Granular

Granular (pronounced gran-yuh-ler)

(1) Of the nature of granules; grainy.

(2) Composed of or bearing granules or grains.

(3) Showing a granulated structure.

(4) In computing, an object existing as a singular form at the level of the file system but which exists at the application level in multiple parts.

(5) Relating to or containing particles having a strong affinity for nuclear stains, as in certain bacteria.

1762 (although use not widespread until 1794): From the Late Latin granulum (granule, a little grain), diminutive of the Latin granum (grain, seed) from the primitive Indo-European gre-no- (grain) + -ar (from the From Latin -āris (of, near, pertaining to), the suffix appended to various words, often nouns, to make the adjectival form; added most often, but not exclusively, to words of Latin origin).  The word seems rather suddenly to have replaced the late fourteenth century granulous.  Granular, granularity, granule & granulation are nouns, granulate is a verb & adjective and granulatory is an adjective.

Terminology describing degrees of granularity

As granular has become a more widely used word, fastidious types have noted the increasing frequency of things being described as "more granular" or "less granular" and this elicits disapproval because it’s imprecise.  Something granular is composed of (usually small), discrete entities as opposed to being continuous and that’s a binary distinction, not a matter of degree so it’s inherently unclear if "more granular" and "less granular" indicate finer or coarser granularity.  For clarity, one should speak only of finer or coarser granularity.

Lindsay Lohan represented in granular art, an artificial intelligence (AI) generated artwork created by Wout from AI Fountain as part of the Curated Community Art initiative (CCAI) and finished in Adobe Photoshop.  Each digital artwork created by this algorithm is unique and made from a set of parameters; process and output are thus both inherently granular.

In computing, the concept of granularity exists in many forks and layers.  Users deal frequently with granular data, most typically when handling what appears to exist in many parts but which is, to the system, at least one layer, a single object.  For system administrators, it’s an especially handy attribute when it’s necessary to recover one small piece of data which has been copied or backed-up as something really huge and there are big machine operators which now routinely handle data sets of a size which only a few years ago were unimaginably large.  For them, the ability to look at the whole and be able to extract pieces, drilling down if need be to individual bytes, makes easily possible what would otherwise require much time and hardware; hence the metaphor of granularity, a mechanism to find a particular grain in a silo of many trillions.

That’s useful but really is just brute-force, the massive up-scaling up of something which has existed since the earliest forms of digital storage.  More intriguing is the recent emergence of Granular computing (GrC), a fork in information processing, the focus of which is information granules, entities created from the processes of data abstraction and derivations from data.  The source and structure of this data is not the imperative; what matters are the relationships (of which there may be many) which can, for example, simultaneously be both the extent of difference and a dependence on indistinguishability.  GrC, as it now exists, is more of a conceptual direction than a coherent process or even a theoretical perspective.  Its most promising implication is perhaps the granules which might form as relationships between previously disparate data sets are explored.  This may allow previously unrealized correlates to be identified, perhaps enabling humanity to mine the accumulate data sets for what Donald Rumsfeld (1932–2021: US Secretary of Defense 1975-1977 & 2001-2006) called the unknown knowns.  Rumsfeld may have been evil but his mind could sparkle and many unknown knowns may await.