web.py ideas, anyone?

[Aaron Swartz]

I'm applying for web.py to participate in this year's Google Summer of
Code. As part of that, we're supposed to have a list of ideas that
we'd like Summer of Code hackers to work on. If you have some ideas or
are perhaps interested in implementing some of them, take a look at:

http://webpy.infogami.com/ideas

Feel free to edit it to add ideas of your own. (If you can't edit it,
send me an email off-list and I'll edit it for you or get you an
account.)

[Kai Hendry]

I am not too fond of the "javascript form support" as this sort of
functionality will be implemented by leading browsers soon-ish.
http://www.w3.org/TR/web-forms-2/
It is already implemented by Opera. Aaron, you should join the HTML WG. :)
http://ln.hixie.ch/?start=1173385976&count=1

user accounts++ gets my vote. with openid and all that jazz.

The web is begging for a decent IRC Web application.
As for the debating idea, I had the same idea! though convinceme.net
has come along and I usually find the Talk page on wikipedia good
enough.

i do like the stats functionality that infogami does. be good if that
was free software.
http://webpy.infogami.com/_special/stats/

Best wishes,

[rkm...@gmail.com]

500 - Internal Server Error
is what i get from webpy.infogami.com

[gregp...@gmail.com]

On Mar 11, 2:35 pm, "rkmr...@gmail.com" <rkmr...@gmail.com> wrote:
> 500 - Internal Server Error
> is what i get from webpy.infogami.com
>

Not really summer of code worthy, but I think we need a new wiki.
Sounds like it's having issues. And it seems important that new users
be able to edit things.

As for summer of code ideas:
1. A whole lot of field types for forms: email address, date, number,
money, etc. The trick is that they get free validation, both server
side and javascript built in to make sure they validate. And date
could have a javscript calandar drop down, etc. (of course this may
deviate from the simplicity philosophy of web.py)
2. All of the other ideas mentioned for form.py in this group
3. I second the user accounts feature
4. Make web.py work in Windows with IIS
5. Make web.py run really fast as a plain old CGI application (this
would make it useful on a whole lot of web hosts that don't have
fastcgi, etc)
6. Built in support for sending emails
7. LISP syntax for templating ... (not sure why ..)
8. Built 10, ready to install sample apps (eg wiki, blog, bug tracker,
etc), they'd have a 5 minute install deal, set up your DB for you and
everything. This would be good for marketing and promoting web.py by
getting some mainstream free software out there running web.py
8.a and of course switch web.py wiki over to said wiki
9. Super-massive-document-everything project!

Hmm, sorry I could only think of 9.
How to think of more ideas:
1. you could see if there are any features that other frameworks have
that would be good to implement in web.py.
2. try to think of pains you guys have developing web app.s and see if
web.py could address it.

-Greg

[rkm...@gmail.com]

On 3/11/07, Kai Hendry <kai.h...@gmail.com> wrote:
>
> i do like the stats functionality that infogami does. be good if that
> was free software.
> http://webpy.infogami.com/_special/stats/

How do I get statistics like this? Is this some part of web.py?
thanks

[Adam Atlas]

Damn, looks like you have to be 18 to participate in Summer of Code.
Still, I'm always happy to contribute!

[Adam Atlas]

On Mar 11, 2:20 pm, "Kai Hendry" <kai.hen...@gmail.com> wrote:
> I am not too fond of the "javascript form support" as this sort of

> functionality will be implemented by leading browsers soon-ish.http://www.w3.org/TR/web-forms-2/

> It is already implemented by Opera.

Don't count on Internet Explorer adopting it "soon-ish". (Microsoft
has never been a good early adopter of useful new open technologies.
Hell, they'll probably come up with some proprietary ActiveX thing
that does this.) And if IE doesn't support it, then we may as well not
bother.

> i do like the stats functionality that infogami does. be good if that
> was free software.http://webpy.infogami.com/_special/stats/

I do similar graphing with Matplotlib (http://
matplotlib.sourceforge.net/). Its API leaves a lot to be desired (it
tries to emulate Matlab too closely -- ugly and non-Pythonic), but it
gets the job done.

[Kai Hendry]

On 3/12/07, Adam Atlas <ad...@atlas.st> wrote:
> On Mar 11, 2:20 pm, "Kai Hendry" <kai.hen...@gmail.com> wrote:
> > I am not too fond of the "javascript form support" as this sort of
> > functionality will be implemented by leading browsers soon-ish.http://www.w3.org/TR/web-forms-2/
> > It is already implemented by Opera.
> Don't count on Internet Explorer adopting it "soon-ish". (Microsoft
> has never been a good early adopter of useful new open technologies.
> Hell, they'll probably come up with some proprietary ActiveX thing
> that does this.) And if IE doesn't support it, then we may as well not
> bother.

A generic piece of JS with provide IE legacy support for Web Forms 2.

Someone is already working on that.

[Aaron Swartz]

The Web Forms stuff looks awesome. It's looked awesome since 2003.
I'll believe it when I see it.

In any event, there's no inconsistency between it and implementing the
functions in web.py. form.py can simply write Web Forms 2-compatible
HTML.

[Tom Berger]

On 3/11/07, gregp...@gmail.com <gregp...@gmail.com> wrote:

Not really summer of code worthy, but I think we need a new wiki.

I think it's definitely worthy of some effort, within or without SOC - web.py's presentation on the web is not as good as it could be, and I'm worried that it's hurting adoption. I know I've had some difficulty 'selling' web.py to clients and collaborators because the initial impression they got from the website. It's sad (that marketing plays such an important role in the adoption of technology) but it's fact, and it's easy to fix.

Tom



b.t.w I never got around to announce it here, but http://travel.dk.com/ has been online for quite some time now, and is the most elaborate website we have constructed using web.py so far. It's basically those Eye Witness guides online, with some web 2.0 goodness added, and a print-on-demand service. Go check it out!



--
http://intellectronica.net/

[Srinivasan R]

On 3/11/07, Kai Hendry <kai.h...@gmail.com> wrote:
>

> i do like the stats functionality that infogami does. be good if that
> was free software.
> http://webpy.infogami.com/_special/stats/
>

I used to like the stats page in infogami.
But, the all time page views/user per day is still showing July.
Is it working?

--
Next mail in %d minutes
cnu - http://cnu.livejournal.com
http://fslog.com - Free Software Blog

[Jeremy]

On Mar 11, 5:36 pm, "Aaron Swartz" <m...@aaronsw.com> wrote:
> I'm applying for web.py to participate in this year's Google Summer of
> Code. As part of that, we're supposed to have a list of ideas that
> we'd like Summer of Code hackers to work on. If you have some ideas or
> are perhaps interested in implementing some of them, take a look at:
>
> http://webpy.infogami.com/ideas

Web unit testing.

This is a really faffy and ultimately quite complex area to get right,
but sorely needed for production websites where proper regression
testing is crucial. If only it could made as easy as writing webpy
code...

Features required:
- Set up a privately accessible server with database settings (if
relevant) to a given test database.
- Resetting Database, and any other values (local files, etc) to a
'base' state.
- Short functions to access a particular URL, with POST data as
required, and then to check the result either against given regexps or
similar if data is provided, or error messages/other header values.
- Very simple syntax to test the same form entries with many different
values (eg. Testing for null values, existing values in the database,
invalid syntax of values, quotes etc)
- Logging of results - reporting of errors in a manner similar to
unittest
- Checking of final data state against an existing value - although
options to ignore particular columns (eg. Auto-incrementing IDs, date
added etc).

To make this a useful generic python tool, it should talk WSGI to the
server (which could be quite fast, possible on a simple development
machine, and all in a single process), although some may want HTTP
support too for completeness (in which case a host and base URL would
be provided).

I'm likely to be doing some of this work anyway over the next couple
of months, but it would probably end up as something quite application-
specific. I can't believe it wouldn't be required more widely...

-jeremy

[Aaron Swartz]

> I think it's definitely worthy of some effort, within or without SOC -
> web.py's presentation on the web is not as good as it could be, and I'm
> worried that it's hurting adoption. I know I've had some difficulty
> 'selling' web.py to clients and collaborators because the initial impression
> they got from the website. It's sad (that marketing plays such an important
> role in the adoption of technology) but it's fact, and it's easy to fix.

That's really good to know. I'm not sure it's as easy to fix as you
suggest -- I'll probably have to hire a designer or something -- but
it's certainly worth fixing.

I can't promise it will happen right away, but things are definitely
in the works.

BTW, that DK site looks awesome. If you want to lend me your designer... :-)

[Tom Berger]

On 3/12/07, Aaron Swartz <m...@aaronsw.com> wrote:

> web.py's presentation on the web is not as good as it could be

 

I can't promise it will happen right away, but things are definitely in the works.

Excellent. I'm no designer but if there's anything I can do to help I'll be glad to try.

BTW, that DK site looks awesome. If you want to lend me your designer... :-)

Thanks! We ( http://www.lshift.net/ ) only did the programming. The company that did the design is http://www.pokelondon.com/ - but if you actually plan to throw money at the problem I think there are cheaper and more hacking-oriented designers out there.

Tom

[blaf]

I'm willing to help too... have you references of things you like for
webpy?

On Mar 12, 5:51 pm, "Tom Berger" <tom.ber...@gmail.com> wrote:
> On 3/12/07, Aaron Swartz <m...@aaronsw.com> wrote:
>
>
>
> > > web.py's presentation on the web is not as good as it could be
>
> > I can't promise it will happen right away, but things are definitely in
> > the works.
>
> Excellent. I'm no designer but if there's anything I can do to help I'll be
> glad to try.
>
> BTW, that DK site looks awesome. If you want to lend me your designer... :-)
>

> Thanks! We (http://www.lshift.net/) only did the programming. The company
> that did the design ishttp://www.pokelondon.com/- but if you actually plan

[Kai Hendry]

On 3/12/07, Jeremy <jeremy...@gmail.com> wrote:
> Web unit testing.

> I'm likely to be doing some of this work anyway over the next couple
> of months, but it would probably end up as something quite application-
> specific. I can't believe it wouldn't be required more widely...

Have you seen twill? http://twill.idyll.org/

I am quite happy with docstrings and my little scripts to load up a test DB etc.

[Srinivasan R]

On 3/11/07, Aaron Swartz <m...@aaronsw.com> wrote:
>

Did web.py get selected to the list of mentoring organisations?
I dont see web.py in the list. Or are you planning to join with python group?

[Aaron Swartz]

> Did web.py get selected to the list of mentoring organisations?
> I dont see web.py in the list. Or are you planning to join with python group?

web.py doesn't seem to have been accepted on its own, but I am hoping
to mentor web.py projects under the PSF application.

[Anand]

On 16-Mar-07, at 12:44 AM, Srinivasan R wrote:

> Did web.py get selected to the list of mentoring organisations?
> I dont see web.py in the list. Or are you planning to join with
> python group?

Already joined the python group.
see http://wiki.python.org/moin/SummerOfCode/Mentors

[Fili...@gmail.com]

I'm interested in the idea of "javascript form support".
and I'm want to attend GSoc 2007 as a student, and work on this idea.

The core problem of this project is the javascript validation in
client side and how to generate the javascript code from the server-
side.

There are two ways to generate the javascript validation from server-
side.

1. written an javascript validation framework alone,
(an good example is http://tetlaw.id.au/view/blog/really-easy-field-validation-with-prototype).
In the server side we just add class_name hook(on customer
attribute for input element), and start up scripts for form
validation.
2. generate the javascript code totally from the server
side,including validation rule, error message.etc.
The first one is good, but it has some disadvantages.
* you have to written the validation rule in javascript yourself
and that force python coder to master javascript.
* It is not convenient to pass arguments from server side to
client side, for example length-validation 's min-length or max-length

So my idea is that, in the client side We write an tiny framework,
only including :
add event listener for element need to be validated.
define some possible action for validation-failed /
validation-success / etc... (for example alert / show msg..)

And in the server-side, we generate the validation rule code,
define the action for each validation.
In this way in the server-side we only generate the necessay code
or dynmatic code.

My English is poor, I hope I have explained my ideas clearly.
I will post a more detailed description of my idea in my Blog.
My Blog : http://blog.filia.cn
or My Summer Of Code Blog : http://filiasoc.blogspot.com/

Please give me some advice for my idea.

Thanks.

---