These are comments received from the review board in response to our PythonTenPaper_

------

The final references must of course be fleshed out as
appropriate URLs.

As you mention the gamut of Python templating, from simple
%-substitutions to full-fledged systems, you might find a space to
mention yaptu, whose claim to fame is to implement embedded
expressions and Python statements in 60 lines of code, see:
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/52305

As yaptu's author, I find it uncanny how many design choices are in
common with Cheetah -- eschewing tags, allowing arbitrary Python
statements with modifiable markers, passing a dictionary for the
substitutions.  Of course, the design spaces are very different
(little cameo versus diamond tiara) and Cheetah is much richer, less
quirky, and architecturally sound (it would be a rather peculiar thing
if it weren't!-).

Some of Cheetah's architectural choices are, it seems to me, a bit
weakly argued in the paper.  I'm basically thinking of the places
where Cheetah's choices are different from Python's (and closer to
other systems, such as JavaScript_ or Basic), such as the decision to
conflate attributes and items, automatically call certain types of
callable (but not all), and the '#block' construct, which I confess is
still a bit mysterious to me after looking at the explanation a few
times.  I'm sure the choices are sound, but adding to the paper's
coverage of these controversial choices (what alternatives were tried,
why was the current choice deemed superior, what possible alternatives
were not explored...) would, I feel, further strengthen your
already-good paper.  Some cursory coverage of other templating systems
you mention, and examples comparing how similar effects are obtained
in a few different frameworks, would really place your paper in
top-class (of course, it's expected you'd select mostly examples
showing Cheetah at its best wrt "competing" templating systems!-)

------

Yet another templating engine.  There are way, *way* too many of these;
[...]

------

A hard to grade paper. The paper itself is good, and rates a '1' on
that account, but the actual subject covered leaves me with an icky
feeling :) It describes yet another templating engine, with different
design decisions than most others. The paper is also a bit weak in the
motivation for some of the choices, and in the technical background of
how Cheetah actually works, but for the average reader/viewer, it
should be about right. I'll keep my biases in check and rate it '1'.

------

The paper was very good. I don't have any obvious suggestions for
changing it. It might be neat to show the generated Python for a
simple template.

------

The domain of template engines for web development is pretty
well-trodden ground; the abstract mentions just four of the best-
known.  So this is not exactly ground-breaking work.

However, there's definitely room in the world for a good template
language for Python.  Cheetah looks like a fairly well-designed entry
in the family of conventional embed-a-programming-language-
in-a-text-markup-language family.  There are a few wonky issues where
it looks faintly like DTML, which is worrying because DTML sucks with
the force of a thousand industrial-strength vacuum cleaners.

The most important hole in the paper is a lack of serious comparison
with other templating engines, especially PSP and DTML (because they
are also Python-based, and presumably compete with Cheetah).

The paper has a few silly formatting problems, in particular a font
change 3 pages in that renders the rest of the paper in Courier.  I
have a few stylistic nits, too.  All in all, pretty well written.