Monday, October 20, 2014

The seach for serendipity

During the final days of working on the first draft of my DBA literature review (that is, last Friday), I found two papers which initially seemed interesting but of doubtful relevance. One [1], a scholarly paper on the subject of serendipity, was downloaded just for the fun of reading the paper. The authors describe serendipity as "making discoveries by accident and sagacity of things which one is not on quest of” [is there not a redundant 'of' in that definition?]. This subject has absolutely nothing to do with my research, but it's interesting, as I've noticed that frequently good ideas seem to come from random events.

The second paper is entitled "Cognitive style factors affecting database query performance" [2], and of course was intended for the section on cognitive style. The paper discusses how cognitive style affects the accuracy of SQL statements; an example appears on page 260
This isn't quite the syntax which the authors use as they neglected to use table identifiers for certain fields. As written, one can see that this statement uses implicit joins (otherwise known as SQL-89 syntax) as opposed to the clearer explicit join (aka SQL-92 syntax); see here for further discussion of this topic. Looking at it now, there is no need whatsoever to include the ITEMSB table in the query. But I digress.

At first, I thought that I would write a few lines about the paper, stating that Priority hides SQL from the end users ("I wouldn't know an SQL statement even if it bit me in the finger") thus awarding the paper a very low significance, but it suddenly struck me that if I ignore the SQL part and show how the research described in the paper examines how people solve database problems at work ("show all the clients who didn't purchase anything in 2014"), then the paper has high relevance. I wrote about this before, when someone needed data about products which could only be retrieved by accessing three different screens. The cognitive style will play an important part in how the data is retrieved and how accurate it is. Of course, I wrote a program/developed a report (using SQL) to do the work, so that users will be able to access accurate data whatever their cognitive style. In the end, I wrote nearly two pages about this paper which suddenly became extremely relevant.

I see that my writing style has been strongly influenced by the literature review; I even include references.

[1] Foster A.E. and Ellis D., (2014), "Serendipity and its study", Journal of Documentation, Vol. 70(6), 1015 - 1038
[2] Bowen, P. L., Ferguson, C. B., Lehmann, T. H. and Rohde, F. H. (2003): "Cognitive style factors affecting database query performance", International Journal of Accounting Information Systems, 4(4), 251-273.

Sunday, October 19, 2014

Literature review: first draft completed

"I wrote ten days ago that I've been using this time to progress on the literature review, and indeed my aim is to complete a draft by the end of this holiday period which I can then send to my supervisor". Yesterday was the end of that holiday period, and at exactly 12:15pm, I finished reviewing the final paper which I wanted to include in the review.

The review opens a general, explanatory section which covers topics such as ERP history, future (mobile devices), production strategies, misfits and training. Some papers are quoted and some are reviewed critically; it's difficult to assign a count for the number of papers which were reviewed for this section, but at least 30 papers were referenced.

Hereon, the picture is clearer. There are 7 previous literature surveys reviewed, 24 ERP implementation case studies, 6 papers on computer self-efficacy, 9 on end user computer satisfaction, 7 on cognitive fit, 3 on user ownership, 4 on perceived organisational support and 9 on cognitive style - a total of 63 papers. The whole review weighs in at 93 pages. Obviously, I wasn't trying to be brief.

Today I'll send it off to my supervisor then wait a week for him to wade through it and produce his feedback. I would have preferred feedback after finishing each section, so that I could correct any systematic errors, but he wanted to receive a completed object. It's clear to me that this is only a first draft; there are probably too many direct quotes and the supervisor may feel that there are sections which need bolstering, especially in what might be termed 'management impact'. There are references to this scattered around, but maybe they need to be collected in one place.

Now I need a week's holiday in order to recuperate. For fun, I printed a paper which I found on serendipity; I've only read the beginning of it but it seems interesting.

I note that several cited papers came from one journal, Computers in Human Behaviour. I was thinking that maybe I should try reading this journal on as each issue is published - just for fun!