web framework advice

Hi guys

I'm looking for a web framework. Does anyone have
experience with any of 
the following ...

     WebWare
     TurboGears
     Karigell
     Others I haven't looked at but you have

The infrastructure I like is Linux, Apache, Kinterbasdb +
Firebird. In 
case you are wondering, the licensing and very low ongoing
admin effort 
for Firebird is behind that choice.

At the moment I'm leaning towards WebWare but that is just
because I 
have looked at it a little bit and it seems theoretically
OK. It has an 
app server which sits behind Apache and replaces CGI for
scalability 
reasons.

Karigell works but uses its own web server and DBMS.
Although I could 
put it behind Apache I'm not sure about the effort to slide
Firebird 
underneath.

I have heard that TurboGears is brilliant but I worry about
"magic" (I 
read that as complex) glue between componentry supplied by
different 
teams. Haven't looked at it really so I shouldn't be
saying that.

I would like to ask about ...

     1. Development complexity (I know, mmmv)
     2. Scalability
     3. Security against black hats

Thanks for any insights

Mike
_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Hi Mike,

 From what I know, TurboGears uses CherryPy at the backend.
It can be 
seen as a site-builder using CherryPy. However, Being
Python-based, 
Karigell and TurboGears have a simplified way to link to
SQLlite and Gadfly.

I've not really done any real web development before.
Nevertheless, 
looking at some other people's experiences, the database
backend is 
irrelevant to the web server or application server.

I am in full support of Firebird, you know that... :P

It may be possible to use mod_python in Apache to do that
link. I 
believe using any forms of data management backend has no
relevance to 
Apache. Can someone comment on this?

Cheers
Maurice

Mike Dewhirst wrote:

>Hi guys
>
>I'm looking for a web framework. Does anyone have
experience with any of 
>the following ...
>
>     WebWare
>     TurboGears
>     Karigell
>     Others I haven't looked at but you have
>
>The infrastructure I like is Linux, Apache, Kinterbasdb
+ Firebird. In 
>case you are wondering, the licensing and very low
ongoing admin effort 
>for Firebird is behind that choice.
>
>At the moment I'm leaning towards WebWare but that is
just because I 
>have looked at it a little bit and it seems
theoretically OK. It has an 
>app server which sits behind Apache and replaces CGI for
scalability 
>reasons.
>
>Karigell works but uses its own web server and DBMS.
Although I could 
>put it behind Apache I'm not sure about the effort to
slide Firebird 
>underneath.
>
>I have heard that TurboGears is brilliant but I worry
about "magic" (I 
>read that as complex) glue between componentry supplied
by different 
>teams. Haven't looked at it really so I shouldn't be
saying that.
>
>I would like to ask about ...
>
>     1. Development complexity (I know, mmmv)
>     2. Scalability
>     3. Security against black hats
>
>Thanks for any insights
>
>Mike
>_______________________________________________
>melbourne-pug mailing list
>melbourne-pugpython.org
>http://mail.python.org/mailman/listinfo/melbourne-pug
>
>  
>

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Hi all,

It looks like the regular meetings have fallen by the
wayside, unless
they've been going on without any associated mailing list
chatter.

I'd quite like to see a little more life in MPUG, I don't
know if anyone
else agrees. I found the regular meetings a bit tough to
attend because
I work full-time and have quite a lot of time demands in my
home life.

To that end, perhaps I could offer to collate and edit some
articles
into an e-magazine, perhaps bringing the first issue into
existence
around June 1? I would be happy to do put such a thing
together, and if
anyone has any articles, cookbook recipes or even simple
ideas, please
feel free to contact me. The existence of a magazine of this
kind might
help people feel a bit more connected? Also, if any of the
more
long-standing members of MPUG would care to contribute an
article on the
history of mpug, or personal accounts, that would be great.

If people think this is a bad idea, feel free to let me know
that, also.
Crocker's rules are fine by me.

Anyone interested in my background can examine my largely
defunct blog
which I maintained actively for a year, at
melbournephilosopher.blogspot.com. I don't have any
webspace currently,
so I haven't got an ego page up right now.

Cheers,
-T

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

> >
> >I'm looking for a web framework. Does anyone have
experience with any of 
> >the following ...
> >
> >     WebWare
> >     TurboGears
> >     Karigell
> >     Others I haven't looked at but you have


 I have used TurboGears to build my personal website (see
link in my
sig) and was very pleased with it.  The combination of Kid
templates and
the CherryPy publishing model was a very good fit for my
head - it all
just made sense.  I cant speak to scalability or performance
really, but
it did find development quick and painless.

 My setup uses sqlite as the relational database simply
because I don't
need anything more powerful - but moving to another database
would be a
one-line configuration change.

 I have the cherrypy standalone server running on
localhost:8080, with
apache forwarding requests to it via mod_rewrite.  Works a
treat.  I
believe other deployment options are possible.

 Don't forget the 20-minute-wiki screencast if you want to
get a feel
for operating  under turbogears:

    ht
tp://files.turbogears.org/video/20MinuteWiki.mov

>
>I have heard that TurboGears is brilliant but I worry
about "magic" (I 
>read that as complex) glue between componentry supplied
by different 
>teams. 
>

 For the most part, read "magic" as "it
just works seamlessly and
painlessly".  I've found the magic to be pretty low
complexity for the
most part.  Quick example of the magic:

   * in vanilla CherryPy, you expose methods  like so:

            expose
            def some_page():
               return "page contents"

      Your methods return strings containing the page
contents.
   
   * in TurboGears, your exposed methods are associated with
a template:

           expose(template="mysite.templates.main")
           def some_page():
               vals = {}
               vals["title"] = "Page
Title"
               vals["body"] = "Body
Contents"
               return vals

      They return a dictionary of values that automatically
appear
      in the namespace of the template, ready to be
inserted.


  I think this is a great combination of two powerful
technologies.

  So, TurboGears gets a thumbs up from me.  My original
website was
written directly using CherryPy + a templating engine
(SimpleTAL), which
was fun but a lot of work.  Moving to TurboGears made things
much
easier.  I haven't tried other frameworks for comparison
purposes.

  A few words of warning though:

   * TurboGears functionality is scattered throughout the
namespaces of
its componenets.  In order to do simple things, you often
need to access
modules under the turbogears, cherrpy, and sqlobject
packages.
Understandable why this is the case, but confusing to
read...

   * TurboGears is growing quickly.  There is an upcoming
0.9 release
featuring loads of wacky new functionality like Widgets
(very neat
looking, by the way), CRUD data management, etc etc etc. 
Expect some
API breakage on the way towards version 1.0

   * According to rumour, some people find Kid templates
abominably
slow, some people find them really fast.  It's been fine
for me, but
then I'm not exactly serving millions of hits on my
personal site .
YMMV, but be prepared for the possibility of a lot of
template tweaking
if you want good performance.


  Please let us know what you end up choosing, and what you
think of it
a few months down the track 


  Cheers,

     Ryan



-- 
Ryan Kelly
http://www.rfk.id.au  | 
This message is digitally signed. Please visit
ryanrfk.id.au        |  http://www.rfk.id
.au/ramblings/gpg/ for details

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Hi,

I'm away from Melbourne for 7 months and will only be back
next month, 
so I must admit that I've not been following up the
happenings.

>To that end, perhaps I could offer to collate and edit
some articles
>into an e-magazine, perhaps bringing the first issue
into existence
>around June 1? I would be happy to do put such a thing
together, and if
>anyone has any articles, cookbook recipes or even simple
ideas, please
>feel free to contact me. The existence of a magazine of
this kind might
>help people feel a bit more connected? Also, if any of
the more
>long-standing members of MPUG would care to contribute
an article on the
>history of mpug, or personal accounts, that would be
great.
>
>If people think this is a bad idea, feel free to let me
know that, also.
>Crocker's rules are fine by me.
>  
>
I am largely in favour of this. I feel that in the OSS
community, there 
is still a substantial lack of documentation and support for
getting 
materials out into the public.

This Melbourne Python Users' Bulletin (or letters) may be a
collection 
of published case studies or personal experiences and
tutorial-like 
articles. From an academic perspective, I believe that at
least the 
technical contents needs to be correct before
"publishing", so even 
though it is not a real academic journal, some forms of
"technical 
peer-reviewing" needs to take place.

I've also just did an initial proposal to Firebird
Foundation about 
starting a peer-reviewed journal. (Let's talk about this
off-list if you 
are interested) But I think the administration and
moderation are the same.

I can offer my assistance to this e-zine as a form of
associate editor 
or sort if needed.

Cheers
Maurice
_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Maurice Ling wrote:
> Hi,
> 
> I'm away from Melbourne for 7 months and will only be
back next month, 
> so I must admit that I've not been following up the
happenings.

I live here and have missed maybe one or two meetings and
the last 
meeting I did attend in Fitzroy didn't quite make a quorum


>> To that end, perhaps I could offer to collate and
edit some articles
>> into an e-magazine, perhaps bringing the first
issue into existence
>> around June 1? I would be happy to do put such a
thing together, 

That is generous - standing ovation from me |||||||

and if
>> anyone has any articles, cookbook recipes or even
simple ideas, please
>> feel free to contact me. The existence of a
magazine of this kind might
>> help people feel a bit more connected? Also, if any
of the more
>> long-standing members of MPUG would care to
contribute an article on the
>> history of mpug, or personal accounts, that would
be great.
>>
>> If people think this is a bad idea, feel free to
let me know that, also.
>> Crocker's rules are fine by me.

Not in python space and especially not in Melbourne! We vote
with our 
feet 
See standing ovation above.

>>  
>>
> I am largely in favour of this. I feel that in the OSS
community, there 
> is still a substantial lack of documentation and
support for getting 
> materials out into the public.

I strongly agree. I'd like to see an open source knowledge
transfer 
system. More on that in a later discussion.

> This Melbourne Python Users' Bulletin (or letters) may
be a collection 
> of published case studies or personal experiences and
tutorial-like 
> articles. From an academic perspective, I believe that
at least the 
> technical contents needs to be correct before
"publishing", so even 
> though it is not a real academic journal, some forms of
"technical 
> peer-reviewing" needs to take place.

Are you suggesting a mail-list review?

Just thinking out loud, I reckon peer review is brilliant in
terms of 
you guys telling me where I'm going wrong when I show you
what I'm 
doing. That is because my objective (not necessarily anyone
else's) is 
to learn and improve my skills in a much more immediate time
frame than 
would be possible within any formal or academic peer review
situation.

Speaking for myself, I would strongly support mail-list
review of 
whatever gets offered. That is where I would gain the most
knowledge. I 
desperately want to see debate and constructive criticism of
everyone's 
offerings because I know that's how I'll learn. Just
seeing the polished 
outcome in the e-zine would not teach me as much. I'm not
against polish 
I just wanna learn.

In fact I would be happy to submit my own puny efforts and
suffer 
enormous embarrassment in the process. The benefit to me
would be 
incalculable.

There is definitely a niche for polished peer reviewed
stuff. Just last 
night I placed an order on Amazon for O'Reilly's Python
Cookbook. The 
cred of the reviewers looks impeccable to me and that is
what persuaded 
me to spend hard-earned money. Even so, I also visit the
ASPN website 
for recipes because I see peer review debate and commentary
there as well.

I think I'm saying we sacrifice bandwidth in the knowledge
transfer game 
when we add formality. IMHO, absolute best and broadest
knowledge 
transfer bandwidth is face-to-face discussion with
whiteboards on hand. 
Next best is either Professor Google or a mail list like
this. The 
narrowest knowledge transfer bandwidth is to buy and read a
book. I'm 
not saying anything about the quality of the knowledge -
just the speed 
and relevance at the time.

A peer reviewed e-zine would be great. The peer review would
be better 
provided people got in and criticised with gusto. Maybe
Crocker's Rules 
should apply (:

> 
> I've also just did an initial proposal to Firebird
Foundation about 
> starting a peer-reviewed journal. (Let's talk about
this off-list if you 
> are interested) 

Don't be shy - I would be more comfortable being included
(or lurking) 
in the discussion.

Run with it ...

Mike


But I think the administration and moderation are the same.
> 
> I can offer my assistance to this e-zine as a form of
associate editor 
> or sort if needed.
> 
> Cheers
> Maurice
> _______________________________________________
> melbourne-pug mailing list
> melbourne-pugpython.org
> http://mail.python.org/mailman/listinfo/melbourne-pug
> 
> 

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Maurice Ling wrote:
> Hi Mike,
> 
>  From what I know, TurboGears uses CherryPy at the
backend. It can be 
> seen as a site-builder using CherryPy. However, Being
Python-based, 
> Karigell and TurboGears have a simplified way to link
to SQLlite and Gadfly.
> 
> I've not really done any real web development before.
Nevertheless, 
> looking at some other people's experiences, the
database backend is 
> irrelevant to the web server or application server.
> 
> I am in full support of Firebird, you know that... :P
> 
> It may be possible to use mod_python in Apache to do
that link. I 
> believe using any forms of data management backend has
no relevance to 
> Apache. Can someone comment on this?

Thanks Maurice

You are right. The application software does the database
work and 
Apache just serves the constructed page.

SQLite or Gadfly would probably do the trick initially.
However I will 
need to think about slipping Firebird in there at some
point.

I have just done a demo version of the app using the file
system for 
persistence all the time thinking I would separate that out
later but I 
now find it is impossible. I took too many yagni shortcuts
telling 
myself it was a throwaway demo. But I had to add a lot of
features I 
hadn't expected and now I'm reluctant to throw it all out.

What I'm saying is the production version has to be better
thought 
through with a view to ongoing maintenance and if I can
predict it needs 
Firebird for scalability, licensing and low support overhead
then I 
should figure that in from the start.

Regards

mike
> 
> Cheers
> Maurice
> 
> Mike Dewhirst wrote:
> 
>> Hi guys
>>
>> I'm looking for a web framework. Does anyone have
experience with any of 
>> the following ...
>>
>>     WebWare
>>     TurboGears
>>     Karigell
>>     Others I haven't looked at but you have
>>
>> The infrastructure I like is Linux, Apache,
Kinterbasdb + Firebird. In 
>> case you are wondering, the licensing and very low
ongoing admin effort 
>> for Firebird is behind that choice.
>>
>> At the moment I'm leaning towards WebWare but that
is just because I 
>> have looked at it a little bit and it seems
theoretically OK. It has an 
>> app server which sits behind Apache and replaces
CGI for scalability 
>> reasons.
>>
>> Karigell works but uses its own web server and
DBMS. Although I could 
>> put it behind Apache I'm not sure about the effort
to slide Firebird 
>> underneath.
>>
>> I have heard that TurboGears is brilliant but I
worry about "magic" (I 
>> read that as complex) glue between componentry
supplied by different 
>> teams. Haven't looked at it really so I shouldn't
be saying that.
>>
>> I would like to ask about ...
>>
>>     1. Development complexity (I know, mmmv)
>>     2. Scalability
>>     3. Security against black hats
>>
>> Thanks for any insights
>>
>> Mike
>> _______________________________________________
>> melbourne-pug mailing list
>> melbourne-pugpython.org
>> http://mail.python.org/mailman/listinfo/melbourne-pug
>>
>>  
>>
> 
> _______________________________________________
> melbourne-pug mailing list
> melbourne-pugpython.org
> http://mail.python.org/mailman/listinfo/melbourne-pug
> 
> 

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Thanks for the 2 strong votes of confidence.

I also vote for some form of peer-review. I'd love to not
to that peer 
review myself, so getting the list to do it makes sense. I
think it's 
(a) nice to have something of yours published and thought
thoroughly 
about, and (b) interesting to see people's opinions of
things. A 
mailing-list review is the easiest. We need a candidate
piece of code + 
short article from the originator, in the form of "An
abstract", "the 
code", "a discussion", "something
chatty". I'm happy to take that as a 
loosely formatted email to myself, edit it in preparation
for peer 
review, then take random mailing list emails and turn them
into some 
kind of coherent narrative.

That could probably form the kernel of the bulletin, around
which an 
e-zine could proudly grow. I believe we have the people here
to do 
something that each of us would be pleased to be associated
with, 
without it being an excessive burden on any one individual.

"In fact I would be happy to submit my own puny
efforts and suffer 
enormous embarrassment in the process. The benefit to me
would be 
incalculable."

Well, that would be great. Perhaps it would get to the point
where 
people *wanted* to submit code for review in order to
improve it. It 
could also spark discussion on hard problems, etc.

As an aside, I bought the ASPN Cookbook and have already
found it 
invaluable. I find it much more useable than the online
version, because 
of the degree of quality control and polish. I am at the
point where I 
don't need a zen master to improve a whole program, I just
need a better 
way to do X, because I can't think how do to X elegantly
without a 
little help. Polish lets you "zoom in" on the
most important component, 
and if you're good enough (most people are), you can take
that and make 
your own code that much better. That's just my 2c, of
course.

A piece on big issues in Open Source and/or Python would be
neat, there 
are lots of things I could think of.

Cheers,
-T

Mike Dewhirst wrote:
> Maurice Ling wrote:
>   
>> Hi,
>>
>> I'm away from Melbourne for 7 months and will only
be back next month, 
>> so I must admit that I've not been following up
the happenings.
>>     
>
> I live here and have missed maybe one or two meetings
and the last 
> meeting I did attend in Fitzroy didn't quite make a
quorum 
>
>   
>>> To that end, perhaps I could offer to collate
and edit some articles
>>> into an e-magazine, perhaps bringing the first
issue into existence
>>> around June 1? I would be happy to do put such
a thing together, 
>>>       
>
> That is generous - standing ovation from me |||||||
>
> and if
>   
>>> anyone has any articles, cookbook recipes or
even simple ideas, please
>>> feel free to contact me. The existence of a
magazine of this kind might
>>> help people feel a bit more connected? Also, if
any of the more
>>> long-standing members of MPUG would care to
contribute an article on the
>>> history of mpug, or personal accounts, that
would be great.
>>>
>>> If people think this is a bad idea, feel free
to let me know that, also.
>>> Crocker's rules are fine by me.
>>>       
>
> Not in python space and especially not in Melbourne! We
vote with our 
> feet  See
standing ovation above.
>
>   
>>>  
>>>
>>>       
>> I am largely in favour of this. I feel that in the
OSS community, there 
>> is still a substantial lack of documentation and
support for getting 
>> materials out into the public.
>>     
>
> I strongly agree. I'd like to see an open source
knowledge transfer 
> system. More on that in a later discussion.
>
>   
>> This Melbourne Python Users' Bulletin (or letters)
may be a collection 
>> of published case studies or personal experiences
and tutorial-like 
>> articles. From an academic perspective, I believe
that at least the 
>> technical contents needs to be correct before
"publishing", so even 
>> though it is not a real academic journal, some
forms of "technical 
>> peer-reviewing" needs to take place.
>>     
>
> Are you suggesting a mail-list review?
>
> Just thinking out loud, I reckon peer review is
brilliant in terms of 
> you guys telling me where I'm going wrong when I show
you what I'm 
> doing. That is because my objective (not necessarily
anyone else's) is 
> to learn and improve my skills in a much more immediate
time frame than 
> would be possible within any formal or academic peer
review situation.
>
> Speaking for myself, I would strongly support mail-list
review of 
> whatever gets offered. That is where I would gain the
most knowledge. I 
> desperately want to see debate and constructive
criticism of everyone's 
> offerings because I know that's how I'll learn. Just
seeing the polished 
> outcome in the e-zine would not teach me as much. I'm
not against polish 
> I just wanna learn.
>
> In fact I would be happy to submit my own puny efforts
and suffer 
> enormous embarrassment in the process. The benefit to
me would be 
> incalculable.
>
> There is definitely a niche for polished peer reviewed
stuff. Just last 
> night I placed an order on Amazon for O'Reilly's
Python Cookbook. The 
> cred of the reviewers looks impeccable to me and that
is what persuaded 
> me to spend hard-earned money. Even so, I also visit
the ASPN website 
> for recipes because I see peer review debate and
commentary there as well.
>
> I think I'm saying we sacrifice bandwidth in the
knowledge transfer game 
> when we add formality. IMHO, absolute best and broadest
knowledge 
> transfer bandwidth is face-to-face discussion with
whiteboards on hand. 
> Next best is either Professor Google or a mail list
like this. The 
> narrowest knowledge transfer bandwidth is to buy and
read a book. I'm 
> not saying anything about the quality of the knowledge
- just the speed 
> and relevance at the time.
>
> A peer reviewed e-zine would be great. The peer review
would be better 
> provided people got in and criticised with gusto. Maybe
Crocker's Rules 
> should apply (:
>
>   
>> I've also just did an initial proposal to Firebird
Foundation about 
>> starting a peer-reviewed journal. (Let's talk
about this off-list if you 
>> are interested) 
>>     
>
> Don't be shy - I would be more comfortable being
included (or lurking) 
> in the discussion.
>
> Run with it ...
>
> Mike
>
>
> But I think the administration and moderation are the
same.
>   
>> I can offer my assistance to this e-zine as a form
of associate editor 
>> or sort if needed.
>>
>> Cheers
>> Maurice
>> _______________________________________________
>> melbourne-pug mailing list
>> melbourne-pugpython.org
>> http://mail.python.org/mailman/listinfo/melbourne-pug
>>
>>
>>     
>
> _______________________________________________
> melbourne-pug mailing list
> melbourne-pugpython.org
> http://mail.python.org/mailman/listinfo/melbourne-pug
>
>   

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

Ryan Kelly wrote:
>>> I'm looking for a web framework. Does anyone
have experience with any of 
>>> the following ...
>>>
>>>     WebWare
>>>     TurboGears
>>>     Karigell
>>>     Others I haven't looked at but you have
> 
> 
>  I have used TurboGears to build my personal website
(see link in my
> sig) and was very pleased with it.  The combination of
Kid templates and
> the CherryPy publishing model was a very good fit for
my head - it all
> just made sense.  I cant speak to scalability or
performance really, but
> it did find development quick and painless.
> 
>  My setup uses sqlite as the relational database simply
because I don't
> need anything more powerful - but moving to another
database would be a
> one-line configuration change.

Presumably it is DB 2.0 compliant.

> 
>  I have the cherrypy standalone server running on
localhost:8080, with
> apache forwarding requests to it via mod_rewrite. 
Works a treat.  I
> believe other deployment options are possible.

WebWare uses a similar approach with mod_webkit which knows
about the 
webkit application server.

> 
>  Don't forget the 20-minute-wiki screencast if you
want to get a feel
> for operating  under turbogears:
> 
>     ht
tp://files.turbogears.org/video/20MinuteWiki.mov

I'll check that out tonight.

> 
>> I have heard that TurboGears is brilliant but I
worry about "magic" (I 
>> read that as complex) glue between componentry
supplied by different 
>> teams. 
>>
> 
>  For the most part, read "magic" as
"it just works seamlessly and
> painlessly".  I've found the magic to be pretty
low complexity for the
> most part.  Quick example of the magic:
> 
>    * in vanilla CherryPy, you expose methods  like so:
> 
>             expose
>             def some_page():
>                return "page contents"
> 
>       Your methods return strings containing the page
contents.
>    
>    * in TurboGears, your exposed methods are associated
with a template:
> 
>            expose(template="mysite.templates.main")
>            def some_page():
>                vals = {}
>                vals["title"] = "Page
Title"
>                vals["body"] = "Body
Contents"
>                return vals
> 
>       They return a dictionary of values that
automatically appear
>       in the namespace of the template, ready to be
inserted.

OK I think I understand this. My home grown efforts are
structured along 
similar lines. I used ordinary html pages (made with
dreamweaver) as 
templates and a dictionary to collect and/or populate cgi
variables. I 
had to build a method to merge the dictionary into the page
which I read 
off disk into a list and then served.

> 
>   I think this is a great combination of two powerful
technologies.
> 
>   So, TurboGears gets a thumbs up from me.  My original
website was
> written directly using CherryPy + a templating engine
(SimpleTAL), which
> was fun but a lot of work.  Moving to TurboGears made
things much
> easier.  I haven't tried other frameworks for
comparison purposes.

It seems to me that WebWare is a complete system built with
all (or 
most) components intended to interface in a more or less
purist python 
OO manner. At first glance it feels sort of heavy duty in
that it is 
layered like the MVC pattern intends. I just don't know
enough python or 
OO or patterns to make valid judgements about the amount of
work. I 
certainly like the *idea* of doing things in a purist way.

> 
>   A few words of warning though:
> 
>    * TurboGears functionality is scattered throughout
the namespaces of
> its componenets.  In order to do simple things, you
often need to access
> modules under the turbogears, cherrpy, and sqlobject
packages.
> Understandable why this is the case, but confusing to
read...

I'm easily confused at the best of times ...

> 
>    * TurboGears is growing quickly.  There is an
upcoming 0.9 release
> featuring loads of wacky new functionality like Widgets
(very neat
> looking, by the way), CRUD data management, etc etc
etc.  Expect some
> API breakage on the way towards version 1.0
> 
>    * According to rumour, some people find Kid
templates abominably
> slow, some people find them really fast.  It's been
fine for me, but
> then I'm not exactly serving millions of hits on my
personal site .
> YMMV, but be prepared for the possibility of a lot of
template tweaking
> if you want good performance.

Maybe I need to learn about kid. Webware mentions them too.
They have a 
kidkit whatever that might be.

> 
>   Please let us know what you end up choosing, and what
you think of it
> a few months down the track 

Certainly. However, I would love to hear about any
comparisons anyone 
has seen around.

Google turned this up for me ...

http://mail.python.org/pipermail/baypiggies
/2005-December/000051.html

http://www.blueskyo
nmars.com/2005/09/21/turbogears-on-day-4-and-some-jeremy-jon
es-commentary/

The bottom line of most things I've seen on the web is that
TurboGears 
and Ruby-on-Rails are the "happening" things.
I'm more interested in 
choosing a technology in which to invest my limited
time/brainspace.

Thanks Ryan

Mike

> 
> 
>   Cheers,
> 
>      Ryan
> 
> 
> 
> 
> 
>
------------------------------------------------------------
------------
> 
> _______________________________________________
> melbourne-pug mailing list
> melbourne-pugpython.org
> http://mail.python.org/mailman/listinfo/melbourne-pug

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug

>>This Melbourne Python Users' Bulletin (or letters)
may be a collection 
>>of published case studies or personal experiences
and tutorial-like 
>>articles. From an academic perspective, I believe
that at least the 
>>technical contents needs to be correct before
"publishing", so even 
>>though it is not a real academic journal, some forms
of "technical 
>>peer-reviewing" needs to take place.
>>    
>>
>
>Are you suggesting a mail-list review?
>  
>
I am not familiar with the dynamics of mail-list review. Do
you mean 
that the articles or codes (in the form of letters) received
by the 
Editor be sent to this list for group reviewing?

>Just thinking out loud, I reckon peer review is
brilliant in terms of 
>you guys telling me where I'm going wrong when I show
you what I'm 
>doing. That is because my objective (not necessarily
anyone else's) is 
>to learn and improve my skills in a much more immediate
time frame than 
>would be possible within any formal or academic peer
review situation.
>  
>
In academic settings, usually the real review takes 2 hours
but the 
paper can take weeks collecting dust on the reviewer's
table.

Yes, a group review or mail-list review (as I understood it)
will be 
beneficial for learning. But I am also concerned about time
needed for a 
fruitful debate in this list. If we are going down this
path, can I 
suggest that the Editor makes the final remarks in 7
calendar days from 
1st post to this list? If anyone had missed out on the
reviewing but 
strongly feels about the finalized article, please submit
another article.

I feel that I have to declare myself here. I vote strongly
in favour of 
technical peer-review because I arise from the academia. For
me, the 
goal is to produce peer-reviewed publications, which
translates to more 
research funds (buying better books). Even if we use a
mail-list 
peer-review model, we just state precisely how we work, and
with proper 
front matter to the e-zine (listing the reviewers and with
ISSN), it is 
usually accepted by the university as peer-reviewed
articles.

Such publications will matter for undergraduates trying to
get scholarships.

On the research fund side, let me illustrate an example. The
Australian 
Undergraduate Students' Computing Conference (AUSCC)
started in 2003 in 
Melbourne University, and although we did not reject any
papers 
(revisions yes, there wasn't any papers that is truely
unreadable or 
technically flawed to be rejected) but with proper front
matter with 
reviewers' names and an ISBN, it was accepted as a
peer-reviewed 
publication and my department received 350 for the paper I
published in 
that conference then.

>There is definitely a niche for polished peer reviewed
stuff. Just last 
>night I placed an order on Amazon for O'Reilly's
Python Cookbook. The 
>cred of the reviewers looks impeccable to me and that is
what persuaded 
>me to spend hard-earned money. Even so, I also visit the
ASPN website 
>for recipes because I see peer review debate and
commentary there as well.
>  
>
By peer reviewing, I least I know those codes work as
intended (and did 
not mix up standard deviation with variance).

>I think I'm saying we sacrifice bandwidth in the
knowledge transfer game 
>when we add formality. IMHO, absolute best and broadest
knowledge 
>transfer bandwidth is face-to-face discussion with
whiteboards on hand. 
>Next best is either Professor Google or a mail list like
this. The 
>narrowest knowledge transfer bandwidth is to buy and
read a book. I'm 
>not saying anything about the quality of the knowledge -
just the speed 
>and relevance at the time.
>  
>
I agree.

>A peer reviewed e-zine would be great. The peer review
would be better 
>provided people got in and criticised with gusto. Maybe
Crocker's Rules 
>should apply (:
>
>  
>
I support with gusto that peer-reviewed e-zine will be great
but I do 
feel that Crocker's Rules do breed arrogance to a certain
extend.

 From some of the rejections I get for my paper submissions,
it is 
obvious that the reviewers do not understand my work at all.
And I got 
rejected on the basis of my paper was poorly written (no
explanations) 
and my maths symbols are not nice!!!

>>I've also just did an initial proposal to Firebird
Foundation about 
>>starting a peer-reviewed journal. (Let's talk about
this off-list if you 
>>are interested) 
>>    
>>
>
>Don't be shy - I would be more comfortable being
included (or lurking) 
>in the discussion.
>
>Run with it ...
>  
>
I think it will be far to off-topic to discuss starting a
peer-reviewed 
Firebird Journal in a Python list. 

Cheers
Maurice

_______________________________________________
melbourne-pug mailing list
melbourne-pugpython.org
http://mail.python.org/mailman/listinfo/melbourne-pug