Ideas RSS

a Space to save my musings in order for discussion?

Persisting of possible avenues of endeavour for them not to get lost in the ether.

Archive

Aug
7th
Thu
permalink
haha, MAINE back-country.  These dudes make Danville okes look tame!

haha, MAINE back-country.  These dudes make Danville okes look tame!

Aug
5th
Tue
permalink

Ramblings on LBS and what the iPhone could mean to it.

Location Based Services and the iPhone

It is widely recognized that LBS is likely the most significant determining factor in providing Geographically relevant information. Such information could in turn be supplied to a myriad of next generation services and could possibly set the stage for many exciting platforms to be explored.

Until the present day there are several factors which have and are prohibiting the extensive use of accurate and timely Location information.

Obstacles to Cellular Network based LBS:

1.) The Cellular Network (eg. Vodacom, Cell C, MTN) are by far the weakest link in the process as they charge exorbitant amounts for their locations (in the vicinity of 20c-40c per location) making it infeasible to perform the regular (large quantities) locations necessary. Network Operators are also very careful at not allowing location data to get into the wrong hands so they necessitate an elaborate consent process. A large barrier with the use of GSM locations (locations that are provided by a Cellular Network) is also the fact that not all Networks have the facilities to provide LBS.

2.) In order to gain access to a Cellular Network’s underlying services (LBS or SMS etc) one either needs to be a WASP (Wireless Application Services Provider) or alternatively go through a WASP. Such a WASP then in turn acts as a broker for the services you would wish to leverage. In South Africa there are very few WASPS that have gained the licenses or developed the infrastructure to supply LBS.

3.) Then of course this brings us to the fact that GSM Network based locations are inherently inaccurate and slow.

For some time now the solution or proverbial carrot dangled in front of would be LBS suppliers have been the promise that South-Africa would at some point be receiving AGPS (Assisted GPS) devices which would solve most / all the obstacles mentioned above. The benefit of AGPS is much more accurate and faster locations than even regular GPS as it uses the Cellular network to seed its location in stead of having to by itself determine the Satellite to use for it’s own computations.

Since “Nine Eleven” in the United States a legislation was passed that all GSM service providers had to be able to at a given time be able to supply accurate location information for their subscribers. This then alluded to the possibility of more devices coming into the market with GPS abilities. This however in a large extent failed to materialize in the South African market apart from a few devices which included but is not limited to the Nokia N95.

The benefit of devices with the ability to provide location information autonomously then naturally open up the playing field significantly. However in order to leverage such device bound GPS functionality has traditionally been supplied by a Java library named JSR179 or alternatively it’s Symbian counterpart. Naturally the problem with this has essentially been the unknown distribution of these devices and therefore it has never been worthwhile to pursue.

However with all the hype surrounding the iPhone it seems as if the stage has been set for device based LBS services to come into their own right. In order to access the GPS functionality on the iPhone it would be necessary to develop a native (Objective C) client. Such a client would additionally be able to leverage other functionality such as the file system (Photos) as well as the accelerometer to determine for example whether the device is standing still or moving.

The killer Application

If one would develop an iPhone application and make it available through the syndicated Apple Store online one could unlock an international market with possibilities to integrate into several other services or platforms.

A hypothetical iPhone LBS application.

A native iPhone application which records the phone’s location during the time this application is running. It could in turn record Latitude & Longitude pairs with the time as well as whether the device was moving at the time or not. Another benefit of this would be that the device would be able to continue recording it’s own location even when there is no Cellular coverage. After such a recorded excursion this device would allow the user to sync up with a web service from where other information can be completed such as commentary on waypoints marked on the route etc.

This could then in turn facilitate a true “brick and mortar life feed” to be generated of the user of the iPhone. For example this feed could perhaps be mashed up with the user’s credit card transactions or proximity to their affiliated friends. This would lend itself to some very interesting information on a user and their own social graph in a geographical sense?

Another possibility would be to continuously update the individuals location held by Yahoo’s Firebird which in turn would act as a location broker to whomever would want to leverage that information further.

Further avenues could also be pursued such as Geo-caching where individuals go on treasure hunts using clues and their GPS skills to find a hidden box (see. http://www.geocaching.com)

As I am sure you can imagine the possibilities are limitless or at least only limited by ones own creativity.

Aug
1st
Fri
permalink
Jul
16th
Wed
permalink

Ideas on making twitter scale better

Given the wild success of what is reported to have been a small side project.  I guess that scale issues could have been anticipated.

Now naturally ensue the debate of how it is to be remedied, where it seems like everyone is an expert?

However this seems to me like the most relevant summary of the situation and options that could prevail:

http://highscalability.com/scaling-twitter-making-twitter-10000-percent-faster

permalink
Despite copious evidence that normalization rarely scales, I find that many software engineers will zealously hold on to total database normalization on principle alone, long after it has ceased to make sense.
Jul
4th
Fri
permalink

crimebat - sms service for Cape Metrorail Passengers

What are the thoughts around this: A sms service which is available publicly and targeted towards Cape Metrorail passengers.

What this service will initially allow is that anyone can sms a short crime report to a nationally registered shortcode (eg. 27911).

An example of such a message might be: “knife holdup Saltriver, White male 5’7” red beany leather jacket”.  This will then allow anyone checking the WAP’ish site to be aware of current criminal activity in the reported vicinity.

Maybe the weight/validity of such reports can be corroborated by multiple submissions?  The data will also be aggregated and published via a RSS feed which can be subscribed to / used by the Metrorail Security officers or the Cops?

mmm, food for thought but in my opinion perhaps a service worthy of building.  It really shouldn’t be that hard to accomplish either.

permalink
Not shaken an definately not stirred

Not shaken an definately not stirred