Jump to content

Talk:Granularity

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Comments

[edit]

This is only my opinion, but this definition doesn't seem to capture the meaning of the term. Granularity is an undesirable property of a system, though in digital systems some degree of granularity is unavoidable. Granularity is what pixellization is in digital imagery, or graininess in traditional photography. Granularity is the opposite of high resolution.

I agree that this article is lacking. It needs expansion, and probably splitting into the very different meanings in different contexts (surely there must also be a geological meaning?) However, granularity is certainly not always undesirable - in taxonomy, thesaurus construction, etc, it has a perfectly good technical meaning which is not at all negative - see Boxes and Arrows --OpenToppedBus - Talk to the driver 14:29, 11 October 2005 (UTC)[reply]

Increased granularity = coarser, not finer

[edit]

The example in the lead is backwards. Granularity is basically the concept that you can tell that things aren't continuous. If granularity increases, that means it's more obvious that they aren't -- so the grains are coarser, not finer.

Anyone want to disagree with that, on reflection? 66.96.28.244 03:21, 4 February 2006 (UTC)[reply]

It seems fine and coarse are used consistently across fields but the term granularity is not. More positions of smaller size means a greater granularity in investments but more grains of smaller size in photographic film or paper means a lower granularity. I think I'll put that in the article. --61.41.222.249 08:07, 17 June 2007 (UTC)[reply]
Yes, you are correct. The essential problem is that granularity does not have a well-defined *degree* from "less" to "more". In computer science related fields, people have been increasingly *misusing* the term "more granular" in the past few years to mean "finer grained". But historically it has had the *opposite* meaning: coarser grained. For example, try searching google scholar for "more granular": http://scholar.google.com/scholar?q=%22more+granular%22 . If you read the citations carefully, you will notice that in some case the meaning is "finer grained" and in other cases "coarser grained": exact opposites. This write-up needs to be updated to be more illuminating and less biased toward the current trend of misuse. -- DBooth (talk) 15:24, 5 January 2010 (UTC)[reply]

I can't understand Fine Grained or Coarse Grain in Design

[edit]

Can anyone describe more detail? —Preceding unsigned comment added by Lehonghai0409 (talkcontribs) 05:30, May 3, 2006

New section, granularity in a Service-Oriented Architecture (IT field)

[edit]

I do not have time for it now but granularity is also a challenge in the field of Service-Oriented Architectures (SOA). Application systems in IT are decomposed to application services (often implemented by Web Services): elementary application activities that can be reused throughout the organization. See also the wiki-page about Service-Oriented Architecture. The question for organizations is, how 'big' should an application service be? More information about this challenge is described in a ZapThink article: http://www.zapthink.com/report.html?id=ZAPFLASH-200783

In Business

[edit]

The existing In Business section badly needs a re-write.

It bizarrely veers off into an opinion about how the term "granularity" is used in "irritating" ways, presumably inaccurately, by business executives. The section then spends a few sentences talking about "Smurfing" which has nothing to do with granularity.

I'm not sure whether Granularity, as it's used in business, requires a separate page.

In business intelligence, and I think business in general, granularity generally refers to the depth at which one is looking at data. Colloquially, one could look at the 30,000-foot level or the level of fine details -- or, obviously, anywhere in between. It's necessary to be able to drill down from the broad data (e.g. a company's overall revenue) to segmented and cross-tabulated data (e.g. product-, geography- and/or period-level data) to better understand what's really happening. Different levels of granularity can show very different pictures -- not necessarily "false" but not entirely "true" either.

(For what it's worth, I would concur with the existing section that the term is thrown around loosely by executives, but I don't think that that is relevant to the wikipedia entry that's supposed to be explaining the term.)Magnabonzo (talk) 23:06, 12 November 2010 (UTC)[reply]

"Grain"

[edit]

The usage of Grain is under discussion, see talk:food grain -- 70.51.202.113 (talk) 05:45, 17 September 2015 (UTC)[reply]