Referee 1 *******************************************

E: Referee Comments (For Author and Editor)  ------------------------------

The submitted paper is in a working state. It needs much more work. This is obvious while
references are not cited and notes like to be completed are included. This paper should
be completed and reviewed again.


Questions:

	Are there any disadvantages of using Python, only advantages are used.

Remarks:

	It will be very difficult for people not knowing the Globus Toolkit to follow in
	certain sections. The author, should put in appropriate references and may want to
	explain the one or other detail in Globus, to make his paper more readable.
More comments in the examples would not hurt either. (see gass copy, where you should
explain better what the functions do.)

Looking at the download of the toolkit, many more Globus features are provided. The paper
seems not to discuss them.

Describe more about the SOAP interfaces. ... As SOAP is not part of Globus, this seems a
valuable contribution.

F: Presentation Changes

Item 4. remove the /dev ...

Toolkit is in general capitalized.

Include an overview image of the supernova study.

Page 1: examle: spellchecker
Page 3: The projet:  Use spellchecker


Page 3: DO not use the phrase "To do this" avoid using the word do altogether.

Page 3: "it is still an object-oriented architecture" change to is designed with an
object-oriented architecture in mind.

Page 4: "in this section we look" -> In this section we examine

Page 4: show a meaningful example for the GlobusException handeling

Page 6: more references are needed, example python and swig references are missing, each
globus component has a reference, refernec to the example are missing.


Referee 2 *******************************************

pyGlobus is a commodity grid kit (CoG), in Python, and as such it
constitutes a high-level interface to the Grid (i.e., Globus) services.
Python is a very popular, high level, object-oriented language, and
consequently, pyGlobus is a very attractive alternative to Java and CORBA
CoGs. Definitely, it is yet another powerful tool for developing Grid
Computing Environments. In this context, the phrase "yet another" should be
taken as a positive term.

Unfortunately the available version of the paper is incomplete (I would
assume it is a draft rather than a final submission): it contains phrases as
"need tie-in paragraph here" or "to be completed". It is a pity, since even
this version of the paper looks interesting, and it describes an important
development and a good work.

My recommendation is to ask the author for the final version. Once
completed, most likely it will well deserve publication in the special issue
of the Concurrency.

Referee 3 *******************************************

E: Referee Comments (For Author and Editor)  
This paper describes the development of a Python based high-level interface to the Grid
services provided by the Globus Toolkit. In the paper the authors explain why rapid
application development using Grid services is important, look briefly at a motivating
examples, and finally look at the design and implementation of the pyGlobus package.

The paper does a fine job of motivation and describing the interface package.

There are a number of missing sections and reference that need to be filled in.