Thursday, April 22, 2010

Discreet Piles of Data

Lately, I have been working in XML and AS3 and have seen how loading data dynamically can keep a website current and easily maintainable for clients. A simple admin shell goes a long way to helping people update their own sites. This begins to shift the point of view that text data can be separated into discreet piles; persistent data (different from static data), up-to-the-moment data, modular data, random data and others that I have yet to encounter. Persistent data is that information that can change but approximates the same type of information and probably in the same format, addresses of a semi-annual event for example. This is different from static data such as the address of a company that hosts the event. Up-to-the-moment data is just that, but can be any type of string of varying lengths and varying subject, format and style. Up-to-the-moment (uttm) requires the most flexible of parameters. Modular data has some format restrictions such as twitter's character count but can be used to string together conversation, narrative, or even poetic structure. Thoughts strung into an idea, like this blog for instance, stanzas in a poem are also a good example, but modular data can be reused in appropriate places with parameters that help it fit into the larger whole of the narrative. Random data may have some length constraints but that is its only physical restriction. Random data has a topical relationship to the site such as quotes of famous authors may be found on Amazon's site but not on Home Depot, yet these quotes may change out regularly.

These categories are my own construct to attempt to make some sense of their practical uses in programming or design terms but they also have some relationship to the site as a group. They do start to communicate the overall impression of a site if followed regularly, or they could give two different impressions by arbitrary viewers. We don't often consider our personal blogs as a set of text snippets that tell a complete story. These are the ramblings of a day in a life scenario. However, if we consider the aspect of brand and our incessant need to brand ourselves we should consider what the sampled view is at any given moment. The data categories I have described begin to talk about what we are, but also who we are, and the importance of texts as they describe our persona for us.

The relationships are complex not just because of their influence on each other but also the context of the site and who is viewing the data. Drunk pictures on a facebook page seen by ones peers seems more like a laugh than it would to an employer. Now consider the drunk pictures next to the text that reads "Just Married!" and a reserved image of the page owner with a nicely written, conservative job description which indicates that he just graduated a year before. The tag at the bottom reads "grandma made me get up and dance", maybe we will cut him some slack. Now take all that out of facebook and put it in a wedding site for the lovely couple, now put it in each of these context taints the viewer to perceive the image of the drunk in different ways. As we assemble these discreet piles and leave the sorting, formatting and proximity to algorithms we are obligated to assess these relationships and at least sample the brand we are attempting to build.