On search and recall
Traditionally, the Web Search space has been needed to fulfil the role of an index for the World Wide Web. That an index was deliberately left out of the design of web was no doubt pragmatism more than oversight. Nevertheless this design decision left the door open for an innovative company to come along a decade later and monopolise a large part of the web's centre ground.
Been there, seen that
Some of the prerequisites of Search are:
Centralisation of indexes.
Documents are in natural language silos, frequented by native-speaking users.
Willingness of the user to inform the SE of their interest – albeit anonymously.
Users' ability to find effective key-words for a particular search on their favourite SE.
Effectiveness of the ranking algorithm
It's alright, we know what you've seen
Web Search implies finding new things, and Search Engines are certainly used for this purpose. Search is also used to recall information that has already been observed by the user. The category that is Search does not necessarily distinguish between these two use-cases. And, in a truly anonymous search, the Search Engine would be unable to distinguish them anyway.
Yet there does exist a field more specialised than Search, typified by bookmarking services like Delicious, where the objective is to allow users to express an intent to retrieve web content at a later time. Perhaps also to annotate the content in order to facilitate subsequent indexed retrieval by the user and/or some other persons or processes.
I got that already, thanks
Inclusion of Recall in Search is a convenience, yet in principle it need not be necessary.
If the user were to have a reliable, accessible, personal index of their previously viewed content, then a significant part of what is now Search could be reclassified as Recall and conducted stand-alone by the user.
The user's personal repository might be either off-line or on-line. Either way, there need be no third-party present. A party whose interest might lie in mining users' data for its own commercial interest or even for more nefarious purposes. Such a third party's activities might sometimes be aligned to a users' own interests, but just as commonly these activities might be against the interest of the user.
Tools such as X1 Professional or Google Desktop permit a user to index their desktop, or a messaging service off-line. However, indexes created are proprietary. Further, even where two users have the same vendor's product, the data in their indexes may not be (partially) integrated. So, while they may be useful stand-alone, they are not good for collaboration.
On the other hand, on-line URL tagging sites such as Delicious permit collaboration, but these are subject to centralisation and associated privacy concerns.
What is it worth to you ?
Having the option of a personal Recall engine, as opposed to a Search engine for some types of query, could put users in a stronger negotiating position when it comes to their attempting to gain access to information held exclusively by Search services. In effect, it could help level the playing field in information exchange.
Whereas a market is now tacitly conducted through apples-to-oranges exchanges, such as free software services in exchange for (use of) personal information, in future high quality information at the margin might be explicitly granted or withheld by the user, according to whether the SE is to be rewarded. If no reward, then an alternative service might be used for recall.
As information becomes ever more commoditised, through factors such as service ubiquity, intense competition, screen-scraping, network effects, third-party sharing and general information leakage, a premium is placed on burgeoning new sources of information over widely held data. Much of this may start out as personal information. Thus the trade between user and Search Engine may transition from being, say, software services in exchange for information towards quantity of information for quality of information. Much closer to a like-for-like exchange.
A significant obstacle to a true information economy has been a suitable currency. One property of a hypothetical currency is that it should be fungible. In other words, a token of information should be capable of being transferable, representing value (to a market participant), a store of that value, a means of reconciling the inherent heterogeneity of information.