>>> json.dumps(datetime.datetime.now())
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.6/json/__init__.py", line 230, in dumps
    return _default_encoder.encode(obj)
  File "/usr/lib/python2.6/json/encoder.py", line 367, in encode
    chunks = list(self.iterencode(o))
  File "/usr/lib/python2.6/json/encoder.py", line 317, in _iterencode
    for chunk in self._iterencode_default(o, markers):
  File "/usr/lib/python2.6/json/encoder.py", line 323, in _iterencode_default
    newobj = self.default(o)
  File "/usr/lib/python2.6/json/encoder.py", line 344, in default
    raise TypeError(repr(o) + " is not JSON serializable")
TypeError: datetime.datetime(2011, 12, 1, 0, 50, 53, 152215) is not JSON serializable

So this means we need to figure out a work around. The trick is to let the json module know what it should do with a date/time object while leaving the rest of it in place. So, no replacing the default handler.

What we need to do is subclass the JSONEncoder and override the default method

class JSONEncoder(json.JSONEncoder):
    def default(self, obj):
        if hasattr(obj, 'isoformat'): #handles both date and datetime objects
            return obj.isoformat()
        else:
            return json.JSONEncoder.default(self, obj)

Using the hasattr and looking for ‘isoformat’ method will allow this to handle both date objects and datetime objects. So all that is left to demonstrate is how to put it together with the json.dumps method.

>>> json.dumps(datetime.datetime.now(), cls=JSONEncoder)
'"2011-12-01T00:58:34.479929"'

Ok, so now you have this ISO formatted string containing date information, how do you get that converted to a javascript date object once it has been transmitted across the abyss and now resides in a javscript callback function?

var d = new Date("2011-12-01T00:58:34.479929");

Happy Data Slinging!

import android  # The SL4A android.py module should be on your sys.path.

ip = '192.168.x.xxxx'
port = 35766
droid = android.Android((ip, port))

What happens is the __getattr__ method of the Android object uses magic to change droid.getenvironment() into an RPC call to the public server and then return the result back as a named tuple. Nice. Being the nosey bugger that I tend to be, I modified the code to add a debug param to the __init__ method, that when set, would print out what was being sent out over RPC and then the raw tuple results. A snippet of the modification goes like this:

  def __getattr__(self, name):
    def rpc_call(*args):
      if self._debug:
          print "droid.%s%s" % (name, str(args))
      res = self._rpc(name, *args)
      if self._debug:
          print "\t%s" % str(res)
      return res
    return rpc_call

You can easily see where I put my “if self._debug” logic in place. Now if I use my modified android.py I can turn on the debug flag and get some 411 on the magic that is going on. It ends up looking like this:

droid.eventWait(3000,)
	Result(id=28, result=None, error=None)
droid.eventWait(3000,)
	Result(id=29, result={u'data': u'@end', u'name': u'dmt:fromClient.speak', u'time': 1322542655069000L}, error=None)
droid.wakeLockRelease()
	Result(id=30, result=None, error=None)

html/js -> python

and from

python -> html/js

Part of this requires that I knock out and hack some html/js code. However, I am using the Remote method with a public server on my android device, since I am too lazy to set up eclipse and a full blown android dev env. The example code they show, uses an html file located on the sdcard of the device. Of course this brings its own problems, since now I have to mount, edit, unmount between each hack cycle. Ick. Well I could just tell it to load the html from an off device server, but being lazy ( I think I mentioned that already. ) I didn’t want to rsync back and forth to my remote server, set up directories, etc. Also, I didn’t want to set up a Django install just to serve a hacky html script.

So I think to myself, man there has to be some light-weight way to serve this up locally while I’m hacking. So I think, hey CherryPy, but then I remember Edna and it hits me, I can serve static pages out of a directory with just python. A little google-fu and this page appears, giving just the needed incantation.

 python -m SimpleHTTPServer 8080

Which happily serves everything out of the directories below it on the specified port (8080 in this case). I make a little adjustment in to my webViewShow and change it from file:/// to http://my-dev-ip:8080/thefile.html and all is good with the world. As I hack, changes to the html are pulled and served immediately.

Did I mention, Python Rocks!

Jason’s reasons for netbook’s demise, underpowered for Windows and underoutfitted for Windows, while true to a degree, are not going to be the reasons for any so called demise of the netbook market.  What he is really talking about is a failure of Windows, both XP and Vista.  Netbooks are not small laptops just as miniature golf is not golf in miniature.  They are a different class of device, bigger than a phone but smaller than a laptop.  Trying to run windows on a netbook is as frustrating as a phone with windows mobile.  Microsoft is becoming less relevant in the fastest growing market segment.

While I agree that what we now call netbooks (screen < 10 inches) may dissappear in three years it won’t be because of consumer apathy but because the device platform will become wildly popular and manufacturers will struggle to differentiate themselves.  Jason goes on to compare netbooks with tablet PCs but that analogy is false.  There have only been a handful of tablet PC platforms, where as the market is exploding with netbook devices.

His closing statement about price and weight is exactly what will sustain the market.  At a price of 1/3 to 1/4 of a full featured light weight laptop, more people can afford to have both a laptop and a netbook.  If they were to buy the MacBook Air or X360 they would have to live with the tradeoffs they bring.  A netbook/laptop combo comes in 1/4 less the cost.

So being an overly curious bugger, I downloaded the deb pkg for Ubuntu32. Yeah, it looks pretty rough — but what the hey, it’s a PoC right? So, I thought maybe a little speed test is called for, so I googled “javascript test” and my cloud brain returned, “Sunspider JavaScript Benchmark“.

So I fired it up in the cxChromium port and then it the standard FF3 on Ubuntu. Here is what I saw for cxChromium. I then swing over with FF3 and get these results.   Javascript in FF3 is 2.44x slower than cxChromium.  Man oh Man, is the v8 javascript engine a hummer.

Using an auto-increment field is not the way to go.  It is viable when you only have 1 database but the datastore for your app is going to/can be replicated out to other machines.  This would mean that their exists times, when datastore’ != datastore” — over time datastore’ would be sync’d with datastore” so that datastore’ == datastore”   — this would lead one to believe that there will be times when the idea of an auto-increment field will not be synchronizable or that the result of the synchronization would be less than satisfactory.  My belief that auto-increment fields are the wrong idea in this environment is strengthened by the fact that they are not offered as an intrinsic datatype.

The way to go, in my opinion, is to use UUIDs. (see links below)
  http://docs.python.org/lib/module-uuid.html
  http://www.faqs.org/rfcs/rfc4122.html

Other Thoughts on the topic:

• data access is very expensive, using a UUID should be faster
• UUID1 or UUID4 would be the types to consider
• UUID1 is preferable as it would introduce some machine significance which should make the chances for a collision to be even more remote than for a UUID4 (random)

#!/usr/bin/env python
"""implement a greedy coin changer, returning the
fewest coins to make the change requested."""
#coin_list can be expanded to include silver dollars
# and 50 cent pieces by just expanding the coin list
# to [100,50,25,10,5,1] the reulting answer
#structure will modify itself to reflect 

coin_list = [25,10,5,1]
change_requested = .71
remaining = change_requested * 100
change_returned = []    #result structure

for coin in coin_list:
    num_coins,remaining =  divmod(remaining,coin)
    change_returned.append(int(num_coins))

print change_returned
print remaining

The benefits of this version, are no conditional logic is needed, the coin structure can be modified and the answer will modify itself accordingly.

I think it is a great opportunity on a smaller scale than anyone might imagine. Sure, this could serve as the platform for the next YouTube type social-2.x site, but what I think this really means, is that Google is rounding out the Google Apps for Domains by giving the ability to create something more than a brochure-ware style site offered by their current Sites for Google Apps.

Many are looking for Google to use this as an opportunity to expand advertising revenue, and that is certainly possible for widely popular webX.x sites but what they really needed is another tool/knife to hold to the competition’s throats. Looking at the tea leaves in the bottom of my glass, I see something more akin to a SharePoint attack; Going after the S in SMB market.

App Engine allows for authenticating users via Google system, how much longer until we can interact with other Google services in a similar fashion?? Calendaring, GTalk, etc — I’m not talking mashups, something much more refined.

Interesting things/thoughts happen when your programmers are sys-admins too.  This idea of translating the framework we use for testing code to testing systems has a number of interesting dimensions to it.  Just like Alton Brown, I insist that my tools multi-task too.

The gist of the article is that when developing APIs are never as complete as we want them when we are developing a new system and if you are using a static language you’ve got lots of ramifications to consider and code to rework when you have to expand an API.  However, dynamic languages have a real advantage here and they go on to give a very real example and how they dealt with it.