Issue with school and Google Apps

[Andrew Harrington]

I was just at De La Salle High School, a Catholic HS in Chicago, where I am due to teach Python/math to an algebra II class about 2 hours a week for about 4-6 weeks.

I picked up along the way, sort of by accident, that their school would not be able to use PyKata on Google appspot at school.  All students have a tablet computer in school that they take home.  At school they filter out all Internet sites except for an explicit list of allowed ones.  Unfortunately for them, they have no way to partition off parts of Google. (I did not get the details of their filter to see why.  I assume they filter sites, not URL's).  It is all or nothing, and they opt for nothing.  If they were to use PyKata, it would need to be a separate installation (could be on a much lower bandwidth local server).  Jeff, with a HS teacher's perspective, had already mentioned such an issue.  This only reinforces the idea that we want to graciously accept the free bandwidth, scalability, and storage of Google, but I do not want our software and data to be tied to them, or open source does not mean much!  Hence it is a big deal to be able to download/upload BigTable data in a usable form and to have an interface layer that can be swapped to work so the whole site works either with BigTable or not.

--
Andrew N. Harrington
 Director of Academic Programs
 Computer Science Department
 Loyola University Chicago
 512B Lewis Towers (office)
 Snail mail to Lewis Towers 416
 820 North Michigan Avenue
 Chicago, Illinois 60611

http://www.cs.luc.edu/~anh
Phone: 312-915-7982
Fax:    312-915-7998
g...@cs.luc.edu for graduate administration
u...@cs.luc.edu for undergrad administration
aha...@luc.edu as professor

[Andre Roberge]

On Mon, Mar 29, 2010 at 6:26 PM, Andrew Harrington <aha...@luc.edu> wrote:

I was just at De La Salle High School, a Catholic HS in Chicago, where I am due to teach Python/math to an algebra II class about 2 hours a week for about 4-6 weeks.

I picked up along the way, sort of by accident, that their school would not be able to use PyKata on Google appspot at school.  All students have a tablet computer in school that they take home.  At school they filter out all Internet sites except for an explicit list of allowed ones.  Unfortunately for them, they have no way to partition off parts of Google. (I did not get the details of their filter to see why.  I assume they filter sites, not URL's).  It is all or nothing, and they opt for nothing.  If they were to use PyKata, it would need to be a separate installation (could be on a much lower bandwidth local server).  Jeff, with a HS teacher's perspective, had already mentioned such an issue.  This only reinforces the idea that we want to graciously accept the free bandwidth, scalability, and storage of Google, but I do not want our software and data to be tied to them, or open source does not mean much!  Hence it is a big deal to be able to download/upload BigTable data in a usable form and to have an interface layer that can be swapped to work so the whole site works either with BigTable or not.

I want to second strongly this last point.  In fact, this is one of the reasons why I abandoned my efforts to port Crunchy to Google app engine. 

If the data is available in a "simple" form, I could possibly write up a Crunchy plugin and have Crunchy available as a local alternative to deliver the content. 

André
 

--
Andrew N. Harrington
 Director of Academic Programs
 Computer Science Department
 Loyola University Chicago
 512B Lewis Towers (office)
 Snail mail to Lewis Towers 416
 820 North Michigan Avenue
 Chicago, Illinois 60611

http://www.cs.luc.edu/~anh
Phone: 312-915-7982
Fax:    312-915-7998
g...@cs.luc.edu for graduate administration
u...@cs.luc.edu for undergrad administration
aha...@luc.edu as professor

--
You received this message because you are subscribed to the Google Groups "PyWhip" group.
To post to this group, send email to pyw...@googlegroups.com.
To unsubscribe from this group, send email to pywhip+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/pywhip?hl=en.

[David MacQuigg]

If it's just a problem that everything is blocked by default, moving to
another site won't help. Catholic schools need to add our current site
(pykata.org) to their whitelist. I assume they have some procedure
making it easy for one of their teachers to add a new site to the
allowed list.

Simple commands for importing and exporting data should be fairly easy
to implement. Re-writing the program to avoid the Google API may be
more of a challenge. We should keep a list of any calls to Google
services we are using, including user accounts and database access. A
larger problem will be getting thousands of users to create new accounts
at a new site.

I'm not seeing the problem with being tied to Google. If they do
something outrageous, like start charging fees for small projects, it
will offend thousands of people in the open-source community. We have
safety in numbers. Is there some other scenario you are worried about?

-- Dave

[Andrew Harrington]

The separate URL takes care of the problem I mentioned at DeLaSalle.  I forgot about that.  Good.  It does not take care of somebody who is really behind a firewall, and wants to copy the installation to a local server.  You may be right that it is not a first tier issue now if it is clear that it can be done reasonably.  In that case your idea of clearly documenting and maybe isolating dependencies would be enough for now from an operational standpoint.  I still would like somebody like Andre to be able to grab all our basic data (problems, tests, exposition pieces, links) and port it to some other local platform.  So the question of the possibility/ease of BigTable dumps remains.

Andy

--

You received this message because you are subscribed to the Google Groups "PyWhip" group.
To post to this group, send email to pyw...@googlegroups.com.
To unsubscribe from this group, send email to pywhip+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/pywhip?hl=en.

[Athar Hameed اطهر حمید‎]

We can have a simple facility for generating a nightly dump of all the problems uploaded to the site. The dump could be in XML or JSON or any other format for that matter. That way anyone can use the problems authored on PyKata for their own purposes. GAE has support for scheduled tasks  (http://code.google.com/appengine/docs/python/config/cron.html).

Another thing that I would like to mention is that the reason GAE was selected in the first place was because of its sandboxing. Since we are allowing arbitrary Python code to be run on our web service, sanboxing is a real winner for us. The other approach is of course what Andre has done with Crunchy.

Athar

[Abhishek Mishra]

Also the recently supported TaskQueues are useful when a single cron
job starts exceeding timeout.

Abhishek

On Mar 31, 3:16 am, Athar Hameed اطهر حمید <ath...@gmail.com> wrote:
> We can have a simple facility for generating a nightly dump of all the
> problems uploaded to the site. The dump could be in XML or JSON or any other
> format for that matter. That way anyone can use the problems authored on
> PyKata for their own purposes. GAE has support for scheduled tasks  (http://code.google.com/appengine/docs/python/config/cron.html).
>
> Another thing that I would like to mention is that the reason GAE was
> selected in the first place was because of its sandboxing. Since we are
> allowing arbitrary Python code to be run on our web service, sanboxing is a
> real winner for us. The other approach is of course what Andre has done with
> Crunchy.
>
> Athar
>
> On Tue, Mar 30, 2010 at 10:14 PM, Andrew Harrington

> <anharring...@gmail.com>wrote:

>
> > The separate URL takes care of the problem I mentioned at DeLaSalle.  I
> > forgot about that.  Good.  It does not take care of somebody who is really
> > behind a firewall, and wants to copy the installation to a local server.
> > You may be right that it is not a first tier issue now if it is clear that
> > it can be done reasonably.  In that case your idea of clearly documenting
> > and maybe isolating dependencies would be enough for now from an operational
> > standpoint.  I still would like somebody like Andre to be able to grab all
> > our basic data (problems, tests, exposition pieces, links) and port it to
> > some other local platform.  So the question of the possibility/ease of
> > BigTable dumps remains.
>
> > Andy
>

> >> pywhip+un...@googlegroups.com<pywhip%2Bunsu...@googlegroups.com>

> >> .
> >> For more options, visit this group at
> >>http://groups.google.com/group/pywhip?hl=en.
>
> > --
> > Andrew N. Harrington
> >  Director of Academic Programs
> >  Computer Science Department
> >  Loyola University Chicago
> >  512B Lewis Towers (office)
> >  Snail mail to Lewis Towers 416
> >  820 North Michigan Avenue
> >  Chicago, Illinois 60611
>

> >http://www.cs.luc.edu/~anh<http://www.cs.luc.edu/%7Eanh>

> > Phone: 312-915-7982
> > Fax:    312-915-7998
> > g...@cs.luc.edu for graduate administration
> > u...@cs.luc.edu for undergrad administration

> > ahar...@luc.edu as professor

>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "PyWhip" group.
> > To post to this group, send email to pyw...@googlegroups.com.
> > To unsubscribe from this group, send email to

> > pywhip+un...@googlegroups.com<pywhip%2Bunsu...@googlegroups.com>