Python/Parrot still alive?

[Berlin Brown]

Is there going to be a python parrot release. I looking on the web and
seeing stuff, stilly pending. That would be cool to see application
servers, via python(I think that could only be with a virtual machine).
Could be wrong of course.

 

[=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=]

Is there going to be a python parrot release. I looking on the web and
seeing stuff, stilly pending. That would be cool to see application
servers, via python(I think that could only be with a virtual machine).
Could be wrong of course.

If you believe that application servers require Parrot because it
has a virtual machine, you are certainly wrong:
- Python does have a virtual machine even without Parrot.
- Applications servers do not fundamentally require virtual
machines.

Regards,
Martin

 

[Paul Prescod]

...

If you believe that application servers require Parrot because it
has a virtual machine, you are certainly wrong:
- Python does have a virtual machine even without Parrot.
- Applications servers do not fundamentally require virtual
machines.

It is interesting how interpreters have been rebranded as "virtual
machines." Python people still use the word "interpreter" and that is
probably not great from a marketing point of view.

Paul Prescod

 

[=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=]

It is interesting how interpreters have been rebranded as "virtual
machines." Python people still use the word "interpreter" and that is
probably not great from a marketing point of view.

I personally make a distinction between an interpreter and a virtual
machine. A virtual machine is one that has a "machine code", i.e. a
set of abstract instructions, operating on machine state. In Python,
the abstract set of instructions is the Python byte code, and the
abstract state is the collection of frame objects, etc.

Some interpreters don't have virtual machines, e.g. Tcl did not
have one until Tcl 8 or so. OTOH, not all virtual machines are
interpreted, e.g. MS .NET is always compiled to native code
("just in time" instead of being interpreted.

Regards,
Martin

 

[Lothar Scholz]

Some interpreters don't have virtual machines, e.g. Tcl did not
have one until Tcl 8 or so. OTOH, not all virtual machines are

So Ruby is an interpreter and not a virutal machine ?

 

[=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=]

So Ruby is an interpreter and not a virutal machine ?

I don't know how Ruby is implemented.

On the language level, there are not interpreted or compiled
languages - there are only interpreters and compilers, and they
are on the level of language implementation. So your question
would be only valid for "Ruby 1.8.1" or some other specific
version. However, I could not answer the question for any Ruby
version.

Regards,
Martin

 

[Peter Hansen]

So Ruby is an interpreter and not a virutal machine ?

Does Ruby have a "byte code" such as Python and Java have?

 

[Lothar Scholz]

Does Ruby have a "byte code" such as Python and Java have?

No. It keeps the parse tree in memory and traverse it during
evaluation. Same as all lisp interpreters(?!?!).

 

[Jeremy Fincher]

No. It keeps the parse tree in memory and traverse it during
evaluation. Same as all lisp interpreters(?!?!).

Definitely not the same as Lisp. In actuality, a Common Lisp
implementation that compiles to bytecode rather than native code is
uncommon; the majority of quality (i.e., complete) implementations
compile to native code.

Jeremy

 

[Michael Hudson]

Definitely not the same as Lisp. In actuality, a Common Lisp
implementation that compiles to bytecode rather than native code is
uncommon; the majority of quality (i.e., complete) implementations
compile to native code.

Well, yeah, but I think what happens to stuff you type into the repl
of e.g. CMUCL could be described as traversing the parse tree. CMUCL
has a bytecode compiler and a native compiler *as well*, but I think
you have to ask for them. SBCL (and some others: Corman, MCL)
compiles absolutely everything to native code, I think.

Cheers,
mwh