Dapper and the fragile web
A colleague put me on to Dapper earlier today (I must have been sleeping when it was all the buzz). It's kinda nifty. Dapper is basically a tool that allows you to create various kinds of dynamic content from existing web pages. You can take any page, define fields of interest, and Dapper will then allow you to produce RSS feeds, alerts, Google Widgets, and generally integrate with all sorts of nifty web 2.0 services.
Dapper allows you to create a "black box" (a Dapp) for any data source on the Internet. The Dapp produces XML which you can use programatically in whatever way you like. Additionally, Dapper provides you with the ability to transform this XML into other formats (e.g.: HTML, Google Maps, RSS, more).
It's designed to be usable by the novice, and to enable the benefits of dynamic content from sites that otherwise are fairly closed. I feel a great tremor in the force.
My issue: It's so very very fragile! As far as I can tell Web 2.0 is screen scraping, there's no way around it. It creates brittle interfaces on top of brittle interfaces. Unless the interface is intended to be used in this way, and you have some sort of service contract, you can't rely on this sort of hack.
I hear lots about of this type of service being promulgated as a solution to so many problems (I hear this particularly in the library world). Scrape this and mashup that. Sure, it's fun, and it may even work, but it's not a sustainable solution. Some guy dots an i on some website somewhere, and kablooie... SOA and well formed (web) services are definitely the way to go for anything serious.
Still, it's a blast to play with. I've always wanted an RSS feed of the CISTI Press releases... so I made one (use with caution... no guarantees as to it's reliability -- see above). I also created one to search the CISTI Library Catalogue and produce nice XML results... but until I can figure a way to provide appropriate paging functionality I didn't dare release it for fear that someone would actually use it.
Comments