Who is going back through prior search interfaces and looking at how things might be implemented differently now (given the local contexts and reasons that led to the dominant forms of search today may have shifted)?
Here is an image from “Sorting out searching” (1998)
musing to myself as I once again write “germane” instead of “relevant”1
What if it were otherwise? If documentation/LIS/IR used “germane”?
s/relevant/germane/ s/relevance/germaneness/ this is what OED might suggest. I imagine it may’ve encouraged more faceted search due to association with the ‘principle of germaneness’ (in legal/legislative realms)?
Given the above and stumbling through prior relevant/germane discussion of such a naming, I find I strongly disagree with Saracevic (2016)’s claim that “relevance will be relevance by any other name”.
Some of the manifestations are referred to by a number of different names, such as pertinent, useful, utility, germane, material, applicable, appropriate, and the like. No matter what, they all connote relevance manifestations, but denote slightly different relations. As already mentioned, relevance will be relevance by any other name. (p. 20) [emphasis added]
Words cannot be so easily wrested from other uses. Would we have such little public control over general-purpose web search engines if ‘utility’, with its connotations of subsidized and regulated monopoly, was the core term?
Or would ‘utility’ have been so dismissed in 1971? Could researchers and designers have focused more on utility rather than being “restricted” to the ‘retrieval system design’? (Nolin, 2009)
May ‘appropriateness’ as the core term have helped us see “fluidity” and “signal how we are interpellated by it” well beyond the system design (de Laet & Mol, 2000)?
Or ensure that the search algorithms were never ones of oppression, but from the start there were “committed and protracted investments in repairing knowledge stores to reflect and recenter all communities” (Noble, 2018).
Or pushed earlier recognition of questions re “real-world messiness” in search engines and their pursuit of “societal relevance” (Haider & Sundin, 2019, Sundin et al., 2021).
With ‘material’ we may have earlier gotten “closer to the fact of the metal”, better asked “who makes”, and seen that “Hardware, the underlying material stuff, turns out to be full of politics and negotiations rather than crisp ontological certainty.” (Brunton & Coleman, 2014)
Or even now better address “the processes that translate spatial-temporal constraints related to material access into orientations” to information and information engines (Robinson, 2009).
Or not forget how some systems, by design, are oriented to base material concerns: “the question of information access and retrieval [] fundamentally a technical problem of connecting some set of resources to a query within the system for advertising purposes” (Hoffmann, 2016).
For each, despite all the ink from academics, the “cognitive boundaries” of the “boundary concept” would have been different as the “original processing of a core concept has an unavoidable side effect of articulating boundaries.” (Nolin, 2009)
As Buckland (1996) writes: “it is germane to draw attention to what was not done, to the road not taken.” (p. 71) [emphasis in original]
Footnotes
an annoying habit re wielding the r-word only where I want my words to appear in my own searches re information retrieval, access, etc.↩︎
Hoffmann, A. L. (2016). Google books, libraries, and self-respect: Information justice beyond distributions. The Library Quarterly, 86(1), 76–92. https://doi.org/10.1086/684141[hoffmann2016google]
Saracevic, T. (2016). The notion of relevance in information science: Everybody knows what relevance is. But, what is it really? Synthesis Lectures on Information Concepts, Retrieval, and Services, 8(3), i–109. https://doi.org/10.2200/S00723ED1V01Y201607ICR050[saracevic2016notion]
Sundin, O., Lewandowski, D., & Haider, J. (2021). Whose relevance? Web search engines as multisided relevance machines. Journal of the Association for Information Science and Technology. https://doi.org/10.1002/asi.24570[sundin2021relevance]
I am wary of over-indulging in prompts and missing the impetus to search in the first place; attending to and cultivating spaces that reward curiosity, doubt, and ignorance; and the various ways to distribute evaluation of results/responses across people, tools, and time.
OK, finally! Here is my first attempt at CF (so slick!)—various roles, only one (overly simplistic) prompt. (I cannot figure out how to link the inspect node…) also the 3.5 and 4 different is so striking.
The thing I don’t understand about asking ChatGPT to pretend to be an expert or a certain persona is that it will never tell you if it actually has no idea what you’re talking about
@danielsgriffin via Twitter on Aug 27, 2023
This makes me wonder if folks have studied differences in ‘hallucination’ under these sorts of prompts (“You are”, “Pretend to be”, etc.). Can you prompt-in humility? Have people looked at this? Perhaps this is something @IanArawjo’s http://chainforge.ai could help analyze?
Are you searching for [aslkfja- stowerk;asndf] or [aslkfjastowerk;asndf]?
@danielsgriffin via Twitter on Oct 19, 2022
Searching for [aslkfja- stowerk;asndf] on Google, Bing, DuckDuckGo, Yandex, @Neeva, @YouSearchEngine, @mojeek
Query parsing choices made by Google make it the only search engine providing no reference to the use of the term (rendered as such) by Golebiewski & boyd (2018).
Currently only a small portion of this tweet is recorded here, see the thread at the link.
@danielsgriffin via Twitter on Feb 1, 2022
*Searching Temporally?*
[Disclaimer: This (the choices and effects of availability/affordance of temporal search operators) isn’t my research focus. But I’d love to read work on it!]
Bonus: @StackOverflow also has a nifty feature whereby pressing enter in an empty search box sends you to a Search page with a handy and clear link to “Advanced Search Tips” (though the screenshot re searching dates above is from a further link to their Help pages).
Ex.: searching entangled & enfolded How does Google Search account for its inflation of any top-ranked results’ later PageRank metric? How do we account for the authority imbued throughout the top results? For the Matthew Effect on the plutocratic web?
Screenshot collected from the linked-tweet.