the website ?
2007/2/25, Stéphane Rollandin <[hidden email]>: > Damien Cassou wrote: > > What you did looks really cool. I will prepare it for next squeak-dev > > image. Did you already use it in one of your projects ? Can you give > > us a little background please ? > > yes, I use it a lot in µO. > > it first started from ideous hacks I did on MenuMorph and MenuItemMorph > so that I could have highly dynamic behavior from menus. I would set the > morphs target as a MessageSend with another MessageSend as receiver and > a custom made evaluation trigger as selector, etc. very ugly ! basically > what I describe in the first part of the notes: how trying to nest > MessageSends naturally leads to LambdaMessageSend. > > now about all menus in µO rely on it. > > and again in µO, I also use subclasses of LambdaMessageSend for > representing plain mathematical functions. see the NFunctionEditor for > example: if you do 'µO objects'->'+ µO tools'->FunctionEditor then click > 'help' in the appearing bookmorph, you will see how functions can be > defined directly by their equation: this is all based on LambdaMessageSend. > > > the lambda calculus stuff just came along as I tried to figure out the > conceptual framework for evaluation, which is a tricky matter. > > > curiously enough I also first started to hack a Lisp interpreter > thinking that would be a good way to test the LambdaMessageSend idea, > but I eventually came up with Lispkit which does not make use of it at > all ! ... still I am now trying to implement a compiler and it seems it > could be quite useful at last. > > > so everything came from very practical problems I had with µO (I'm no > theorist). trouble is, I could not stop digging and eventually went very > astray, since my primary goal is to compose music. I would have been > better off learning guitar :) > > > regards, > > Stef > > > > -- Damien Cassou |
Damien Cassou wrote:
> Thank you very much for your explanation. Why not writing it into the > website ? I will.. Stef |
Free forum by Nabble | Edit this page |