tag:blogger.com,1999:blog-4717220359532645973.post8183409191806587180..comments2023-04-09T05:54:18.997-04:00Comments on Learning Complexity: #el30: The Complexity of Data?keith.hamonhttp://www.blogger.com/profile/08404376705918243534noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-4717220359532645973.post-80996936739409483572019-03-24T23:02:22.910-04:002019-03-24T23:02:22.910-04:00Stephen, thanks for the careful reading and genero...Stephen, thanks for the careful reading and generous response. I suspect that you are correct that I read too much into "data", but then that's one of the things I love about MOOCs. Rather, the MOOCs I enjoy are those that lead me into too much reading. EL30 did that for me.<br /><br />Still, I'm confused by your comment that data cannot be connected as can other entities, if I read you correctly. Data is always connected, even if only in a database. The data in a database where the change in state of one data point can indeed result in a change in state of the other data. <br /><br />For me, the big advantages of electronic data, especially when compared to traditional documents, is that, whereas the data in documents (whether on tablets, papyrus, paper, or disk) tend to have static relationships, the relationships among the data in a database can be more easily rearranged and redrawn to fit any emerging situation. Moreover, modern databases can manage and manipulate so much more data than can traditional documents; thus, any given writing of the data allows an author to deal with much larger chunks of reality--chunks that are not readily apparent to the unaided human eye. Like telescopes and telegraphs, big data lets us see patterns we could not otherwise see.<br /><br />But to my mind, the data in any database is already connected if by nothing other than proximity. Since we cannot collect all data, then any data collected by humans or algorithms is selected on some basis, and that basis establishes a connection among the datum and operates within the dataset even before we perform any other operations upon it.<br /><br />Anyway, that makes sense to me now, but I'm still reading. Thanks again.keith.hamonhttps://www.blogger.com/profile/08404376705918243534noreply@blogger.comtag:blogger.com,1999:blog-4717220359532645973.post-60684231342590569262019-03-18T16:11:05.184-04:002019-03-18T16:11:05.184-04:00Nice article.
I think you read maybe too much int...Nice article.<br /><br />I think you read maybe too much into my use of the term 'data' to start this course. I do not intend data to be a primitive in connectivism or in any other sense (in connectivism, the primitives (if we can call them that) are connected entities, where an entity may be whatever we want it to be (a human, a cricket, a neuron, etc) and where a connection exists between them if a change of state in one can result in a change of state in another. I don't think that 'data' (properly so-called) can be connected in this way (though I could be proven wrong).<br /><br />My use of 'data' here is best understood by the contrast with 'document'. I mean, at best, by 'data', a physical presentation of some information (this is what it has in common with a document) presented as a set of facts, list, table, etc. (this is how it differs from a document). I might also say that a document is typically centrally produced, while a data repository has multiple sources. But that's not a hard-and-fast distinction.<br /><br />The concept of 'physical presentation of some information' is important here. The concept of 'data' should be distinguished from the concept of 'fact', which is what some purport that an instance of data represents. But a fact is a non-physical entity, roughly analagous in status to a proposition (indeed, some would say 'propositions are facts'). As a physical entity, a datum is situated in time; propositions are not (though the truth of a proposition may be).<br /><br />I think you are quite right to point out that the meaning (or maybe more technically, the interpretation) of data is context-dependent. Some physical thing (like a period, say) can represent information (the end of a sentence, say) only if interpreted as such. Data, in and of itself, is not epistemologically foundational (though arguable, from a realist perspective, no epistemology could exist without data).<br /><br />So what I am saying in this first module is this: the way we are representing (or transmitting, or storing) facts and information is changing. Whereas formerly, these were contained in longer, more coherent, linear and formally structured artifacts (ie., documents), we are now beginning to employ relatively unstructured, more finely-grained and non-linear collections of artifacts (data). And (thus) knowledge is not a semantic (and inherent) property of the artifact, but rather an emergent (and hence, recognition-dependent) property of the artifact.<br /><br />You *could* set up a database whereby one datum can change the state of another datum (that is precisely what a neural network is (if we remain neutral on the question of whether the datum actually represents anything) but in databases more broadly conceived (such as, say, a financial ledger) changing the state of a datum is frowned upon, and the preference is to create new data for each new event or state of affairs.<br />Stephen Downeshttps://www.blogger.com/profile/06140591903467372209noreply@blogger.com