[squeak-dev] what is holding back Smalltalk?

David Mitchell david.mitchell at gmail.com
Sat Nov 22 13:42:38 UTC 2008


Some guys at IBM are working on exactly this. File-based Smalltalk
under Eclipse.

Based on IBM Smalltalk, replacing Envy with CVS. Originally called
project Black Knight and Wrath, I think it is now known by the more
pedestrian Smalltalk Development Tools (stdt):

http://www.ibm.com/developerworks/blogs/page/stdt

It sounds to me like it is a hybrid system. There is an in-memory
image but all of the tools work on the parallel source. The image
isn't loaded on startup, though.

IBM uses it internally (sounded like one of the reasons was in
producing the J9 Java VM, which makes sense historically). Someone
from the team presented at a conference. It and a separate interview
are both available from the Industry Misinterpretations podcast.

I believe there is an effort to open-source stdt, but I'm sure we can
appreciate the complications of a task like that!

If Eclipse had support, I could see NetBeans might someday have
support for Smalltalk. Tor Norbe made a joke about it on the Java
Posse podcast after discussing Squeak on the JVM. There certainly are
a lot of ex- and closet Smalltalkers at Sun. If Seaside gets enough
buzz with the web developer community...



On Sat, Nov 22, 2008 at 5:40 AM, Martin Beck
<martin.beck at hpi.uni-potsdam.de> wrote:
> Bert Freudenberg wrote:
>>
>> Smalltalk is powerful precisely because it is different than today's
>> popular programming environments. The idea of a "live system" is too strange
>> for the masses.
>>
>> The thing with the popular languages is that they all are used in pretty
>> much the same way - write source code in a file in an editor or IDE of your
>> choice, build your program, run it, debug it, ship it. This makes it
>> relatively easy to switch to a new language, it's basically just a different
>> syntax and a change in the makefiles. You can easily replace parts of your
>> project with pieces in another language. The SCM can work the same. You can
>> continue to use the editor you know in your sleep.
>>
> While this still leaves the question, what is holding back "scripty"
> Smalltalks like GNU Smalltalk, one possibility is to build up an IDE similar
> to Squeak around a text-based Smalltalk: Using the existing sources, a "live
> system" containing the IDE classes and objects is built around them. Changed
> sources are propagated to the file system and vice versa.
>
> Such a hybrid system (which perhaps has already been done - I don't know)
> would combine advantages from both worlds. A python scripter can still use
> text editors, a Smalltalk world user can still image based development (for
> instance, GNU ST also has image saving), while both work on the same system.
> This also (almost) solves the deployment issues in Squeak, like having to
> hide all IDE facilities from my end-users.
>
> IMHO, if - and only *if* - the Smalltalk open source community wants to
> reach a wider range of possibile users, it has to open itself and take some
> step towards existing solutions, whether they might be inferior or not. This
> does not necessarily mean, to give up Smalltalk's special characteristics.
> For now, I see no reason why a Pythonist or Rubyist should completely throw
> away almost everything he has learnt about file-based software development
> in order to use an admittedly more expressive and powerful, but incomplete
> (where is my USB support? :( ) and unmodular system? Why does everything
> have to be replaced, like SCM, UI, editors...? Squeak does not have the
> manpower to provide bug-free tool-support for everything, which is BTW
> called "reinvention". I believe, things like "image as a file-system",
> SqueakSVN, SqueakGTK and similar projects are the right direction but still
> not enough...
>
> Best regards,
> Martin
>
>



More information about the Squeak-dev mailing list