It is very long post, so i decided to put it on my blog.
http://computeradventures.wordpress.com/ i more or less realizing the scale of changes in order to achieve it, but i also wonder why we do not pay attention to this aspect.. maybe i miss something, maybe we don't need it at all. or maybe there's less painful alternatives... -- Best regards, Igor Stasenko. |
On Tue, Mar 6, 2012 at 4:35 PM, Igor Stasenko <[hidden email]> wrote: It is very long post, so i decided to put it on my blog. My main task this year at Cadence is to make Newspeak embeddable. So I will have cycles to work on this :) I've done it before with VisualWorks, although it never made production. I'd *like* to use this as cover for working on the new object representation too, but that's proving harder to sell.
best, Eliot |
On 7 March 2012 01:51, Eliot Miranda <[hidden email]> wrote:
> > > > On Tue, Mar 6, 2012 at 4:35 PM, Igor Stasenko <[hidden email]> wrote: >> >> It is very long post, so i decided to put it on my blog. >> >> http://computeradventures.wordpress.com/ >> >> i more or less realizing the scale of changes in order to achieve it, >> but i also wonder why we do not pay attention to this aspect.. >> maybe i miss something, maybe we don't need it at all. or maybe >> there's less painful alternatives... > > > My main task this year at Cadence is to make Newspeak embeddable. So I will have cycles to work on this :) I've done it before with VisualWorks, although it never made production. I'd *like* to use this as cover for working on the new object representation too, but that's proving harder to sell. > yeah, it is hard to sell things which "should be done yesterday", especially when there's other things which should be done yesterday. I just hoping that they do not make same mistake(s) of the past of pursuing the project without developing the user base, because it is doomed to fail. A Squeak and Pharo could provide such user base, as well as serve as a source of developers and even improvements. I'd like to know if there are understanding of this point. -- Best regards, Igor Stasenko. |
In reply to this post by Igor Stasenko
I don't think embeding should be the focus. I think that the focus should be an improvement in FFI and thread support . When I look at python I see a language that started with a rather huge emphasis on embedding, its was a way to simplify your c/c++ app but nowdays embedding python is frowned upon and generally considered a bad practice among the python developers (python mantra -> dont embed , extend) . And if you see how much successful ctypes have been for python its quick to figure out what direction python is going. Lua on the other hand though it always
prioritised embedding never became really popular. The main problem that any coder will have with smalltalk is not that he or she wont like it, but the fact that he / she will have to make a rather big compromise considering the libraries. Python became so popular not only because its a great simple easy to use language but also because has a great amount of mature libraries. But those libraries did not came out of the blue , they are not really "python" libraries, the vast majoriry of them are ports from c/c++ libraries and I think smalltalk could use the same recipe of success. So anything that can give a big boost to the libraries of smalltalk is something that its my personal opinion the one thing , the only thing , that smalltalk community should concentrate on. So no I dont think embedding is worth our
attention. But as I said, that's just my personal opinion. From: Igor Stasenko <[hidden email]> To: Pharo Development <[hidden email]>; The general-purpose Squeak developers list <[hidden email]> Sent: Wednesday, 7 March 2012, 2:35 Subject: [squeak-dev] Smalltalk as embedded platform/language? It is very long post, so i decided to put it on my blog. http://computeradventures.wordpress.com/ i more or less realizing the scale of changes in order to achieve it, but i also wonder why we do not pay attention to this aspect.. maybe i miss something, maybe we don't need it at all. or maybe there's less painful alternatives... -- Best regards, Igor Stasenko. |
On 7 March 2012 10:09, dimitris chloupis <[hidden email]> wrote:
> I don't think embeding should be the focus. I think that the focus should be > an improvement in FFI and thread support . > > When I look at python I see a language that started with a rather huge > emphasis on embedding, its was a way to simplify your c/c++ app but > nowdays embedding python is frowned upon and generally considered a bad > practice among the python developers (python mantra -> dont embed , extend) > . And if you see how much successful ctypes have been for python its quick > to figure out what direction python is going. Lua on the other hand though > it always prioritised embedding never became really popular. > > The main problem that any coder will have with smalltalk is not that he or > she wont like it, but the fact that he / she will have to make a rather big > compromise considering the libraries. Python became so popular not only > because its a great simple easy to use language but also because has a great > amount of mature libraries. But those libraries did not came out of the blue > , they are not really "python" libraries, the vast majoriry of them are > ports from c/c++ libraries and I think smalltalk could use the same recipe > of success. So anything that can give a big boost to the libraries of > smalltalk is something that its my personal opinion the one thing , the only > thing , that smalltalk community should concentrate on. > well, this is road in both directions: the more people using smalltalk, the more ports will appear, and the more ports appear, the more people will use smalltalk. But i don't agree that embedding is something which don't have to be there, while i agree that the most of developers will never use it, except in cases when you already having a big code base in language like C, and want to add some scripting features, this is where embedding helps. It will be always harder to use language as embedded , than just using VM "out of the box". This is for sure. But from other side, it is much easier to control your life in a way like that, and if we think about commercial apps, i think they would prefer to use it as embedded or not use it at all. > So no I dont think embedding is worth our attention. But as I said, that's > just my personal opinion. > > ________________________________ > From: Igor Stasenko <[hidden email]> > To: Pharo Development <[hidden email]>; The > general-purpose Squeak developers list > <[hidden email]> > Sent: Wednesday, 7 March 2012, 2:35 > Subject: [squeak-dev] Smalltalk as embedded platform/language? > > It is very long post, so i decided to put it on my blog. > > http://computeradventures.wordpress.com/ > > i more or less realizing the scale of changes in order to achieve it, > but i also wonder why we do not pay attention to this aspect.. > maybe i miss something, maybe we don't need it at all. or maybe > there's less painful alternatives... > > -- > Best regards, > Igor Stasenko. > -- Best regards, Igor Stasenko. |
Free forum by Nabble | Edit this page |