I'm New Here

John Hinsley jhinsley at telinco.co.uk
Sun Jul 22 01:04:56 UTC 2001


Andreas Raab wrote:
> 
> Frank wrote:
> > It is difficult to argue with a manager today to use
> > Smalltalk - Smalltalk has it's biggest advantages durin
> > development time -  deployment is still weak.
> >
> > So the translation to solve the deployment problem is attractive.
> 
> Personally, I don't see any relation at all between translating Smalltalk
> code and "solving" deployment issues.

And, surely, if this translation issue was seen to be so big a deal,
Eiffel would be much more popular.

> 
> > And also a problem with almost every Smalltalk if you use
> > an image for deployment - it is decompilable - so you give
> > potentially your sources away.
> 
> This argument surely doesn't hold. You can decompile everything (even your
> translated code) and the only difference is that selectors, temps and
> instance variables give you a pretty good idea what some class is going to
> do. You could trivially replace those names that don't interact with your
> environment (the API) and change them into something utterly unreadable. And
> I want to see the person who can figure out how the balloon engine works if
> I rename all the classes to "T1 ... T1000" the selectors to #t1: ... #t1000
> and all instVars to t1 ... t1000 ;-) It would even be trivial to make it so
> that the names are held in some external map which you can hook up to some
> deployed image for debugging.

IIRC there is some discussion of this in one of the appendices to Ivan
Tomek's "Joy of Smalltalk" and probably elsewhere. Of course, as a
"problem" it doesn't only apply to Smalltalk -- PHP jumps rapidly to
mind.

Cheers

John
> 
> Cheers,
>   - Andreas

-- 
******************************************************************************
Marx: "Why do Anarchists only drink herbal tea?"
Proudhon: "Because all proper tea is theft."
******************************************************************************




More information about the Squeak-dev mailing list