Thursday, January 13, 2005

Turn any iPod into an iPod Shuffle in 3 easy steps! on Flickr - Photo Sharing!

Friday, January 07, 2005

Text messages aid disaster recovery

BBC NEWS | Technology | Text messages aid disaster recovery

SMS is one of the most robust telecommunication media available and the authors in the above article got that part right. The challenges faced by the 'Alert Retrieval Cache' concept, as it was loosely defined, is sketched out below, and the recommended solutions are attached as well.

1: Objective of usage: The technology solves what problem? As far as the article outlined, the problem seems to be an inability for needs to be quickly communicated to "needproviders". Needs are identified by "needspotters" (sorry...is there a better term?), and have to be sent to "needproviders".

2: Challenge: How are the needs qualified? What is a valid need and what isn't? Does the system take all needs, from "need an airplane ticket" to "need antibacterial ointment"? Are needs deemed qualified through the qualification of the identity of the needspotter? As with all news data, of which need news is one, there is the primary issue of authentication and qualification. Is it real, and is it relevant? There needs to be some form of filtering involved, though not necessarily just one. Filtering can take place at multiple points, the outcome, or end product of which, is a verified and qualified need.

3: Challenge: How to optimally match need against needprovider? A need for antibacterial ointment in Sri Lanka may not be optimally met by a provider of ointment in Aceh, Indonesia. There needs to be a qualification of the matching process based on relevant criteria such as timeliness, logistics of meeting needs, availability, willingness, among others. This must be overlaid, in a disaster scenario, by the fact that optimization is a preferred though not required attribute, as "something worth doing is worth doing poorly" (G.K. Chesterton), and any help is better than none.

To keep this brief, I hope to start this discussion of a more optimal approach with suggestions on some attributes of the system that should be implemented.

1: Relevance of need messages. There should be some layers of scoring provided to the need message in the following ways: 1-Granted by system. This would be the case of preregistering the mobile numbers of the team members at Medicin Sans Frontiers Aceh field operations, such that any sms's from those numbers get automatically tagged for the relevant layer of authority plus relevance. 2-Qualified by popularity. This would not involve authorizing individual messages, but rather the nature of the message must be extracted and compared with other messages to see if there is popular citing of this need. This would involve text parsing ability and the existence of some intelligence on the server side of the messaging system. This is eminently available in existing technology. 3-Intelligent tagging. Where did the message come from? Any location data from the service provider by cell site? If possible, save room in the precious 160 character message length and obviate the need for sender to state location, by getting participation from message providers. This helps inform the need message with necessary data for addressing the need.

2: Navigation of need messages. Need messages have to be accessible and relevant. A spam of irrelevant need messages would not be useful in these circumstances. There are two primary ways to tag these messages for relevance. 1-provide key words in the messages; these keywords would be searchable and would provide guidance to need providers when scanning through and searching these messages. Structured text messages, with the structure format made available (through web, email or sms broadcasts), would make this easily possible when combined with the right server side components. 2-need messages can be submitted to subscription based distribution (email or otherwise), such that need providers get the right messages on a timely basis. This also can be done with structured text messages.

3: Matching. Need messages should not disappear off the face of the earth once it has been 'matched' against a provider. There should be a bulletin board type of information dissemination that tracks needs and matches. Both the needprovider and the needspotter are identified and can easily be held more accountable given the public nature of this bulletin board information. Similar to user ranking, these parties can also be evaluated based on the performance of their respective roles in the matches made.

Summary: This is a potentially very important system that can save people, and thus justifies some serious thinking going into the design thereof. The basic technology components are available and affordable to put such a system in place. Hopefully this can be done quickly and accurately.

Monday, January 03, 2005

The South-East Asia Earthquake and Tsunami Blog

The South-East Asia Earthquake and Tsunami Blog

This is a valuable use of technology. Shout out to the tech providers who do this free of charge to the end user.